notebooks/02 - AWS-Batch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Transcriptome annotation on AWS\n",
    "\n",
    "This notebook execute our workflow using AWS. We explore different [machine types]( https://aws.amazon.com/ec2/instance-types/), from the second generation M5D and M5DN instances that feature 2nd generation Intel Xeon Scalable Processors (Cascade Lake) with a sustained all-core Turbo CPU frequency of 3.1 GHz. M5DN is ideal for network intensive applications because they include a custom Nitro card and Elastic Network Adapter (ENA) device to deliver up to 100 Gbps of network throughput to a single instance. Three number of CPUs for each machine type: 16, 32, and 64 were tested. We will call `configuration` to the combination of `MACHINE_TYPE/CPU`\n",
    "\n",
    "The input data, transcriptome in FASTA file, is partitioned in the [01 - Data Partitioning.ipynb]( 01%20-%20Data%20Partitioning.ipynb). 20 FASTA files are created for each query size and are uploaded to a S3 storage bucket with prefixes: `nopal-2000-<UUID>` for input data and `nopal-results-2000-<UUID>` for the results files (2000 is for the query size processed and will be 2000, 6000 and 10000).\n",
    "\n",
    "This notebook will create an organizational structure for storing the jobs descriptions and logs files. It includes directories following the convention: `QUERY_SIZE/MACHINE_TYPE/CPU`. See next directory tree for more information:\n",
    "\n",
    "```bash\n",
    "└── ./PRJNA320545/\n",
    "      ├── 10000\n",
    "      │   ├── m5d\n",
    "      │   │   ├── 64\n",
    "      │   │   ├── 32\n",
    "      │   │   └── 16\n",
    "      │   ├── m5dn\n",
    "      │   │   ├── 64\n",
    "      │   │   ├── 32\n",
    "      │   │   └── 16\n",
    "      │   └── fasta\n",
    "      ├── 6000\n",
    "      │   ├── m5d\n",
    "      │   │   ├── 64\n",
    "      │   │   ├── 32\n",
    "      │   │   └── 16\n",
    "      │   ├── m5dn\n",
    "      │   │   ├── 64\n",
    "      │   │   ├── 32\n",
    "      │   │   └── 16\n",
    "      │   └── fasta\n",
    "      └── 2000\n",
    "          ├── m5d\n",
    "          │   ├── 64\n",
    "          │   ├── 32\n",
    "          │   └── 16\n",
    "          ├── m5dn\n",
    "          │   ├── 64\n",
    "          │   ├── 32\n",
    "          │   └── 16\n",
    "          └── fasta\n",
    "\n",
    "30 directories\n",
    "```\n",
    "\n",
    "After completion, each configuration, `QUERY_SIZE/MACHINE_TYPE/CPU`, will include 40 files, two per job with extensions: \n",
    "\n",
    " * _job.json.gz: Json with the job description returned with **aws batch describe-jobs**\n",
    " * _log.json.gz: Json with the job logs returned with **aws logs get-log-events --log-group-name /aws/batch/job**\n",
    "\n",
    "```bash\n",
    "└── 2000\n",
    "    ├── m5d\n",
    "        ├── 64\n",
    "        ├── 1bb53a49-cdb3-4434-8839-0263a8f07ea4_job.json.gz\n",
    "        ├── 1bb53a49-cdb3-4434-8839-0263a8f07ea4_log.json.gz\n",
    "        ├── 1ce4660e-c0e7-4cc7-b06d-50bdcfaba5c6_job.json.gz\n",
    "        ├── 1ce4660e-c0e7-4cc7-b06d-50bdcfaba5c6_log.json.gz\n",
    "        ├── 2a375778-8a9f-449a-98bc-9904571d774f_job.json.gz\n",
    "        ├── 2a375778-8a9f-449a-98bc-9904571d774f_log.json.gz\n",
    "        ├── 39219810-ae2d-4f02-93dc-42a4523838e1_job.json.gz\n",
    "        ├── 39219810-ae2d-4f02-93dc-42a4523838e1_log.json.gz\n",
    "        ├── 39ea074e-00c0-4eaa-b9ba-a0dbaca093f0_job.json.gz\n",
    "        ├── 39ea074e-00c0-4eaa-b9ba-a0dbaca093f0_log.json.gz\n",
    "        ├── 409c9a06-70fa-42f5-9ccd-5cfe150139cf_job.json.gz\n",
    "        ├── 409c9a06-70fa-42f5-9ccd-5cfe150139cf_log.json.gz\n",
    "        ├── 5774c8b0-5c43-4ed6-9a09-92f0b9b41683_job.json.gz\n",
    "        ├── 5774c8b0-5c43-4ed6-9a09-92f0b9b41683_log.json.gz\n",
    "        ├── 72f396dd-431a-4214-af97-55bdfc91f0ae_job.json.gz\n",
    "        ├── 72f396dd-431a-4214-af97-55bdfc91f0ae_log.json.gz\n",
    "        ├── 7333d168-9def-4526-8a7f-1598ac6309dc_job.json.gz\n",
    "        ├── 7333d168-9def-4526-8a7f-1598ac6309dc_log.json.gz\n",
    "        ├── 87b38638-de1d-4b20-945d-851f592a7575_job.json.gz\n",
    "        ├── 87b38638-de1d-4b20-945d-851f592a7575_log.json.gz\n",
    "        ├── 908dbc68-2d40-407d-abe3-67fddeb23d07_job.json.gz\n",
    "        ├── 908dbc68-2d40-407d-abe3-67fddeb23d07_log.json.gz\n",
    "        ├── 93b5459b-b2c4-409d-9dd4-e1db8f9b70fb_job.json.gz\n",
    "        ├── 93b5459b-b2c4-409d-9dd4-e1db8f9b70fb_log.json.gz\n",
    "        ├── 97093d57-5cce-4962-a6f7-040eddabc4fd_job.json.gz\n",
    "        ├── 97093d57-5cce-4962-a6f7-040eddabc4fd_log.json.gz\n",
    "        ├── a2421aeb-d5e8-40ad-a353-af6c0adc93c6_job.json.gz\n",
    "        ├── a2421aeb-d5e8-40ad-a353-af6c0adc93c6_log.json.gz\n",
    "        ├── cf25552a-a29c-4818-ba67-1002ee96533b_job.json.gz\n",
    "        ├── cf25552a-a29c-4818-ba67-1002ee96533b_log.json.gz\n",
    "        ├── d278f30f-7437-4290-b892-ef17e8f403dc_job.json.gz\n",
    "        ├── d278f30f-7437-4290-b892-ef17e8f403dc_log.json.gz\n",
    "        ├── d34e4005-73a2-4b6e-9830-91d9fcf77795_job.json.gz\n",
    "        ├── d34e4005-73a2-4b6e-9830-91d9fcf77795_log.json.gz\n",
    "        ├── e285f19b-ad47-41bc-a90e-0776ffcabe4a_job.json.gz\n",
    "        ├── e285f19b-ad47-41bc-a90e-0776ffcabe4a_log.json.gz\n",
    "        ├── f201e0b8-20a2-4cbc-9233-baaf661b27c6_job.json.gz\n",
    "        ├── f201e0b8-20a2-4cbc-9233-baaf661b27c6_log.json.gz\n",
    "        ├── fd230900-874c-40ab-b2cd-f69ecd9597cb_job.json.gz\n",
    "        └── fd230900-874c-40ab-b2cd-f69ecd9597cb_log.json.gz\n",
    "\n",
    "0 directories, 40 files\n",
    "```\n",
    "\n",
    "### AWS Configuration\n",
    "\n",
    "### AWS Access Key\n",
    "\n",
    "This notebook assumes that you created a AWS access key as described [here](https://docs.aws.amazon.com/IAM/latest/UserGuide/id_credentials_access-keys.html#Using_CreateAccessKey)\n",
    "\n",
    "To configure the AWS-cli execute **aws configure**\n",
    "\n",
    "```bash\n",
    "$ aws configure\n",
    "AWS Access Key ID [None]: ................W7MTD\n",
    "AWS Secret Access Key [None]: ....................4dypPikIFg\n",
    "Default region name [None]: us-east1\n",
    "Default output format [None]: json\n",
    "$\n",
    "```\n",
    "\n",
    "### Configuration created in this Notebook \n",
    "\n",
    "This notebook will use the [AWS Batch System](https://aws.amazon.com/batch/) to create 6 [compute environments](https://docs.aws.amazon.com/batch/latest/userguide/compute_environments.html), one for each `configuration` of `MACHINE_TYPE/CPU`. Then, 6 [job queues](https://docs.aws.amazon.com/batch/latest/userguide/job_queues.html) and 6 [job definitions](https://docs.aws.amazon.com/batch/latest/userguide/job_definitions.html) will be created for each compute environment.  \n",
    "\n",
    "Automatically, AWS will create a [Amazon Elastic Container Service](https://aws.amazon.com/ecs/) (ECs) for each compute envionment associated to 6 [Auto Scalling groups](https://docs.aws.amazon.com/autoscaling/ec2/userguide/AutoScalingGroup.html)\n",
    "\n",
    "The notebook creates a complete Amazon Virtual Private Cloud (VPC), including ACL, routes and subnets that will be used for all resources created under this setup. \n",
    "\n",
    "Then, a custom [Amazon Machine Image](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AMIs.html), which is public available with ID: **ami-0dac0383cac1dc96e** in the **us-east-1** zone, will be used by the ECs for launching on-demand instances. However, if you are working in a different zone than the **us-east-1** you should copy the AMI to your working zone following this [tutorial](https://scrapoxy.readthedocs.io/en/master/standard/providers/awsec2/copy_ami_to_region/). The final AMI ID should be defined in the third cell.\n",
    "\n",
    "### Deleting resources\n",
    "\n",
    "After completion (last cell of this notebook), all AWS Batch compute environments, queues and job definitions will be deleted.\n",
    "\n",
    "#### Delete VPC \n",
    "\n",
    "User should go to the AWS console and delete the created [VPV](https://console.aws.amazon.com/vpc/). \n",
    "\n",
    "#### Copying and deleting S3 buckets\n",
    "\n",
    "S3 bucktes can be accessed and deleted from the [AWS S3 console](https://s3.console.aws.amazon.com/s3/) \n",
    "\n",
    "### Set AWS tags for resources\n",
    "\n",
    "We defined two json files for defining Tags for all the AWS resources created by this notebook. This help to control cost and identify resources. Edit these files and chage the tags accordingly to your needs.\n",
    "\n",
    " 1. [aws-tags-s3.json](/edit/config/aws/aws-tags-s3.json) Tags used by S3\n",
    " 2. [aws-tags.json](/edit/config/aws/aws-tags.json) Tags used by all other resources\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading packages from config/init.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run ../config/init.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing AWS-cli configuration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using access_key: ........XVULYDAZGO\n"
     ]
    }
   ],
   "source": [
    "access_key = !aws configure get aws_access_key_id\n",
    "secret_access_key = !aws configure get aws_secret_access_key\n",
    "if access_key and secret_access_key:\n",
    "    print('Using access_key: ........{}'.format(access_key[0][10:]))\n",
    "else:\n",
    "    print('Please, configure AWS-cli before running this notebook')\n",
    "    print('Open a Terminal and run: aws configure')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Defining variables\n",
    "\n",
    "Edit AWS zone and region variable accordingly to your geographical location."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "USE_SPOT = False\n",
    "REGION = 'us-east-1'\n",
    "ZONE = 'us-east-1a'\n",
    "AMI = 'ami-0dac0383cac1dc96e'\n",
    "!aws configure set region {REGION}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Defining global Tags for identifying resources\n",
    "\n",
    "Associate AWS Tags with each resource created by this notebook helps to compile total cost used by AWS.\n",
    "The notebook will use, if exists, a file in the **CONFIG/aws** folder named: **aws-tags.json**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using project tag: cbb-research-dl\n"
     ]
    }
   ],
   "source": [
    "TAGFILE_S3 = None\n",
    "TAGFILE = None\n",
    "TAGDIR = None\n",
    "if os.path.exists(os.path.join(CONFIG, \"aws\", \"aws-tags-s3.json\")):\n",
    "    TAGFILE_S3 = os.path.join(CONFIG, \"aws\", \"aws-tags-s3.json\")\n",
    "if os.path.exists(os.path.join(CONFIG, \"aws\", \"aws-tags.json\")):\n",
    "    TAGFILE = os.path.join(CONFIG, \"aws\", \"aws-tags.json\")\n",
    "    with open(TAGFILE) as fin:\n",
    "        TAGDICT = json.loads(fin.read())\n",
    "    PROJECT = None\n",
    "    for k in TAGDICT['Tags']:\n",
    "        if k['Key'] == 'Project':\n",
    "            PROJECT = k['Value']\n",
    "    if PROJECT:\n",
    "        print(\"Using project tag: {}\".format(PROJECT))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AWS machine types\n",
    "\n",
    "| Instance Size | vCPU | Memory (GiB) | Instance Storage (GiB) | Network Bandwidth (Gbps) | EBS Bandwidth (Mbps) |  \\$/Hour | SPOT \\$/Hour |\n",
    "|---------|----------|----------|-------------|---------------|---------------|-----------|-----------|\n",
    "| m5d.4xlarge | 16 | 64 | 2 x 300 NVMe SSD | Up to 10 | 4,750 | 0.904 | 0.1783 |\n",
    "| m5d.8xlarge | 32 | 128 | 2 x 600 NVMe SSD | 10 | 6,800 | 1.808 | 0.3233 |\n",
    "| m5d.16xlarge | 64 | 256 | 4 x 600 NVMe SSD | 20 | 13,600 | 3.616 | 0.6941 |\n",
    "| m5dn.4xlarge | 16 | 64 | 2 x 300 NVMe SSD | Up to 25 | 4,750 | 1.088 | 0.2058 |\n",
    "| m5dn.8xlarge | 32 | 128 | 2 x 600 NVMe SSD | 25 | 6,800| 2.176 | 0.3193 |\n",
    "| m5dn.16xlarge | 64 | 256 | 4 x 600 NVMe SSD | 75 | 13,600 | 4.352 | 0.6551 |\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "QUERY_SIZES = [2000, 6000, 10000]\n",
    "\n",
    "MACHINE_TYPES = ['m5d', 'm5dn']\n",
    "CPUs = [\n",
    "    {\n",
    "        'name':'4xlarge',\n",
    "        'CPUs': 16,\n",
    "        'memory': 64 * 1000 - 5000\n",
    "    }, \n",
    "    {\n",
    "        'name':'8xlarge',\n",
    "        'CPUs': 32,\n",
    "        'memory': 128 * 1000 - 5000\n",
    "    }, \n",
    "    {\n",
    "        'name':'16xlarge',\n",
    "        'CPUs': 64,\n",
    "        'memory': 256 * 1000 - 5000\n",
    "    }]\n",
    "\n",
    "# Prices from 03/04/2020\n",
    "# EC2 PRICE\n",
    "EC2_PRICE = {\n",
    "    'm5d':{\n",
    "        16: 0.904,\n",
    "        32: 1.808,\n",
    "        64: 3.616\n",
    "    },\n",
    "    'm5dn':{\n",
    "        16: 1.088,\n",
    "        32: 2.176,\n",
    "        64: 4.352\n",
    "    }    \n",
    "}\n",
    "#SPOT PRICE\n",
    "SPOT_PRICE = {\n",
    "    'm5d':{\n",
    "        16: 0.1783,\n",
    "        32: 0.3233,\n",
    "        64: 0.6941\n",
    "    },\n",
    "    'm5dn':{\n",
    "        16: 0.2058,\n",
    "        32: 0.3193,\n",
    "        64: 0.6551\n",
    "    }    \n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using as output directory: /panfs/pan1.be-md.ncbi.nlm.nih.gov/alt_splicing/cloud-transcriptome-annotation/results/PRJNA320545\n"
     ]
    }
   ],
   "source": [
    "result_dir = os.path.join(RESULTS, DATASET)\n",
    "if not os.path.exists(result_dir):\n",
    "    os.mkdir(result_dir) \n",
    "os.chdir(result_dir)\n",
    "print('Using as output directory: {}'.format(result_dir))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create or retrieve AWS S3 storage bucket"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query size: 2000\n",
      "\tin-bucket: nopal-2000-a3f0609c-a795-4750-b46b-917fec2f53ab\n",
      "\tout-bucket: nopal-results2000-a3f0609c-a795-4750-b46b-917fec2f53ab\n",
      "Query size: 6000\n",
      "\tin-bucket: nopal-6000-3ab7042a-20ff-4c77-8d33-b2ef677bd185\n",
      "\tout-bucket: nopal-results6000-3ab7042a-20ff-4c77-8d33-b2ef677bd185\n",
      "Query size: 10000\n",
      "\tin-bucket: nopal-10000-d801420a-be79-44a2-85f3-171eb754bd79\n",
      "\tout-bucket: nopal-results10000-d801420a-be79-44a2-85f3-171eb754bd79\n"
     ]
    }
   ],
   "source": [
    "bucket_list = !aws s3 ls | awk '{print $3}'\n",
    "buckets = {}\n",
    "for q in QUERY_SIZES:\n",
    "    prefix = 'nopal-' + str(q) + '-'\n",
    "    suffix = None\n",
    "    for l in bucket_list:\n",
    "        if prefix in l:\n",
    "            suffix = l.replace('nopal-' + str(q) + '-','')\n",
    "            break\n",
    "    if suffix:\n",
    "        buckets[q] = suffix    \n",
    "\n",
    "for q in QUERY_SIZES:\n",
    "    if q not in buckets:\n",
    "        suffix = str(uuid.uuid4())\n",
    "        inbucket = 'nopal-' + str(q) + '-' + suffix\n",
    "        outbucket = 'nopal-results-' + str(q) + '-' + suffix\n",
    "        buckets[q] = suffix \n",
    "        \n",
    "        !aws s3 mb s3://{inbucket} --region {REGION}          \n",
    "        !aws s3 mb s3://{outbucket} --region {REGION}\n",
    "        if TAGFILE_S3:\n",
    "            !aws s3api put-bucket-tagging --bucket {inbucket} --tagging file://{TAGFILE_S3} \n",
    "            !aws s3api put-bucket-tagging --bucket {outbucket} --tagging file://{TAGFILE_S3} \n",
    "        !aws s3 cp {q}/fasta/ s3://{inbucket}/ --recursive\n",
    "        \n",
    "for q in buckets:\n",
    "    print('Query size: {0}\\n\\tin-bucket: nopal-{0}-{1}\\n\\tout-bucket: nopal-results{0}-{1}'.format(q, buckets[q]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating AWS EC2 components"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating an Internet gateway.\n",
    "\n",
    "https://docs.aws.amazon.com/cli/latest/reference/ec2/create-internet-gateway.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using Internet Gateway: igw-00916243d902a86ed\n"
     ]
    }
   ],
   "source": [
    "igw = !aws ec2 describe-internet-gateways --filters Name=tag:Project,Values={PROJECT}\n",
    "igw = json.loads(''.join(igw))   \n",
    "if 'InternetGateways' in igw and len(igw['InternetGateways']) > 0:\n",
    "    igw = igw['InternetGateways'][0]\n",
    "    print('Using Internet Gateway: {}'.format(igw['InternetGatewayId']))\n",
    "else:\n",
    "    igw = !aws ec2 create-internet-gateway\n",
    "    igw = json.loads(''.join(igw))\n",
    "    if 'InternetGateway' in igw and 'InternetGatewayId' in igw['InternetGateway']:\n",
    "        igw = igw['InternetGateway']\n",
    "        print('Created Internet Gateway: {}'.format(igw['InternetGatewayId']))\n",
    "        if TAGFILE:\n",
    "            igw_id = igw['InternetGatewayId']\n",
    "            !aws ec2 create-tags --resources {igw_id} --cli-input-json file://{TAGFILE}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating the Amazon Virtual Private Cloud  (VPC) and all its componets\n",
    "\n",
    "* VPC: https://docs.aws.amazon.com/cli/latest/reference/ec2/create-vpc.html\n",
    "* ACL: https://docs.aws.amazon.com/cli/latest/reference/ec2/create-network-acl.html\n",
    "* Route Table: https://docs.aws.amazon.com/cli/latest/reference/ec2/create-route-table.html\n",
    "* Subnet: https://docs.aws.amazon.com/cli/latest/reference/ec2/create-subnet.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using VPC: vpc-0c0a9fdd87030446d\n",
      "Subnet subnet-0882777bc17300040 attached to VPC vpc-0c0a9fdd87030446d\n"
     ]
    }
   ],
   "source": [
    "vpc = !aws ec2 describe-vpcs --filters Name=tag:Project,Values={PROJECT}\n",
    "vpc = json.loads(''.join(vpc))\n",
    "if 'Vpcs' in vpc and len(vpc['Vpcs']) > 0:\n",
    "    vpc = vpc['Vpcs'][0]\n",
    "    vpc_id = vpc['VpcId']\n",
    "    igw_id = igw['InternetGatewayId']\n",
    "    print('Using VPC: {}'.format(vpc['VpcId']))\n",
    "    subnet = !aws ec2 describe-subnets --filters \"Name=vpc-id,Values={vpc_id}\"\n",
    "    subnet = json.loads(''.join(subnet))\n",
    "    if 'Subnets' in subnet:\n",
    "        subnet = subnet['Subnets'][0]   \n",
    "        subnet_id = subnet['SubnetId']\n",
    "\n",
    "        print('Subnet {} attached to VPC {}'.format(subnet_id, vpc_id)) \n",
    "        \n",
    "    route = !aws ec2 describe-route-tables  --filters \"Name=vpc-id,Values={vpc_id}\"\n",
    "    route = json.loads(''.join(route))\n",
    "    if 'RouteTables' in route and len(route['RouteTables']) == 1:\n",
    "        route = route['RouteTables'][0]\n",
    "        route_id = route['RouteTableId']\n",
    "    \n",
    "    acl = !aws ec2 describe-network-acls --filters Name=vpc-id,Values={vpc_id}\n",
    "    acl = json.loads(''.join(acl))\n",
    "    if 'NetworkAcls' in acl and len(acl['NetworkAcls']) == 1:\n",
    "        acl = acl['NetworkAcls'][0]\n",
    "        acl_id = acl['NetworkAclId'] \n",
    "    \n",
    "else:\n",
    "    print('No VPC, creating it ..... ')\n",
    "    vpc = !aws ec2 create-vpc --cidr-block 10.0.0.0/16 --amazon-provided-ipv6-cidr-block \n",
    "    vpc = json.loads(''.join(vpc))\n",
    "    if 'Vpc' in vpc:\n",
    "        vpc = vpc['Vpc']\n",
    "        vpc_id = vpc['VpcId']\n",
    "        print('Created VPC: {}'.format(vpc_id))\n",
    "        # adding Tags if file exists\n",
    "        if TAGFILE:        \n",
    "            !aws ec2 create-tags --resources {vpc_id} --cli-input-json file://{TAGFILE}\n",
    "\n",
    "        # Attaching igw\n",
    "        igw_id = igw['InternetGatewayId']\n",
    "        print('Attaching IGW {} to the VPC: {}'.format(igw_id, vpc_id))\n",
    "        !aws ec2 attach-internet-gateway --internet-gateway-id {igw_id} --vpc-id {vpc_id}\n",
    "\n",
    "        # Retrieving created ACL\n",
    "        acl = !aws ec2 describe-network-acls --filters Name=vpc-id,Values={vpc_id}\n",
    "        acl = json.loads(''.join(acl))\n",
    "        if 'NetworkAcls' in acl and len(acl['NetworkAcls']) == 1:\n",
    "            acl = acl['NetworkAcls'][0]\n",
    "            acl_id = acl['NetworkAclId']\n",
    "            # adding Tags if file exists\n",
    "            if TAGFILE:                \n",
    "                print('Tagging ACL {}'.format(acl_id))\n",
    "                !aws ec2 create-tags --resources {acl_id} --cli-input-json file://{TAGFILE}\n",
    "\n",
    "        # Retrieving created routes\n",
    "        route = !aws ec2 describe-route-tables --filters Name=vpc-id,Values={vpc_id}\n",
    "        route = json.loads(''.join(route))\n",
    "        if 'RouteTables' in route and len(route['RouteTables']) == 1:\n",
    "            route = route['RouteTables'][0]\n",
    "            route_id = route['RouteTableId']\n",
    "                \n",
    "            route_igw = !aws ec2 create-route --route-table-id {route_id} --destination-cidr-block 0.0.0.0/0 --gateway-id {igw_id}\n",
    "            route_igw = json.loads(''.join(route_igw))\n",
    "            if 'Return' in route_igw and route_igw['Return']:\n",
    "                print('IGW {} attached to route {}'.format(igw_id, route_id))\n",
    "            \n",
    "            # adding Tags if file exists\n",
    "            if TAGFILE:\n",
    "                print('Tagging Route {}'.format(route_id))\n",
    "                !aws ec2 create-tags --resources {route_id} --cli-input-json file://{TAGFILE}\n",
    "\n",
    "        # Creating Subnets\n",
    "        subnet = !aws ec2 create-subnet --vpc-id {vpc_id} --cidr-block 10.0.0.0/16 --availability-zone {ZONE}\n",
    "        subnet = json.loads(''.join(subnet))\n",
    "        if 'Subnet' in subnet:\n",
    "            subnet = subnet['Subnet']   \n",
    "            subnet_id = subnet['SubnetId']\n",
    "            \n",
    "            print('Subnet {} attached to VPC {}'.format(subnet_id, vpc_id))\n",
    "            !aws ec2 modify-subnet-attribute --subnet-id {subnet_id} --map-public-ip-on-launch\n",
    "            print('Public IPs enable on subnet {}'.format(subnet_id))\n",
    "            \n",
    "            # adding Tags if file exists\n",
    "            if TAGFILE:\n",
    "                print('Tagging subnet {}'.format(subnet_id))\n",
    "                !aws ec2 create-tags --resources {subnet_id} --cli-input-json file://{TAGFILE}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating AWS Batch \n",
    "#### Getting AWSBatchServiceRole information\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/iam/list-roles.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using AmazonEC2SpotFleetRole: arn:aws:iam::036922703339:role/AmazonEC2SpotFleetRole\n",
      "Using AWSBatchServiceRole: arn:aws:iam::036922703339:role/AWSBatchServiceRole\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    awsBatchRole = None\n",
    "    spotFleetRole = None\n",
    "    output = !aws iam list-roles\n",
    "    roles = json.loads(''.join(output))\n",
    "    if 'Roles' in roles:\n",
    "        for r in roles['Roles']:\n",
    "            if 'AWSBatchServiceRole' == r['RoleName']:\n",
    "                awsBatchRole = r\n",
    "                print('Using AWSBatchServiceRole: {}'.format(awsBatchRole['Arn']))\n",
    "            elif 'AmazonEC2SpotFleetRole' == r['RoleName']:\n",
    "                spotFleetRole = r\n",
    "                print('Using AmazonEC2SpotFleetRole: {}'.format(spotFleetRole['Arn']))\n",
    "    if not awsBatchRole:\n",
    "        print('Error: AWSBatchServiceRole not available in your account.')\n",
    "        print('Please, creates the AWSBatchServiceRole before continue as it is described:')\n",
    "        print('https://docs.aws.amazon.com/batch/latest/userguide/service_IAM_role.html')\n",
    "    if not spotFleetRole and USE_SPOT:\n",
    "        print('Error: AmazonEC2SpotFleetRole not available in your account.')\n",
    "        print('Please, creates the AmazonEC2SpotFleetRole before continue as it is described:')\n",
    "        print('https://docs.aws.amazon.com/batch/latest/userguide/spot_fleet_IAM_role.html')\n",
    "except:\n",
    "    print('Error: {}'.format(output))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creation of the AWS Batch System    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating a security group\n",
    " \n",
    " * https://docs.aws.amazon.com/cli/latest/reference/ec2/create-security-group.html\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/ec2/describe-security-groups.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using sg sg-02e8ee80022548d68\n"
     ]
    }
   ],
   "source": [
    "sg = !aws ec2 describe-security-groups --filters Name=tag:Project,Values={PROJECT}\n",
    "sg = json.loads(''.join(sg))\n",
    "if 'SecurityGroups' in sg and len(sg['SecurityGroups']) >= 1:\n",
    "    sg = sg['SecurityGroups'][0]\n",
    "    sg_id = sg['GroupId']\n",
    "    print('Using sg {}'.format(sg_id))\n",
    "else:\n",
    "    sg_name = PROJECT\n",
    "    sg_descr = 'Security Group for project: ' + PROJECT\n",
    "    sg = !aws ec2 create-security-group --group-name {sg_name}  --description \"{sg_descr}\" --vpc-id {vpc_id}\n",
    "    sg = json.loads(''.join(sg))\n",
    "    if 'GroupId' in sg:\n",
    "        sg_id = sg['GroupId']\n",
    "        print('SG {} created'.format(sg_id))\n",
    "        \n",
    "        print('Adding SSH inbound to the sg {}'.format(sg_id))\n",
    "        !aws ec2 authorize-security-group-ingress --group-id {sg_id} --protocol tcp --port 22 --cidr 0.0.0.0/0\n",
    "        \n",
    "        # adding Tags if file exists\n",
    "        if TAGFILE:\n",
    "            print('Tagging sg {}'.format(sg_id))\n",
    "            !aws ec2 create-tags --resources {sg_id} --cli-input-json file://{TAGFILE}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating Compuational Environment\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/create-compute-environment.html\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/describe-compute-environments.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using Computational env: m5d_4xlarge_managed\n",
      "Using Computational env: m5d_8xlarge_managed\n",
      "Using Computational env: m5d_16xlarge_managed\n",
      "Using Computational env: m5dn_4xlarge_managed\n",
      "Using Computational env: m5dn_8xlarge_managed\n",
      "Using Computational env: m5dn_16xlarge_managed\n"
     ]
    }
   ],
   "source": [
    "for m in MACHINE_TYPES:  \n",
    "    for c in CPUs:\n",
    "        instance_type = '{}.{}'.format(m,c['name'])\n",
    "        comp_env_name = '{}_{}_managed'.format(m,c['name'])\n",
    "        comp_env = None\n",
    "        try:\n",
    "            output = !aws batch describe-compute-environments --compute-environments {comp_env_name}\n",
    "            comp_env = json.loads(''.join(output))\n",
    "            if 'computeEnvironments' in comp_env and len(comp_env['computeEnvironments']) == 1:\n",
    "                comp_env = comp_env['computeEnvironments'][0]\n",
    "                print('Using Computational env: {}'.format(comp_env['computeEnvironmentName']))\n",
    "            else:\n",
    "                comp_env = None\n",
    "        except:\n",
    "            print('Error: {}'.format(output))\n",
    "        if not comp_env:\n",
    "            try:\n",
    "                print('Computational env: {} does not exist. Creating it ...'.format(comp_env_name))\n",
    "                create_comp_env_dict(comp_env_name, c['CPUs'], instance_type, AMI, subnet_id, sg_id, awsBatchRole['Arn'], spotFleetRole['Arn'], USE_SPOT, TAGDICT['Tags'], 'aws-nopal')\n",
    "                comp_env_json = os.path.join(CONFIG, \"aws\", 'compute-env-{}.json'.format(comp_env_name))            \n",
    "                output = !aws batch create-compute-environment --cli-input-json file://{comp_env_json}\n",
    "                comp_env = json.loads(''.join(output))\n",
    "                print('Computational env: {} created'.format(comp_env['computeEnvironmentName']))                             \n",
    "            except:\n",
    "                print('Error: {}'.format(output))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating Job queue\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/create-job-queue.html\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/describe-job-queues.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using queue: m5d_4xlarge_queue\n",
      "Using queue: m5d_8xlarge_queue\n",
      "Using queue: m5d_16xlarge_queue\n",
      "Using queue: m5dn_4xlarge_queue\n",
      "Using queue: m5dn_8xlarge_queue\n",
      "Using queue: m5dn_16xlarge_queue\n"
     ]
    }
   ],
   "source": [
    "for m in MACHINE_TYPES:  \n",
    "    for c in CPUs:\n",
    "        queue_name = '{}_{}_queue'.format(m,c['name'])\n",
    "        comp_env_name = '{}_{}_managed'.format(m,c['name'])\n",
    "        queue = None\n",
    "        try:\n",
    "            output = !aws batch describe-job-queues --job-queues {queue_name}\n",
    "            queue = json.loads(''.join(output))\n",
    "            if 'jobQueues' in queue and len(queue['jobQueues']) == 1:\n",
    "                queue = queue['jobQueues'][0]\n",
    "                print('Using queue: {}'.format(queue['jobQueueName']))\n",
    "            else:\n",
    "                queue = None\n",
    "\n",
    "        except:\n",
    "            print('Error: {}'.format(output))\n",
    "        if not queue:\n",
    "            try:\n",
    "                print('Queue: {} does not exist. Creating it ...'.format(queue_name))\n",
    "                queue_template_json = os.path.join(CONFIG, \"aws\", 'queue-template.json')  \n",
    "                queue_json = os.path.join(CONFIG, \"aws\", 'queue-{}.json'.format(queue_name))   \n",
    "                with open(queue_template_json) as fin:\n",
    "                    with open(queue_json, \"w\") as fout:\n",
    "                        queue = json.load(fin)\n",
    "                        queue['jobQueueName'] = queue_name\n",
    "                        queue['computeEnvironmentOrder'][0]['computeEnvironment'] = comp_env_name\n",
    "                        fout.write(json.dumps(queue, indent=4))\n",
    "                output = !aws batch create-job-queue --cli-input-json file://{queue_json}\n",
    "                queue = json.loads(''.join(output))\n",
    "                print('Queue: {} created'.format(queue['jobQueueName']))                             \n",
    "            except:\n",
    "                print('Error: {}'.format(output))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Registering jobs\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/register-job-definition.html\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/describe-job-definitions.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using job definition: m5d_4xlarge\n",
      "Using job definition: m5d_8xlarge\n",
      "Using job definition: m5d_16xlarge\n",
      "Using job definition: m5dn_4xlarge\n",
      "Using job definition: m5dn_8xlarge\n",
      "Using job definition: m5dn_16xlarge\n"
     ]
    }
   ],
   "source": [
    "for m in MACHINE_TYPES: \n",
    "    for c in CPUs:\n",
    "        job_def_name = '{}_{}'.format(m,c['name'])\n",
    "        try:\n",
    "            output = !aws batch describe-job-definitions --job-definition-name {job_def_name}\n",
    "            job = json.loads(''.join(output))\n",
    "            if 'jobDefinitions' in job and len(job['jobDefinitions']) >= 1:\n",
    "                job = job['jobDefinitions'][0]\n",
    "                print('Using job definition: {}'.format(job['jobDefinitionName']))\n",
    "            else:\n",
    "                job = None\n",
    "\n",
    "        except:\n",
    "            print('Error: {}'.format(output))\n",
    "        if not job:       \n",
    "            try:\n",
    "                print('Job: {} does not exist. Creating it ...'.format(job_def_name))\n",
    "                job_template_json = os.path.join(CONFIG, \"aws\", 'job-definition-template.json')  \n",
    "                jon_json = os.path.join(CONFIG, \"aws\", 'job-definition-{}.json'.format(job_def_name))   \n",
    "                with open(job_template_json) as fin:\n",
    "                    with open(jon_json, \"w\") as fout:\n",
    "                        job = json.load(fin)\n",
    "                        job['jobDefinitionName'] = job_def_name\n",
    "                        job['containerProperties']['vcpus'] = c['CPUs']\n",
    "                        job['containerProperties']['memory'] = c['memory']\n",
    "                        job['containerProperties']['environment'].append({'name': 'CPUs' , 'value': str(c['CPUs'])})\n",
    "                        job['containerProperties']['environment'].append({'name': 'MACHINE_TYPE' , 'value': '{}_{}'.format(m,c['name'])})\n",
    "                        fout.write(json.dumps(job, indent=4))\n",
    "                    output = !aws batch register-job-definition --cli-input-json file://{jon_json}\n",
    "                    job = json.loads(''.join(output))\n",
    "                print('Job: {} created'.format(job['jobDefinitionName']))   \n",
    "            except:\n",
    "                print('Error: {}'.format(output))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting the jobs to the AWS Batch\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/submit-job.html\n",
    " \n",
    "Once jobs are submitted you can view the queues in the AWS console:\n",
    "\n",
    " * https://console.aws.amazon.com/batch/\n",
    " \n",
    "This code check if the job id file from AWS exists. If the not exists the job will be marked to be submitted again.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 jobs to be submitted\n"
     ]
    }
   ],
   "source": [
    "bar_length = 20\n",
    "jobs = {}\n",
    "jobs_to_submit = 0\n",
    "for q in QUERY_SIZES:\n",
    "    jobs[q] = {}\n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    for m in MACHINE_TYPES:\n",
    "        jobs[q][m] = {}\n",
    "        for c in CPUs:\n",
    "            jobs[q][m][c['CPUs']] = {}\n",
    "            instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "            if not os.path.exists(instace_dir):\n",
    "                os.makedirs(instace_dir)\n",
    "            os.chdir(instace_dir)\n",
    "            for i in range(1,21):\n",
    "                if os.path.exists('{}.id'.format(i)):                    \n",
    "                    jobs[q][m][c['CPUs']][i] = {'completed': True}\n",
    "                else:\n",
    "                    jobs_to_submit += 1\n",
    "                    jobs[q][m][c['CPUs']][i] = {'completed': False}\n",
    "print('{} jobs to be submitted'.format(jobs_to_submit))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting jobs to AWS Batch\n",
    "\n",
    "Skip the next cell if the jobs were already submitted and the **operations.tsv** file is in each output directory.\n",
    "The notebook will use the job ids in the **operations.tsv** to download the logs and check the job status."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n",
      "10000_20_m5dn_16xlarge    [#-------------------] 5.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##------------------] 10.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###-----------------] 15.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####----------------] 20.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#####---------------] 25.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [######--------------] 30.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#######-------------] 35.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [########------------] 40.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#########-----------] 45.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##########----------] 50.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###########---------] 55.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [############--------] 60.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#############-------] 65.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##############------] 70.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###############-----] 75.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [################----] 80.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [#################---] 85.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [##################--] 90.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [###################-] 95.0% jobs submitted   0\r",
      "10000_20_m5dn_16xlarge    [####################] 100.0% jobs submitted   0\r\n"
     ]
    }
   ],
   "source": [
    "for m in MACHINE_TYPES: \n",
    "    for c in CPUs:\n",
    "        job_def_name = '{}_{}'.format(m,c['name'])\n",
    "        for q in QUERY_SIZES: \n",
    "            q_dir = os.path.join(result_dir, str(q))\n",
    "            instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "            os.chdir(instace_dir)\n",
    "            inbucket = 'nopal-' + str(q) + '-' + buckets[q]\n",
    "            outbucket = 'nopal-results-' + str(q) + '-' + buckets[q]\n",
    "            data = []\n",
    "            submitted = 0\n",
    "            for i in range(1,21):\n",
    "                if i in jobs[q][m][c['CPUs']] and not jobs[q][m][c['CPUs']][i]['completed']:\n",
    "                    submitted += 1\n",
    "                    sample = '{}_{}'.format(q, i)\n",
    "                    job_name = '{}_{}'.format(sample, job_def_name)             \n",
    "                    cmd = 'aws batch submit-job --job-name {} '.format(job_name)\n",
    "                    cmd += '--job-queue {}_queue '.format(job_def_name)\n",
    "                    cmd += '--job-definition {} '.format(job_def_name)\n",
    "                    cmd += '--container-overrides \\'environment=['\n",
    "                    cmd += '{name=AWS_ACCESS_KEY_ID,value=' + access_key[0] + '},'\n",
    "                    cmd += '{name=AWS_SECRET_ACCESS_KEY,value=' + secret_access_key[0] + '},'\n",
    "                    cmd += '{name=SAMPLE,value=' + sample + '},'\n",
    "                    cmd += '{name=INBUCKET,value=' + inbucket + '},'\n",
    "                    cmd += '{name=OUTBUCKET,value=' + outbucket + '}]\\''\n",
    "\n",
    "                    output = !{cmd}\n",
    "                    job_id = json.loads(''.join(output))\n",
    "                    with open('{}.id'.format(i), 'w') as fout:\n",
    "                        fout.write(job_id['jobId'])\n",
    "                progress = i * 100/20\n",
    "                text = \"{0:25s} [{1}] {2:3.1f}% jobs submitted {3:3d}\".format(job_name, \"#\" * i + \"-\" * (bar_length - i), progress, submitted)\n",
    "                print(text, end='\\r')   \n",
    "            print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Retrieving Job description and logs\n",
    "\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/batch/describe-jobs.html\n",
    " * https://docs.aws.amazon.com/cli/latest/reference/logs/get-log-events.html\n",
    " \n",
    "Here we check the AWS logs and mark jobs as completed, failed or not completed. You can run the  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "m5d_4xlarge   [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 2000 Completed  20 Failed   0 No completed   0\n",
      "m5d_4xlarge   [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5d_4xlarge   [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 10000 Completed  20 Failed   0 No completed   0\n",
      "Total jobs 360/360\n"
     ]
    }
   ],
   "source": [
    "total = 0\n",
    "done = 0\n",
    "errors = []\n",
    "total_no_completed = 0\n",
    "for q in QUERY_SIZES:\n",
    "    if q not in jobs:\n",
    "        jobs[q] = {}\n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    for m in MACHINE_TYPES:    \n",
    "        if m not in jobs[q]:\n",
    "            jobs[q][m] = {}\n",
    "        for c in CPUs:    \n",
    "            instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "            job_def_name = '{}_{}'.format(m,c['name'])            \n",
    "            os.chdir(instace_dir)   \n",
    "            \n",
    "            if c['CPUs'] not in jobs[q][m]:\n",
    "                jobs[q][m][c['CPUs']] = {}\n",
    "            \n",
    "            completed = 0\n",
    "            failed = 0\n",
    "            no_completed = 0\n",
    "\n",
    "            for i in range(1,21):\n",
    "                if os.path.exists('{}.id'.format(i)):\n",
    "                    mark_to_delete = False\n",
    "                    with open('{}.id'.format(i)) as fin:\n",
    "                        sample = '{}_{}'.format(q, i)\n",
    "                        job_name = '{}_{}'.format(sample, job_def_name) \n",
    "                        job_id = fin.read()\n",
    "\n",
    "                        # The jobs ids loaded from the operation file are used to download the logs from AWS\n",
    "                        # The logs are check to see if they completed without any error. \n",
    "                        total += 1 \n",
    "                        \n",
    "                        job_file = '{}_job.json.gz'.format(i)\n",
    "                        log_file = '{}_log.json.gz'.format(i)\n",
    "\n",
    "                        if os.path.exists(job_file):\n",
    "                            try:\n",
    "                                with gzip.GzipFile(job_file, 'r') as fin:  \n",
    "                                    job = json.loads(fin.read().decode('utf-8'))\n",
    "                                with gzip.GzipFile(log_file, 'r') as fin:  \n",
    "                                    log = json.loads(fin.read().decode('utf-8'))\n",
    "                                    if aws_log_completed(log['events']):\n",
    "                                        completed += 1\n",
    "                                        done += 1\n",
    "                                        jobs[q][m][c['CPUs']][job_name] = {}\n",
    "                                        jobs[q][m][c['CPUs']][job_name]['job'] = job\n",
    "                                        jobs[q][m][c['CPUs']][job_name]['log'] = log\n",
    "                                    else:\n",
    "                                        failed += 1\n",
    "                                        mark_to_delete = True\n",
    "                            except:\n",
    "                                failed += 1\n",
    "                                mark_to_delete = True\n",
    "                                \n",
    "\n",
    "                        else:\n",
    "                            output = !aws batch describe-jobs --jobs {job_id}\n",
    "                            l = json.loads(''.join(output))\n",
    "                            if 'SUCCEEDED' == l['jobs'][0]['status']:\n",
    "                                logStreamName = l['jobs'][0]['container']['logStreamName']\n",
    "                                output = !aws logs get-log-events --log-group-name /aws/batch/job --log-stream-name {logStreamName}\n",
    "                                log = json.loads(''.join(output))\n",
    "                                if aws_log_completed(log['events']):\n",
    "                                    completed += 1\n",
    "                                    done += 1\n",
    "                                    jobs[q][m][c['CPUs']][job_name] = {}\n",
    "                                    jobs[q][m][c['CPUs']][job_name]['job'] = l\n",
    "                                    with gzip.GzipFile(job_file, 'w') as fout:\n",
    "                                        fout.write(json.dumps(l, indent=2).encode('utf-8'))\n",
    "                                    jobs[q][m][c['CPUs']][job_name]['log'] = log\n",
    "                                    with gzip.GzipFile(log_file, 'w') as fout:\n",
    "                                        fout.write(json.dumps(log, indent=2).encode('utf-8'))\n",
    "                                else:\n",
    "                                    mark_to_delete = True\n",
    "                                    failed += 1\n",
    "\n",
    "                            elif 'FAILED' == l['jobs'][0]['status']:\n",
    "                                mark_to_delete = True\n",
    "                                failed += 1\n",
    "                            else:\n",
    "                                no_completed += 1                        \n",
    "                    if mark_to_delete:\n",
    "                        if os.path.exists(job_file):\n",
    "                            os.remove(job_file)\n",
    "                        if os.path.exists(log_file):\n",
    "                            os.remove(log_file)\n",
    "                        os.remove('{}.id'.format(i))\n",
    "                progress = i * 100/20\n",
    "                text = \"{0:13s} [{1}] {2:3.1f}%\\tQuery: {3} Completed {4:3d} Failed {5:3d} No completed {6:3d}\".format(job_def_name, \"#\" * i + \"-\" * (bar_length - i), progress, q, completed, failed, no_completed)\n",
    "                print(text, end='\\r')\n",
    "            print()\n",
    "print('Total jobs {}/{}'.format(done, total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Processing the results\n",
    "\n",
    "Run this cell after all jobs are terminated to create the data frames for plotting. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "m5d_4xlarge   [#-------------------] 5.0%\tQuery: 2000\r",
      "m5d_4xlarge   [##------------------] 10.0%\tQuery: 2000\r",
      "m5d_4xlarge   [###-----------------] 15.0%\tQuery: 2000\r",
      "m5d_4xlarge   [####----------------] 20.0%\tQuery: 2000\r",
      "m5d_4xlarge   [#####---------------] 25.0%\tQuery: 2000\r",
      "m5d_4xlarge   [######--------------] 30.0%\tQuery: 2000\r",
      "m5d_4xlarge   [#######-------------] 35.0%\tQuery: 2000\r",
      "m5d_4xlarge   [########------------] 40.0%\tQuery: 2000\r",
      "m5d_4xlarge   [#########-----------] 45.0%\tQuery: 2000\r",
      "m5d_4xlarge   [##########----------] 50.0%\tQuery: 2000\r",
      "m5d_4xlarge   [###########---------] 55.0%\tQuery: 2000\r",
      "m5d_4xlarge   [############--------] 60.0%\tQuery: 2000\r",
      "m5d_4xlarge   [#############-------] 65.0%\tQuery: 2000\r",
      "m5d_4xlarge   [##############------] 70.0%\tQuery: 2000\r",
      "m5d_4xlarge   [###############-----] 75.0%\tQuery: 2000\r",
      "m5d_4xlarge   [################----] 80.0%\tQuery: 2000\r",
      "m5d_4xlarge   [#################---] 85.0%\tQuery: 2000\r",
      "m5d_4xlarge   [##################--] 90.0%\tQuery: 2000\r",
      "m5d_4xlarge   [###################-] 95.0%\tQuery: 2000\r",
      "m5d_4xlarge   [####################] 100.0%\tQuery: 2000\r\n",
      "m5d_8xlarge   [#-------------------] 5.0%\tQuery: 2000\r",
      "m5d_8xlarge   [##------------------] 10.0%\tQuery: 2000\r",
      "m5d_8xlarge   [###-----------------] 15.0%\tQuery: 2000\r",
      "m5d_8xlarge   [####----------------] 20.0%\tQuery: 2000\r",
      "m5d_8xlarge   [#####---------------] 25.0%\tQuery: 2000\r",
      "m5d_8xlarge   [######--------------] 30.0%\tQuery: 2000\r",
      "m5d_8xlarge   [#######-------------] 35.0%\tQuery: 2000\r",
      "m5d_8xlarge   [########------------] 40.0%\tQuery: 2000\r",
      "m5d_8xlarge   [#########-----------] 45.0%\tQuery: 2000\r",
      "m5d_8xlarge   [##########----------] 50.0%\tQuery: 2000\r",
      "m5d_8xlarge   [###########---------] 55.0%\tQuery: 2000\r",
      "m5d_8xlarge   [############--------] 60.0%\tQuery: 2000\r",
      "m5d_8xlarge   [#############-------] 65.0%\tQuery: 2000\r",
      "m5d_8xlarge   [##############------] 70.0%\tQuery: 2000\r",
      "m5d_8xlarge   [###############-----] 75.0%\tQuery: 2000\r",
      "m5d_8xlarge   [################----] 80.0%\tQuery: 2000\r",
      "m5d_8xlarge   [#################---] 85.0%\tQuery: 2000\r",
      "m5d_8xlarge   [##################--] 90.0%\tQuery: 2000\r",
      "m5d_8xlarge   [###################-] 95.0%\tQuery: 2000\r",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 2000\r\n",
      "m5d_16xlarge  [#-------------------] 5.0%\tQuery: 2000\r",
      "m5d_16xlarge  [##------------------] 10.0%\tQuery: 2000\r",
      "m5d_16xlarge  [###-----------------] 15.0%\tQuery: 2000\r",
      "m5d_16xlarge  [####----------------] 20.0%\tQuery: 2000\r",
      "m5d_16xlarge  [#####---------------] 25.0%\tQuery: 2000\r",
      "m5d_16xlarge  [######--------------] 30.0%\tQuery: 2000\r",
      "m5d_16xlarge  [#######-------------] 35.0%\tQuery: 2000\r",
      "m5d_16xlarge  [########------------] 40.0%\tQuery: 2000\r",
      "m5d_16xlarge  [#########-----------] 45.0%\tQuery: 2000\r",
      "m5d_16xlarge  [##########----------] 50.0%\tQuery: 2000\r",
      "m5d_16xlarge  [###########---------] 55.0%\tQuery: 2000\r",
      "m5d_16xlarge  [############--------] 60.0%\tQuery: 2000\r",
      "m5d_16xlarge  [#############-------] 65.0%\tQuery: 2000\r",
      "m5d_16xlarge  [##############------] 70.0%\tQuery: 2000\r",
      "m5d_16xlarge  [###############-----] 75.0%\tQuery: 2000\r",
      "m5d_16xlarge  [################----] 80.0%\tQuery: 2000\r",
      "m5d_16xlarge  [#################---] 85.0%\tQuery: 2000\r",
      "m5d_16xlarge  [##################--] 90.0%\tQuery: 2000\r",
      "m5d_16xlarge  [###################-] 95.0%\tQuery: 2000\r",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 2000\r\n",
      "m5dn_4xlarge  [#-------------------] 5.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [##------------------] 10.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [###-----------------] 15.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [####----------------] 20.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [#####---------------] 25.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [######--------------] 30.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [#######-------------] 35.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [########------------] 40.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [#########-----------] 45.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [##########----------] 50.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [###########---------] 55.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [############--------] 60.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [#############-------] 65.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [##############------] 70.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [###############-----] 75.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [################----] 80.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [#################---] 85.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [##################--] 90.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [###################-] 95.0%\tQuery: 2000\r",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 2000\r\n",
      "m5dn_8xlarge  [#-------------------] 5.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [##------------------] 10.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [###-----------------] 15.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [####----------------] 20.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [#####---------------] 25.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [######--------------] 30.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [#######-------------] 35.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [########------------] 40.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [#########-----------] 45.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [##########----------] 50.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [###########---------] 55.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [############--------] 60.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [#############-------] 65.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [##############------] 70.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [###############-----] 75.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [################----] 80.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [#################---] 85.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [##################--] 90.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [###################-] 95.0%\tQuery: 2000\r",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 2000\r\n",
      "m5dn_16xlarge [#-------------------] 5.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##------------------] 10.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###-----------------] 15.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####----------------] 20.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#####---------------] 25.0%\tQuery: 2000\r",
      "m5dn_16xlarge [######--------------] 30.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#######-------------] 35.0%\tQuery: 2000\r",
      "m5dn_16xlarge [########------------] 40.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#########-----------] 45.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##########----------] 50.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###########---------] 55.0%\tQuery: 2000\r",
      "m5dn_16xlarge [############--------] 60.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#############-------] 65.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##############------] 70.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###############-----] 75.0%\tQuery: 2000\r",
      "m5dn_16xlarge [################----] 80.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#################---] 85.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##################--] 90.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###################-] 95.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 2000\r\n",
      "m5d_4xlarge   [#-------------------] 5.0%\tQuery: 6000\r",
      "m5d_4xlarge   [##------------------] 10.0%\tQuery: 6000\r",
      "m5d_4xlarge   [###-----------------] 15.0%\tQuery: 6000\r",
      "m5d_4xlarge   [####----------------] 20.0%\tQuery: 6000\r",
      "m5d_4xlarge   [#####---------------] 25.0%\tQuery: 6000\r",
      "m5d_4xlarge   [######--------------] 30.0%\tQuery: 6000\r",
      "m5d_4xlarge   [#######-------------] 35.0%\tQuery: 6000\r",
      "m5d_4xlarge   [########------------] 40.0%\tQuery: 6000\r",
      "m5d_4xlarge   [#########-----------] 45.0%\tQuery: 6000\r",
      "m5d_4xlarge   [##########----------] 50.0%\tQuery: 6000\r",
      "m5d_4xlarge   [###########---------] 55.0%\tQuery: 6000\r",
      "m5d_4xlarge   [############--------] 60.0%\tQuery: 6000\r",
      "m5d_4xlarge   [#############-------] 65.0%\tQuery: 6000\r",
      "m5d_4xlarge   [##############------] 70.0%\tQuery: 6000\r",
      "m5d_4xlarge   [###############-----] 75.0%\tQuery: 6000\r",
      "m5d_4xlarge   [################----] 80.0%\tQuery: 6000\r",
      "m5d_4xlarge   [#################---] 85.0%\tQuery: 6000\r",
      "m5d_4xlarge   [##################--] 90.0%\tQuery: 6000\r",
      "m5d_4xlarge   [###################-] 95.0%\tQuery: 6000\r",
      "m5d_4xlarge   [####################] 100.0%\tQuery: 6000\r\n",
      "m5d_8xlarge   [#-------------------] 5.0%\tQuery: 6000\r",
      "m5d_8xlarge   [##------------------] 10.0%\tQuery: 6000\r",
      "m5d_8xlarge   [###-----------------] 15.0%\tQuery: 6000\r",
      "m5d_8xlarge   [####----------------] 20.0%\tQuery: 6000\r",
      "m5d_8xlarge   [#####---------------] 25.0%\tQuery: 6000\r",
      "m5d_8xlarge   [######--------------] 30.0%\tQuery: 6000\r",
      "m5d_8xlarge   [#######-------------] 35.0%\tQuery: 6000\r",
      "m5d_8xlarge   [########------------] 40.0%\tQuery: 6000\r",
      "m5d_8xlarge   [#########-----------] 45.0%\tQuery: 6000\r",
      "m5d_8xlarge   [##########----------] 50.0%\tQuery: 6000\r",
      "m5d_8xlarge   [###########---------] 55.0%\tQuery: 6000\r",
      "m5d_8xlarge   [############--------] 60.0%\tQuery: 6000\r",
      "m5d_8xlarge   [#############-------] 65.0%\tQuery: 6000\r",
      "m5d_8xlarge   [##############------] 70.0%\tQuery: 6000\r",
      "m5d_8xlarge   [###############-----] 75.0%\tQuery: 6000\r",
      "m5d_8xlarge   [################----] 80.0%\tQuery: 6000\r",
      "m5d_8xlarge   [#################---] 85.0%\tQuery: 6000\r",
      "m5d_8xlarge   [##################--] 90.0%\tQuery: 6000\r",
      "m5d_8xlarge   [###################-] 95.0%\tQuery: 6000\r",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 6000\r\n",
      "m5d_16xlarge  [#-------------------] 5.0%\tQuery: 6000\r",
      "m5d_16xlarge  [##------------------] 10.0%\tQuery: 6000\r",
      "m5d_16xlarge  [###-----------------] 15.0%\tQuery: 6000\r",
      "m5d_16xlarge  [####----------------] 20.0%\tQuery: 6000\r",
      "m5d_16xlarge  [#####---------------] 25.0%\tQuery: 6000\r",
      "m5d_16xlarge  [######--------------] 30.0%\tQuery: 6000\r",
      "m5d_16xlarge  [#######-------------] 35.0%\tQuery: 6000\r",
      "m5d_16xlarge  [########------------] 40.0%\tQuery: 6000\r",
      "m5d_16xlarge  [#########-----------] 45.0%\tQuery: 6000\r",
      "m5d_16xlarge  [##########----------] 50.0%\tQuery: 6000\r",
      "m5d_16xlarge  [###########---------] 55.0%\tQuery: 6000\r",
      "m5d_16xlarge  [############--------] 60.0%\tQuery: 6000\r",
      "m5d_16xlarge  [#############-------] 65.0%\tQuery: 6000\r",
      "m5d_16xlarge  [##############------] 70.0%\tQuery: 6000\r",
      "m5d_16xlarge  [###############-----] 75.0%\tQuery: 6000\r",
      "m5d_16xlarge  [################----] 80.0%\tQuery: 6000\r",
      "m5d_16xlarge  [#################---] 85.0%\tQuery: 6000\r",
      "m5d_16xlarge  [##################--] 90.0%\tQuery: 6000\r",
      "m5d_16xlarge  [###################-] 95.0%\tQuery: 6000\r",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 6000\r\n",
      "m5dn_4xlarge  [#-------------------] 5.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [##------------------] 10.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [###-----------------] 15.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [####----------------] 20.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [#####---------------] 25.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [######--------------] 30.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [#######-------------] 35.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [########------------] 40.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [#########-----------] 45.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [##########----------] 50.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [###########---------] 55.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [############--------] 60.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [#############-------] 65.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [##############------] 70.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [###############-----] 75.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [################----] 80.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [#################---] 85.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [##################--] 90.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [###################-] 95.0%\tQuery: 6000\r",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 6000\r\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "m5dn_8xlarge  [#-------------------] 5.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [##------------------] 10.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [###-----------------] 15.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [####----------------] 20.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [#####---------------] 25.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [######--------------] 30.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [#######-------------] 35.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [########------------] 40.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [#########-----------] 45.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [##########----------] 50.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [###########---------] 55.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [############--------] 60.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [#############-------] 65.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [##############------] 70.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [###############-----] 75.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [################----] 80.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [#################---] 85.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [##################--] 90.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [###################-] 95.0%\tQuery: 6000\r",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 6000\r\n",
      "m5dn_16xlarge [#-------------------] 5.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##------------------] 10.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###-----------------] 15.0%\tQuery: 6000\r",
      "m5dn_16xlarge [####----------------] 20.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#####---------------] 25.0%\tQuery: 6000\r",
      "m5dn_16xlarge [######--------------] 30.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#######-------------] 35.0%\tQuery: 6000\r",
      "m5dn_16xlarge [########------------] 40.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#########-----------] 45.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##########----------] 50.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###########---------] 55.0%\tQuery: 6000\r",
      "m5dn_16xlarge [############--------] 60.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#############-------] 65.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##############------] 70.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###############-----] 75.0%\tQuery: 6000\r",
      "m5dn_16xlarge [################----] 80.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#################---] 85.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##################--] 90.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###################-] 95.0%\tQuery: 6000\r",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 6000\r\n",
      "m5d_4xlarge   [#-------------------] 5.0%\tQuery: 10000\r",
      "m5d_4xlarge   [##------------------] 10.0%\tQuery: 10000\r",
      "m5d_4xlarge   [###-----------------] 15.0%\tQuery: 10000\r",
      "m5d_4xlarge   [####----------------] 20.0%\tQuery: 10000\r",
      "m5d_4xlarge   [#####---------------] 25.0%\tQuery: 10000\r",
      "m5d_4xlarge   [######--------------] 30.0%\tQuery: 10000\r",
      "m5d_4xlarge   [#######-------------] 35.0%\tQuery: 10000\r",
      "m5d_4xlarge   [########------------] 40.0%\tQuery: 10000\r",
      "m5d_4xlarge   [#########-----------] 45.0%\tQuery: 10000\r",
      "m5d_4xlarge   [##########----------] 50.0%\tQuery: 10000\r",
      "m5d_4xlarge   [###########---------] 55.0%\tQuery: 10000\r",
      "m5d_4xlarge   [############--------] 60.0%\tQuery: 10000\r",
      "m5d_4xlarge   [#############-------] 65.0%\tQuery: 10000\r",
      "m5d_4xlarge   [##############------] 70.0%\tQuery: 10000\r",
      "m5d_4xlarge   [###############-----] 75.0%\tQuery: 10000\r",
      "m5d_4xlarge   [################----] 80.0%\tQuery: 10000\r",
      "m5d_4xlarge   [#################---] 85.0%\tQuery: 10000\r",
      "m5d_4xlarge   [##################--] 90.0%\tQuery: 10000\r",
      "m5d_4xlarge   [###################-] 95.0%\tQuery: 10000\r",
      "m5d_4xlarge   [####################] 100.0%\tQuery: 10000\r\n",
      "m5d_8xlarge   [#-------------------] 5.0%\tQuery: 10000\r",
      "m5d_8xlarge   [##------------------] 10.0%\tQuery: 10000\r",
      "m5d_8xlarge   [###-----------------] 15.0%\tQuery: 10000\r",
      "m5d_8xlarge   [####----------------] 20.0%\tQuery: 10000\r",
      "m5d_8xlarge   [#####---------------] 25.0%\tQuery: 10000\r",
      "m5d_8xlarge   [######--------------] 30.0%\tQuery: 10000\r",
      "m5d_8xlarge   [#######-------------] 35.0%\tQuery: 10000\r",
      "m5d_8xlarge   [########------------] 40.0%\tQuery: 10000\r",
      "m5d_8xlarge   [#########-----------] 45.0%\tQuery: 10000\r",
      "m5d_8xlarge   [##########----------] 50.0%\tQuery: 10000\r",
      "m5d_8xlarge   [###########---------] 55.0%\tQuery: 10000\r",
      "m5d_8xlarge   [############--------] 60.0%\tQuery: 10000\r",
      "m5d_8xlarge   [#############-------] 65.0%\tQuery: 10000\r",
      "m5d_8xlarge   [##############------] 70.0%\tQuery: 10000\r",
      "m5d_8xlarge   [###############-----] 75.0%\tQuery: 10000\r",
      "m5d_8xlarge   [################----] 80.0%\tQuery: 10000\r",
      "m5d_8xlarge   [#################---] 85.0%\tQuery: 10000\r",
      "m5d_8xlarge   [##################--] 90.0%\tQuery: 10000\r",
      "m5d_8xlarge   [###################-] 95.0%\tQuery: 10000\r",
      "m5d_8xlarge   [####################] 100.0%\tQuery: 10000\r\n",
      "m5d_16xlarge  [#-------------------] 5.0%\tQuery: 10000\r",
      "m5d_16xlarge  [##------------------] 10.0%\tQuery: 10000\r",
      "m5d_16xlarge  [###-----------------] 15.0%\tQuery: 10000\r",
      "m5d_16xlarge  [####----------------] 20.0%\tQuery: 10000\r",
      "m5d_16xlarge  [#####---------------] 25.0%\tQuery: 10000\r",
      "m5d_16xlarge  [######--------------] 30.0%\tQuery: 10000\r",
      "m5d_16xlarge  [#######-------------] 35.0%\tQuery: 10000\r",
      "m5d_16xlarge  [########------------] 40.0%\tQuery: 10000\r",
      "m5d_16xlarge  [#########-----------] 45.0%\tQuery: 10000\r",
      "m5d_16xlarge  [##########----------] 50.0%\tQuery: 10000\r",
      "m5d_16xlarge  [###########---------] 55.0%\tQuery: 10000\r",
      "m5d_16xlarge  [############--------] 60.0%\tQuery: 10000\r",
      "m5d_16xlarge  [#############-------] 65.0%\tQuery: 10000\r",
      "m5d_16xlarge  [##############------] 70.0%\tQuery: 10000\r",
      "m5d_16xlarge  [###############-----] 75.0%\tQuery: 10000\r",
      "m5d_16xlarge  [################----] 80.0%\tQuery: 10000\r",
      "m5d_16xlarge  [#################---] 85.0%\tQuery: 10000\r",
      "m5d_16xlarge  [##################--] 90.0%\tQuery: 10000\r",
      "m5d_16xlarge  [###################-] 95.0%\tQuery: 10000\r",
      "m5d_16xlarge  [####################] 100.0%\tQuery: 10000\r\n",
      "m5dn_4xlarge  [#-------------------] 5.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [##------------------] 10.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [###-----------------] 15.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [####----------------] 20.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [#####---------------] 25.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [######--------------] 30.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [#######-------------] 35.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [########------------] 40.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [#########-----------] 45.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [##########----------] 50.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [###########---------] 55.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [############--------] 60.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [#############-------] 65.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [##############------] 70.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [###############-----] 75.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [################----] 80.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [#################---] 85.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [##################--] 90.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [###################-] 95.0%\tQuery: 10000\r",
      "m5dn_4xlarge  [####################] 100.0%\tQuery: 10000\r\n",
      "m5dn_8xlarge  [#-------------------] 5.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [##------------------] 10.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [###-----------------] 15.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [####----------------] 20.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [#####---------------] 25.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [######--------------] 30.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [#######-------------] 35.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [########------------] 40.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [#########-----------] 45.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [##########----------] 50.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [###########---------] 55.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [############--------] 60.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [#############-------] 65.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [##############------] 70.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [###############-----] 75.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [################----] 80.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [#################---] 85.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [##################--] 90.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [###################-] 95.0%\tQuery: 10000\r",
      "m5dn_8xlarge  [####################] 100.0%\tQuery: 10000\r\n",
      "m5dn_16xlarge [#-------------------] 5.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##------------------] 10.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###-----------------] 15.0%\tQuery: 10000\r",
      "m5dn_16xlarge [####----------------] 20.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#####---------------] 25.0%\tQuery: 10000\r",
      "m5dn_16xlarge [######--------------] 30.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#######-------------] 35.0%\tQuery: 10000\r",
      "m5dn_16xlarge [########------------] 40.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#########-----------] 45.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##########----------] 50.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###########---------] 55.0%\tQuery: 10000\r",
      "m5dn_16xlarge [############--------] 60.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#############-------] 65.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##############------] 70.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###############-----] 75.0%\tQuery: 10000\r",
      "m5dn_16xlarge [################----] 80.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#################---] 85.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##################--] 90.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###################-] 95.0%\tQuery: 10000\r",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 10000\r\n"
     ]
    }
   ],
   "source": [
    "for q in QUERY_SIZES:\n",
    "    boxdata_total = []\n",
    "    boxdata_blast = []\n",
    "    boxdata_creation = []\n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    for m in MACHINE_TYPES:        \n",
    "        for c in CPUs:   \n",
    "            data = []\n",
    "            job_def_name = '{}_{}'.format(m,c['name'])\n",
    "            platform = 'AWS\\n{}\\n{}'.format(m.upper(), c['CPUs'])\n",
    "            for i in range(1,21):  \n",
    "                progress = i * 100/20\n",
    "                text = \"{0:13s} [{1}] {2:.1f}%\\tQuery: {3}\".format(job_def_name, \"#\" * i + \"-\" * (bar_length - i), progress, q)\n",
    "                print(text, end='\\r')\n",
    "                sample = '{}_{}'.format(q, i)\n",
    "                job_name = '{}_{}'.format(sample, job_def_name) \n",
    "                if job_name in jobs[q][m][c['CPUs']]:\n",
    "                    if 'job' in jobs[q][m][c['CPUs']][job_name]:\n",
    "                        s = jobs[q][m][c['CPUs']][job_name]['job']['jobs'][0]['startedAt']\n",
    "                        e = jobs[q][m][c['CPUs']][job_name]['job']['jobs'][0]['stoppedAt']                        \n",
    "                        total = (e-s)/(1000*60)\n",
    "                        db, cwl = aws_log_times(jobs[q][m][c['CPUs']][job_name]['log']['events'])\n",
    "                        data.append([sample, total, db, cwl])\n",
    "                        \n",
    "                        boxdata_total.append([total, platform])\n",
    "                        boxdata_blast.append([db, platform])\n",
    "                        boxdata_creation.append([total - (cwl + db), platform])\n",
    "            jobs[q][m][c['CPUs']]['aws'] = pandas.DataFrame(data, columns=['id', 'AWS', 'BlastDB', 'CWL'])\n",
    "            print()\n",
    "    os.chdir(q_dir)\n",
    "    jobs[q]['total'] = pandas.DataFrame(boxdata_total, columns=['Time', 'Platform'])\n",
    "    jobs[q]['total'].to_csv('aws_total.tsv', sep='\\t', index=None)\n",
    "    jobs[q]['blast'] = pandas.DataFrame(boxdata_blast, columns=['Time', 'Platform'])\n",
    "    jobs[q]['blast'].to_csv('aws_blast.tsv', sep='\\t', index=None)\n",
    "    jobs[q]['creation'] = pandas.DataFrame(boxdata_creation, columns=['Time', 'Platform'])\n",
    "    jobs[q]['creation'].to_csv('aws_creation.tsv', sep='\\t', index=None)\n",
    "             "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for q in QUERY_SIZES:\n",
    "    df_total = jobs[q]['total']\n",
    "    df_blast = jobs[q]['blast']\n",
    "    df_creation = jobs[q]['creation']\n",
    "\n",
    "    fig, axs = plt.subplots(figsize=(18,10), ncols=3)\n",
    "\n",
    "    fig.suptitle(locale.format_string(\"%d\", int(q), grouping=True)\n",
    "                         + ' transcripts per query for a total processed: ' \n",
    "                         + locale.format_string(\"%d\", int(q) * 20, grouping=True)\n",
    "                 + '\\n\\n', fontsize=16)\n",
    "\n",
    "    j=0\n",
    "\n",
    "    top = []\n",
    "    low = []\n",
    "    billed = []\n",
    "    cost = []\n",
    "    for p in df_total['Platform'].unique():\n",
    "        f = p.lower().split('\\n')\n",
    "        top.append(df_total[df_total['Platform'] == p]['Time'].max())\n",
    "        low.append(df_total[df_total['Platform'] == p]['Time'].min())\n",
    "        billed.append('$ {}\\n$ {}'.format(\n",
    "            round(df_total[df_total['Platform'] == p]['Time'].sum()/60 * EC2_PRICE[f[1]][int(f[2])], 2),\n",
    "            round(df_total[df_total['Platform'] == p]['Time'].sum()/60 * SPOT_PRICE[f[1]][int(f[2])], 2)\n",
    "        ))\n",
    "        cost.append('{:.2f}¢\\n{:.2f}¢'.format(\n",
    "            df_total[df_total['Platform'] == p]['Time'].sum()/60 * EC2_PRICE[f[1]][int(f[2])] * 100/(int(q) * 20),\n",
    "            df_total[df_total['Platform'] == p]['Time'].sum()/60 * SPOT_PRICE[f[1]][int(f[2])] * 100/(int(q) * 20)\n",
    "        ))\n",
    "\n",
    "\n",
    "    box_plot = sns.boxplot(x=\"Platform\", y=\"Time\", data=df_total, ax=axs[j])\n",
    "    \n",
    "    top_tick = 0.8\n",
    "    low_tick = 2.6\n",
    "    if q == 6000:\n",
    "        top_tick += 1.0\n",
    "        low_tick += 2.4\n",
    "    elif q == 10000:\n",
    "        top_tick += 2.2\n",
    "        low_tick += 7.2\n",
    "\n",
    "    for xtick in box_plot.get_xticks():\n",
    "        box_plot.text(xtick,top[xtick] + top_tick,billed[xtick], \n",
    "                horizontalalignment='center',size='small',weight='normal')\n",
    "\n",
    "    for xtick in box_plot.get_xticks():\n",
    "        box_plot.text(xtick,low[xtick] - low_tick,cost[xtick], \n",
    "                horizontalalignment='center',size='small',weight='normal')\n",
    "\n",
    "    sns.stripplot(x=\"Platform\", y=\"Time\", data=df_total, jitter=True, color=\".3\", ax=axs[j])\n",
    "    axs[j].set_ylabel(\"Total Time (Minutes)\")\n",
    "    axs[j].set_xlabel(\"\\nPlatform\\nMachine Type\\nDisk Type\")\n",
    "    axs[j].set_title(\"AWS Total Billed time\\n per machine type\", fontsize=16)\n",
    "\n",
    "    j = 1\n",
    "\n",
    "    top = []\n",
    "    billed = []\n",
    "    for p in df_creation['Platform'].unique():\n",
    "        f = p.lower().split('\\n')\n",
    "        top.append(df_creation[df_creation['Platform'] == p]['Time'].max())\n",
    "        billed.append('$ {}\\n$ {}'.format(\n",
    "            round(df_creation[df_creation['Platform'] == p]['Time'].sum()/60 * EC2_PRICE[f[1]][int(f[2])], 2),\n",
    "            round(df_creation[df_creation['Platform'] == p]['Time'].sum()/60 * SPOT_PRICE[f[1]][int(f[2])], 2)\n",
    "        ))\n",
    "        \n",
    "    box_plot = sns.boxplot(x=\"Platform\", y=\"Time\", data=df_creation, ax=axs[j])\n",
    "\n",
    "    for xtick in box_plot.get_xticks():\n",
    "        box_plot.text(xtick,top[xtick] + 0.01,billed[xtick], \n",
    "                horizontalalignment='center',size='small',weight='normal')\n",
    "\n",
    "    sns.stripplot(x=\"Platform\", y=\"Time\", data=df_creation, jitter=True, color=\".3\", ax=axs[j])\n",
    "    axs[j].set_ylabel(\"\")\n",
    "    axs[j].set_xlabel(\"\\nPlatform\\nMachine Type\\nDisk Type\")\n",
    "    axs[j].set_title(\"AWS Time for VM\\ncreation, setup and release\", fontsize=16)\n",
    "\n",
    "    j = 2\n",
    "\n",
    "    top = []\n",
    "    billed = []\n",
    "    for p in df_blast['Platform'].unique():\n",
    "        f = p.lower().split('\\n')\n",
    "        top.append(df_blast[df_blast['Platform'] == p]['Time'].max())\n",
    "        billed.append('$ {}\\n$ {}'.format(\n",
    "            round(df_blast[df_blast['Platform'] == p]['Time'].sum()/60 * EC2_PRICE[f[1]][int(f[2])], 2),\n",
    "            round(df_blast[df_blast['Platform'] == p]['Time'].sum()/60 * SPOT_PRICE[f[1]][int(f[2])], 2)\n",
    "        ))\n",
    "        \n",
    "    box_plot = sns.boxplot(x=\"Platform\", y=\"Time\", data=df_blast, ax=axs[j])\n",
    "\n",
    "    for xtick in box_plot.get_xticks():\n",
    "        box_plot.text(xtick,top[xtick] + 0.2,billed[xtick], \n",
    "                horizontalalignment='center',size='small',weight='normal')\n",
    "\n",
    "    sns.stripplot(x=\"Platform\", y=\"Time\", data=df_blast, jitter=True, color=\".3\", ax=axs[j])\n",
    "    axs[j].set_ylabel(\"\")\n",
    "    axs[j].set_xlabel(\"\\nPlatform\\nMachine Type\\nDisk Type\")\n",
    "    axs[j].set_title(\"BlastDB transfer\\n per machine type\", fontsize=16)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Processing 120k transcripts using query sizes of 2k, 6k and 10k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_dir = os.path.join(RESULTS, DATASET, '120k')\n",
    "if not os.path.exists(result_dir):\n",
    "    os.mkdir(result_dir) \n",
    "os.chdir(result_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query size: 2000\n",
      "\tin-bucket: nopal-120k-2000-9f03d34f-ffac-4037-b5b8-da9fae170cd0\n",
      "\tout-bucket: nopal-results-120k-2000-9f03d34f-ffac-4037-b5b8-da9fae170cd0\n",
      "Query size: 6000\n",
      "\tin-bucket: nopal-120k-6000-a9bf8e59-fe75-4b5a-9b99-97c305814f87\n",
      "\tout-bucket: nopal-results-120k-6000-a9bf8e59-fe75-4b5a-9b99-97c305814f87\n",
      "Query size: 10000\n",
      "\tin-bucket: nopal-120k-10000-36858e5d-8b7c-4f69-a730-8ecf3514e287\n",
      "\tout-bucket: nopal-results-120k-10000-36858e5d-8b7c-4f69-a730-8ecf3514e287\n"
     ]
    }
   ],
   "source": [
    "bucket_list = !aws s3 ls | awk '{print $3}'\n",
    "buckets = {}\n",
    "for q in QUERY_SIZES:\n",
    "    prefix = 'nopal-120k-' + str(q) + '-'\n",
    "    suffix = None\n",
    "    for l in bucket_list:\n",
    "        if prefix in l:\n",
    "            suffix = l.replace('nopal-120k-' + str(q) + '-','')\n",
    "            break\n",
    "    if suffix:\n",
    "        buckets[q] = suffix    \n",
    "\n",
    "for q in QUERY_SIZES:\n",
    "    if q not in buckets:\n",
    "        suffix = str(uuid.uuid4())\n",
    "        inbucket = 'nopal-120k-' + str(q) + '-' + suffix\n",
    "        outbucket = 'nopal-results-120k-' + str(q) + '-' + suffix\n",
    "        buckets[q] = suffix \n",
    "        \n",
    "        !aws s3 mb s3://{inbucket} --region {REGION}          \n",
    "        !aws s3 mb s3://{outbucket} --region {REGION}\n",
    "        if TAGFILE_S3:\n",
    "            !aws s3api put-bucket-tagging --bucket {inbucket} --tagging file://{TAGFILE_S3} \n",
    "            !aws s3api put-bucket-tagging --bucket {outbucket} --tagging file://{TAGFILE_S3} \n",
    "        !aws s3 cp {q}/fasta/ s3://{inbucket}/ --recursive\n",
    "        \n",
    "for q in buckets:\n",
    "    print('Query size: {0}\\n\\tin-bucket: nopal-120k-{0}-{1}\\n\\tout-bucket: nopal-results-120k-{0}-{1}'.format(q, buckets[q]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 jobs to be submitted\n"
     ]
    }
   ],
   "source": [
    "os.chdir(result_dir)\n",
    "m = 'm5dn'\n",
    "c = {\n",
    "        'name':'16xlarge',\n",
    "        'CPUs': 64,\n",
    "        'memory': 256 * 1000 - 5000\n",
    "    }\n",
    "\n",
    "bar_length = 20\n",
    "jobs = {}\n",
    "jobs_to_submit = 0\n",
    "for q in QUERY_SIZES:\n",
    "    jobs[q] = {}\n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    samples = [ f.replace('.fa', '') for ds, dr, files in os.walk(os.path.join(q_dir, 'fasta')) for f in files if f.endswith('.fa') ]\n",
    "    \n",
    "    instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "    if not os.path.exists(instace_dir):\n",
    "        os.makedirs(instace_dir)\n",
    "    os.chdir(instace_dir)\n",
    "    j = 1\n",
    "    for s in samples:\n",
    "        if os.path.exists('{}.id'.format(j)):\n",
    "            jobs[q][s] = {'completed': True}\n",
    "        else:\n",
    "            jobs_to_submit += 1\n",
    "            jobs[q][s] = {'completed': False}\n",
    "        j += 1\n",
    "print('{} jobs to be submitted'.format(jobs_to_submit))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2000_60_m5dn_16xlarge     [############################################################] 300.0% jobs submitted  60\n",
      "6000_20_m5dn_16xlarge     [####################] 100.0% jobs submitted  20\n",
      "10000_12_m5dn_16xlarge    [############--------] 60.0% jobs submitted  12\n"
     ]
    }
   ],
   "source": [
    "job_def_name = '{}_{}'.format(m,c['name'])\n",
    "for q in QUERY_SIZES: \n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "    os.chdir(instace_dir)\n",
    "    inbucket = 'nopal-120k-' + str(q) + '-' + buckets[q]\n",
    "    outbucket = 'nopal-results-120k-' + str(q) + '-' + buckets[q]\n",
    "    \n",
    "    samples = [ f.replace('.fa', '') for ds, dr, files in os.walk(os.path.join(q_dir, 'fasta')) for f in files if f.endswith('.fa') ]\n",
    "    bar_length = len(samples)\n",
    "    data = []\n",
    "    submitted = 0\n",
    "    i = 1\n",
    "    for s in samples:\n",
    "        if s in jobs[q] and not jobs[q][s]['completed']:\n",
    "            submitted += 1\n",
    "            job_name = '{}_{}'.format(s, job_def_name)             \n",
    "            cmd = 'aws batch submit-job --job-name {} '.format(job_name)\n",
    "            cmd += '--job-queue {}_queue '.format(job_def_name)\n",
    "            cmd += '--job-definition {} '.format(job_def_name)\n",
    "            cmd += '--container-overrides \\'environment=['\n",
    "            cmd += '{name=AWS_ACCESS_KEY_ID,value=' + access_key[0] + '},'\n",
    "            cmd += '{name=AWS_SECRET_ACCESS_KEY,value=' + secret_access_key[0] + '},'\n",
    "            cmd += '{name=SAMPLE,value=' + s + '},'\n",
    "            cmd += '{name=INBUCKET,value=' + inbucket + '},'\n",
    "            cmd += '{name=OUTBUCKET,value=' + outbucket + '}]\\''\n",
    "\n",
    "            output = !{cmd}\n",
    "            job_id = json.loads(''.join(output))\n",
    "            with open('{}.id'.format(i), 'w') as fout:\n",
    "                fout.write(job_id['jobId'])\n",
    "        progress = i * 100/bar_length\n",
    "        text = \"{0:25s} [{1}] {2:3.1f}% jobs submitted {3:3d}\".format(job_name, \"#\" * i + \"-\" * (bar_length - i), progress, submitted)\n",
    "        print(text, end='\\r')  \n",
    "        i += 1\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "m5dn_16xlarge [############################################################] 100.0%\tQuery: 2000 Completed  60 Failed   0 No completed   0\n",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 6000 Completed  20 Failed   0 No completed   0\n",
      "m5dn_16xlarge [############] 100.0%\tQuery: 10000 Completed  12 Failed   0 No completed   0\n",
      "Total jobs 92/92\n"
     ]
    }
   ],
   "source": [
    "os.chdir(result_dir)\n",
    "total = 0\n",
    "done = 0\n",
    "errors = []\n",
    "total_no_completed = 0\n",
    "for q in QUERY_SIZES:\n",
    "    q_dir = os.path.join(result_dir, str(q))\n",
    "    instace_dir = os.path.join(q_dir, m, str(c['CPUs']))\n",
    "    job_def_name = '{}_{}'.format(m,c['name'])            \n",
    "    os.chdir(instace_dir)  \n",
    "    \n",
    "    samples = [ f.replace('.fa', '') for ds, dr, files in os.walk(os.path.join(q_dir, 'fasta')) for f in files if f.endswith('.fa') ]\n",
    "    bar_length = len(samples)\n",
    "    \n",
    "    completed = 0\n",
    "    failed = 0\n",
    "    no_completed = 0\n",
    "    \n",
    "    j = 1\n",
    "    for s in samples:\n",
    "        if os.path.exists('{}.id'.format(j)):\n",
    "            mark_to_delete = False\n",
    "            with open('{}.id'.format(j)) as fin:\n",
    "                job_name = '{}_{}'.format(s, job_def_name) \n",
    "                job_id = fin.read()\n",
    "\n",
    "                # The jobs ids loaded from the operation file are used to download the logs from AWS\n",
    "                # The logs are check to see if they completed without any error. \n",
    "                total += 1 \n",
    "\n",
    "                job_file = '{}_job.json.gz'.format(s)\n",
    "                log_file = '{}_log.json.gz'.format(s)\n",
    "\n",
    "                if os.path.exists(job_file):\n",
    "                    try:\n",
    "                        with gzip.GzipFile(job_file, 'r') as fin:  \n",
    "                            job = json.loads(fin.read().decode('utf-8'))\n",
    "                        with gzip.GzipFile(log_file, 'r') as fin:  \n",
    "                            log = json.loads(fin.read().decode('utf-8'))\n",
    "                            if aws_log_completed(log['events']):\n",
    "                                completed += 1\n",
    "                                done += 1\n",
    "                                jobs[q][s][job_name] = {}\n",
    "                                jobs[q][s][job_name]['job'] = job\n",
    "                                jobs[q][s][job_name]['log'] = log\n",
    "                            else:\n",
    "                                failed += 1\n",
    "                                mark_to_delete = True\n",
    "                    except:\n",
    "                        failed += 1\n",
    "                        mark_to_delete = True\n",
    "\n",
    "\n",
    "                else:\n",
    "                    output = !aws batch describe-jobs --jobs {job_id}\n",
    "                    l = json.loads(''.join(output))\n",
    "                    if 'SUCCEEDED' == l['jobs'][0]['status']:\n",
    "                        logStreamName = l['jobs'][0]['container']['logStreamName']\n",
    "                        output = !aws logs get-log-events --log-group-name /aws/batch/job --log-stream-name {logStreamName}\n",
    "                        log = json.loads(''.join(output))\n",
    "                        if aws_log_completed(log['events']):\n",
    "                            completed += 1\n",
    "                            done += 1\n",
    "                            jobs[q][s][job_name] = {}\n",
    "                            jobs[q][s][job_name]['job'] = l\n",
    "                            with gzip.GzipFile(job_file, 'w') as fout:\n",
    "                                fout.write(json.dumps(l, indent=2).encode('utf-8'))\n",
    "                            jobs[q][s][job_name]['log'] = log\n",
    "                            with gzip.GzipFile(log_file, 'w') as fout:\n",
    "                                fout.write(json.dumps(log, indent=2).encode('utf-8'))\n",
    "                        else:\n",
    "                            mark_to_delete = True\n",
    "                            failed += 1\n",
    "\n",
    "                    elif 'FAILED' == l['jobs'][0]['status']:\n",
    "                        mark_to_delete = True\n",
    "                        failed += 1\n",
    "                    else:\n",
    "                        no_completed += 1                        \n",
    "            if mark_to_delete:\n",
    "                if os.path.exists(job_file):\n",
    "                    os.remove(job_file)\n",
    "                if os.path.exists(log_file):\n",
    "                    os.remove(log_file)\n",
    "                os.remove('{}.id'.format(j))\n",
    "        progress = j * 100/bar_length\n",
    "        text = \"{0:13s} [{1}] {2:3.1f}%\\tQuery: {3} Completed {4:3d} Failed {5:3d} No completed {6:3d}\".format(job_def_name, \"#\" * j + \"-\" * (bar_length - j), progress, q, completed, failed, no_completed)\n",
    "        print(text, end='\\r')\n",
    "        j += 1\n",
    "    print()\n",
    "print('Total jobs {}/{}'.format(done, total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "m5dn_16xlarge [#-----------------------------------------------------------] 5.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##----------------------------------------------------------] 10.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###---------------------------------------------------------] 15.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####--------------------------------------------------------] 20.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#####-------------------------------------------------------] 25.0%\tQuery: 2000\r",
      "m5dn_16xlarge [######------------------------------------------------------] 30.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#######-----------------------------------------------------] 35.0%\tQuery: 2000\r",
      "m5dn_16xlarge [########----------------------------------------------------] 40.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#########---------------------------------------------------] 45.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##########--------------------------------------------------] 50.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###########-------------------------------------------------] 55.0%\tQuery: 2000\r",
      "m5dn_16xlarge [############------------------------------------------------] 60.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#############-----------------------------------------------] 65.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##############----------------------------------------------] 70.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###############---------------------------------------------] 75.0%\tQuery: 2000\r",
      "m5dn_16xlarge [################--------------------------------------------] 80.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#################-------------------------------------------] 85.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##################------------------------------------------] 90.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###################-----------------------------------------] 95.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####################----------------------------------------] 100.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#####################---------------------------------------] 105.0%\tQuery: 2000\r",
      "m5dn_16xlarge [######################--------------------------------------] 110.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#######################-------------------------------------] 115.0%\tQuery: 2000\r",
      "m5dn_16xlarge [########################------------------------------------] 120.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#########################-----------------------------------] 125.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##########################----------------------------------] 130.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###########################---------------------------------] 135.0%\tQuery: 2000\r",
      "m5dn_16xlarge [############################--------------------------------] 140.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#############################-------------------------------] 145.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##############################------------------------------] 150.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###############################-----------------------------] 155.0%\tQuery: 2000\r",
      "m5dn_16xlarge [################################----------------------------] 160.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#################################---------------------------] 165.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##################################--------------------------] 170.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###################################-------------------------] 175.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####################################------------------------] 180.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#####################################-----------------------] 185.0%\tQuery: 2000\r",
      "m5dn_16xlarge [######################################----------------------] 190.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#######################################---------------------] 195.0%\tQuery: 2000\r",
      "m5dn_16xlarge [########################################--------------------] 200.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#########################################-------------------] 205.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##########################################------------------] 210.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###########################################-----------------] 215.0%\tQuery: 2000\r",
      "m5dn_16xlarge [############################################----------------] 220.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#############################################---------------] 225.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##############################################--------------] 230.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###############################################-------------] 235.0%\tQuery: 2000\r",
      "m5dn_16xlarge [################################################------------] 240.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#################################################-----------] 245.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##################################################----------] 250.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###################################################---------] 255.0%\tQuery: 2000\r",
      "m5dn_16xlarge [####################################################--------] 260.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#####################################################-------] 265.0%\tQuery: 2000\r",
      "m5dn_16xlarge [######################################################------] 270.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#######################################################-----] 275.0%\tQuery: 2000\r",
      "m5dn_16xlarge [########################################################----] 280.0%\tQuery: 2000\r",
      "m5dn_16xlarge [#########################################################---] 285.0%\tQuery: 2000\r",
      "m5dn_16xlarge [##########################################################--] 290.0%\tQuery: 2000\r",
      "m5dn_16xlarge [###########################################################-] 295.0%\tQuery: 2000\r",
      "m5dn_16xlarge [############################################################] 300.0%\tQuery: 2000\r\n",
      "m5dn_16xlarge [#-------------------] 5.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##------------------] 10.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###-----------------] 15.0%\tQuery: 6000\r",
      "m5dn_16xlarge [####----------------] 20.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#####---------------] 25.0%\tQuery: 6000\r",
      "m5dn_16xlarge [######--------------] 30.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#######-------------] 35.0%\tQuery: 6000\r",
      "m5dn_16xlarge [########------------] 40.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#########-----------] 45.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##########----------] 50.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###########---------] 55.0%\tQuery: 6000\r",
      "m5dn_16xlarge [############--------] 60.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#############-------] 65.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##############------] 70.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###############-----] 75.0%\tQuery: 6000\r",
      "m5dn_16xlarge [################----] 80.0%\tQuery: 6000\r",
      "m5dn_16xlarge [#################---] 85.0%\tQuery: 6000\r",
      "m5dn_16xlarge [##################--] 90.0%\tQuery: 6000\r",
      "m5dn_16xlarge [###################-] 95.0%\tQuery: 6000\r",
      "m5dn_16xlarge [####################] 100.0%\tQuery: 6000\r\n",
      "m5dn_16xlarge [#-----------] 5.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##----------] 10.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###---------] 15.0%\tQuery: 10000\r",
      "m5dn_16xlarge [####--------] 20.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#####-------] 25.0%\tQuery: 10000\r",
      "m5dn_16xlarge [######------] 30.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#######-----] 35.0%\tQuery: 10000\r",
      "m5dn_16xlarge [########----] 40.0%\tQuery: 10000\r",
      "m5dn_16xlarge [#########---] 45.0%\tQuery: 10000\r",
      "m5dn_16xlarge [##########--] 50.0%\tQuery: 10000\r",
      "m5dn_16xlarge [###########-] 55.0%\tQuery: 10000\r",
      "m5dn_16xlarge [############] 60.0%\tQuery: 10000\r\n"
     ]
    }
   ],
   "source": [
    "boxdata_total = []\n",
    "for q in QUERY_SIZES:\n",
    "    q_dir = os.path.join(result_dir, str(q))    \n",
    "    data = []\n",
    "    job_def_name = '{}_{}'.format(m,c['name'])\n",
    "    platform = 'AWS\\n{}\\n{}'.format(m.upper(), c['CPUs'])\n",
    "    \n",
    "    samples = [ f.replace('.fa', '') for ds, dr, files in os.walk(os.path.join(q_dir, 'fasta')) for f in files if f.endswith('.fa') ]\n",
    "    bar_length = len(samples)\n",
    "    \n",
    "    j = 1\n",
    "    for s in samples:\n",
    "        progress = j * 100/20\n",
    "        text = \"{0:13s} [{1}] {2:.1f}%\\tQuery: {3}\".format(job_def_name, \"#\" * j + \"-\" * (bar_length - j), progress, q)\n",
    "        print(text, end='\\r')\n",
    "        \n",
    "        job_name = '{}_{}'.format(s, job_def_name) \n",
    "        if job_name in jobs[q][s]:\n",
    "            if 'job' in jobs[q][s][job_name]:\n",
    "                st = jobs[q][s][job_name]['job']['jobs'][0]['startedAt']\n",
    "                e = jobs[q][s][job_name]['job']['jobs'][0]['stoppedAt']                        \n",
    "                total = (e-st)/(1000*60)\n",
    "                db, cwl = aws_log_times(jobs[q][s][job_name]['log']['events'])\n",
    "                data.append([s, total, db, cwl])\n",
    "                boxdata_total.append([total, q])\n",
    "        j += 1\n",
    "    jobs[q]['aws'] = pandas.DataFrame(data, columns=['id', 'AWS', 'BlastDB', 'CWL'])        \n",
    "    print()\n",
    "    os.chdir(q_dir)\n",
    "jobs['total'] = pandas.DataFrame(boxdata_total, columns=['Time', 'Query'])\n",
    "jobs['total'].to_csv(os.path.join(result_dir,'aws_total_120k.tsv'), sep='\\t', index=None)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAAI4CAYAAAB3HEhGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdeZhdVZ3v//e3hqQyQuZOEzDQCaCCoJS2P2xskNYWpYVGQFQ0DC3tGBRBordVtKVJO4DGoTVXoMMFGRrlMoktRtDmqkjCJIiQQCAEM1QGIHNVJd/fH3USKkml6lRl1zlVxfv1POfZZ6+9197fwkeKT621147MRJIkSZK052qqXYAkSZIkDRQGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgddUuYE+MHTs2J0+eXO0yJEkFmz9//srMHFfOuf4ukKSBpzu/B/qafh2wJk+ezLx586pdhiSpYBHxTLnn+rtAkgae7vwe6GucIihJkiRJBTFgSZIkSVJBDFiSJEmSVJB+/QyWJEmS1Ne1tLSwZMkSNm3aVO1S+pyGhgYmTZpEfX19tUspjAFLkiRJ6kVLlixhxIgRTJ48mYiodjl9RmayatUqlixZwv7771/tcgrjFEFJkiSpF23atIkxY8YYrnYSEYwZM2bAjewZsCRJkqReZrjq2ED852LAkiRJkqSC+AyWJEmSVEGTZ9xe6PWenvnOLs+JCM477zy+8Y1vAPD1r3+ddevWcdFFFxVaS2fOOOMMjj/+eE4++eSK3bMaHMGSJEmSBrjBgwfzk5/8hJUrV/aof2tra8EVDVyOYEmSJEkDXF1dHeeccw6XXXYZF1988Q7HnnnmGc466yyampoYN24cV155Jfvttx9nnHEGo0eP5oEHHuB1r3sdI0aMYNGiRSxdupQnnniCSy+9lN/97nfccccd7LPPPtx6663U19fz5S9/mVtvvZWNGzdy5JFH8oMf/GBAPmu1O45gSZIkSS8DH/vYx7jmmmt44YUXdmj/+Mc/zgc/+EEefvhh3v/+9zN9+vTtx5544gl+8YtfbJ9a+OSTT3L77bdz8803c/rpp3PMMcfwhz/8gSFDhnD77bdvv959993HI488wsaNG7ntttsq90P2AQYsSZIk6WVg5MiRfPCDH2TWrFk7tP/2t7/lfe97HwAf+MAHuOeee7YfO+WUU6itrd2+f9xxx1FfX8+hhx7Kli1bePvb3w7AoYceytNPPw3AXXfdxV//9V9z6KGH8stf/pJHH320l3+yvsWAJUmSJL1MfPKTn+Tyyy9n/fr1uz2n/XS+YcOG7XBs8ODBANTU1FBfX7/93JqaGlpbW9m0aRMf/ehHufHGG/nDH/7Ahz70oQH3nquuGLAkSZKkl4nRo0dz6qmncvnll29vO/LII7nuuusAuOaaa/ibv/mbHl9/W5gaO3Ys69at48Ybb9yzgvshF7mQJEmSKqicZdV706c//Wm+853vbN+fNWsWZ511Fl/72te2L3LRU3vvvTcf+tCHOPTQQ5k8eTKvf/3riyi5X4nMrHYNPdbY2Jjz5s2rdhmSpIJFxPzMbCznXH8XSOrrHnvsMV75yldWu4w+q6N/Pt35PdDX9NoUwYi4IiJWRMQj7dpGR8SdEbGgtB3V7thnI2JhRDweEX/fW3VJkiRJUm/pzWew/hN4+05tM4C5mTkVmFvaJyJeBZwGvLrU53sRUYskSZIk9SO9FrAy89fA6p2aTwDmlL7PAU5s135dZm7OzEXAQuANvVWbJEmSJPWGSq8iOCEzlwKUtuNL7fsAz7Y7b0mpbRcRcU5EzIuIeU1NTb1arCRJkiR1R19Zpj06aOtw9Y3MnJ2ZjZnZOG7cuF4uS5IqLzOZP38+d9xxB/4hSZKk/qXSy7Qvj4iJmbk0IiYCK0rtS4B92503CfhzhWuTpD7hsssu4+677wZg0KBBfOUrX+Hggw+ublHq0OQZt1ft3tVe5lmS1LFKB6xbgGnAzNL25nbtP4qIS4G/BKYCv69wbZJUllmzZrFw4cJeuXZLSwvPPffc9v3m5mYuuugixo8f30mvnpsyZQrTp0/vlWtLknbjor0Kvt4LZZ128cUX86Mf/Yja2lpqamr4wQ9+wIUXXsjSpUtpaGhg+PDhXHHFFRx00EE0Nzfzmc98hltvvZWamhpe9apX8d3vfpchQ4Zw7LHHArBs2TJqa2vZNqvs97//PYMGDSr2Z+uHei1gRcS1wNHA2IhYAnyRtmB1Q0ScDSwGTgHIzEcj4gbgj0Ar8LHM3NJbtUlSX9XRuwn78/sKJUl9w29/+1tuu+027r//fgYPHszKlStpbm4G4JprrqGxsZHZs2dzwQUXcMstt/C5z32OtWvX8sQTT1BbW8uVV17JSSedxL333suDDz4IwEUXXcTw4cM5//zzq/mj9Tm9FrAy8727OXTsbs6/GLi4t+qRpKL09ojP5z//eR566CEAampq+MxnPsMRRxzRq/eUJA1sS5cuZezYsQwePBiAsWPH7nLOm9/8Zr75zW+yYcMGrrzyShYtWkRtbdubk84880yuuOIKfvnLX24fwVLH+soiF5Kkkn/5l39hzJgxjBw5kq997WuGK0nSHnvb297Gs88+y4EHHshHP/pRfvWrX+1yzq233sqhhx7KwoUL2W+//Rg5cuQOxxsbG3n00UcrVXK/VelnsCRJXRg8eDAjRowAYOrUqVWuRpI0EAwfPpz58+fzP//zP9x111285z3vYebMmQC8//3vZ8iQIUyePJlvf/vbrF69mohdF/nOzA7btSMDliRJkvQyUFtby9FHH83RRx/NoYceypw5c4CXnsHaZvTo0TzzzDOsXbt2+x/8AO6//37+4R/+oeJ19zdOEZSkCtm8eTNXX301n//857nuuutoaWkBul7EYt26daxbt64SJUqSBqjHH3+cBQsWbN9/8MEHecUrXtHhucOGDWPatGmcd955bNnStu7cVVddxYYNG3jLW95SkXr7M0ewJKlCvvOd72yf8/7QQw/x4IMP8txzz7Fx40aOO+44zjzzTGpqXvq7V3NzM1/4whd4+OGHgbb58x/+8Id3OEeS1A+Vuax6kdatW8cnPvEJnn/+eerq6pgyZQqzZ8/m5JNP7vD8Sy65hPPPP58DDzyQmpoaDj74YG666SanCJYh+vPyv42NjTlv3rxqlyGpD+nNd1TticzkmWee6fScESNGMHjwYIYMGcLTTz9NTU3NLr/Ixo0bx7Bhw3qz1ML15F1bETE/Mxu7PrO6vwt80bCkcjz22GO88pWvrHYZfVZH/3y683ugr3EES9KAsnDhQh549AHYu9qV7GokI6mldrfH165dy9q1a9nKVqImOvwr4eJVi9n0/KbeLLNYz1e7AEmSKsuAJWng2Ru2Hr212lXsYn3Tekb8YQSxJcjahC0Q7Bqiaqgho4MXDpNsbtzM1r363s+2OzV3O51RkvTyYsCSpAppGdfCmqPWULuultYRrQxqGsTQhUOJlqBmy45BZMvQLdRteOlf0VsGb2HDgRvYsteWSpctSSqAS5x3rD8/rrQ7BixJqqCsT1pHtQLQPLGZ5onNsBX2/u3e1G5omz6YNcm6Q9ZR/0I9tWtraRnV0naev5clqV9qaGhg1apVjBkzxpDVTmayatUqGhoaql1KoQxYklRtNfDC619g8HODqWmtYfPEzWwZvsXRKkkaICZNmsSSJUtoamqqdil9TkNDA5MmTap2GYUyYElSH5CDkk3796PFKyRJZauvr2f//fevdhmqEAOWpAFlyZIl8IKLK/QZz8OSXFLtKiRJqhj/C0SSJEmSCuIIlqQBZdKkSTRFU59cpv3lqObuGibtM7Dm1kuS1BkDliRVSX1TPfVr6mkd2UrzBFcJlCRpIDBgSVIVNDzTwLAnhm3f3/jCRjYctKGKFUmSpCL4DJYkVUHD4h3f+dHwbAM4q1GSpH7PgCVJ1bDzdMCEmk3+K1mSpP7O3+aSVAUte7XssB8Eg5YPqlI1kiSpKAYsSaqCljEtu7TloKxCJZIkqUgGLEmqguYJzTuMYrWObGXzX2yuYkWSJKkIriIoSZWWUL+mns37bGbj5I1kbdI6urVHy7Q3LG5gyFNDIGHTvpvYOGVj8fVKkqSyGbAkqcJGPDSCQU1tz1ttrdvKi69/sUfhqu6FOoY9/tJS70MXDaV1ZCst43edfihJkirDgCVp4Hkeau7umzOga7OWQa0vLWZR01rDkN8NYUPdTu/AWlfaDt/9teq31O/SNugPg9hSu6WASgvyPLBPtYuQJKlyDFiSBpQpU6ZUu4RObdq0iWXLlu3QNnroaKaOmcr69evZunUrw4YNY9GGRQBM3Wfqbq+1efNmli5dukPbfmP3Y+jQocUX3lP79P3/TSRJKpIBS9KAMn369GqX0KmtW7fy6U9/mieffBKA2tpaZsyYwRVXXMGqVasAaGlpYfLkydTX1zNr1qxOr3fHHXdw/fXX09rayrve9S5OPfXUXv8ZJEnS7hmwJKmCampq+MpXvsKdd97JmjVrOOqoo1i9ejWLFi3afs66devITAYPHswzzzzDK17xit1e77jjjuO4446rROmSJKkMffMhBUkawIYNG8aJJ57ImWeeudvpc+vXr2f16tWce+65/PrXv65whZIkqaccwZKkbpo1axYLFy4s7HqZSX19PS0tu67+t3XrVr71rW9x4403Fna/9qZMmdLnp1VKktSRiLgCOB5YkZmHlNpGA9cDk4GngVMzc03p2GeBs4EtwPTM/O/eqMsRLEmqsohg4sSJjBkzhr333nuX45lZhaokSerz/hN4+05tM4C5mTkVmFvaJyJeBZwGvLrU53sRUdsbRTmCJUnd1NsjPt///vf56U9/un3/n/7pn3jHO97Rq/eUJKm/ycxfR8TknZpPAI4ufZ8D3A1cWGq/LjM3A4siYiHwBuC3RddlwJKkPuacc87h1a9+NU899RSHHXYYhx9+eLVLkiSpv5iQmUsBMnNpRIwvte8D/K7deUvopTc1GrAkqY+pqanhqKOO4qijjqp2KZIkVcvYiJjXbn92Zs7eg+tFB229MgffgCVJkiSpr1mZmY096Lc8IiaWRq8mAitK7UuAfdudNwn4854W2REXuZAkSZI0UNwCTCt9nwbc3K79tIgYHBH7A1OB3/dGAY5gSZIkSep3IuJa2ha0GBsRS4AvAjOBGyLibGAxcApAZj4aETcAfwRagY9l5pbeqMuAJUmSJKnfycz37ubQsbs5/2Lg4t6rqI1TBCVJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIJUJWBFxLkR8UhEPBoRnyy1jY6IOyNiQWk7qhq1SZIkSVJPVTxgRcQhwIeANwCHAcdHxFRgBjA3M6cCc0v7kiRJktRvVGME65XA7zJzQ2a2Ar8C/hE4AZhTOmcOcGIVapMkSZKkHqtGwHoEeHNEjImIocA7gH2BCZm5FKC0Hd9R54g4JyLmRcS8pqamihUtSZIkSV2peMDKzMeAfwfuBH4GPAS0dqP/7MxszMzGcePG9VKVkiRJktR9VVnkIjMvz8zXZeabgdXAAmB5REwEKG1XVKM2SZIkSeqpaq0iOL603Q84CbgWuAWYVjplGnBzNWqTJEmSpJ6qq9J9fxwRY4AW4GOZuSYiZgI3RMTZwGLglCrVJkmSJEk9UpWAlZlHddC2Cji2CuVIkiRJUiGqMkVQkiRJkgYiA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFaSu2gVIkiT1F5Nn3F7V+z89851Vvb+krjmCJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBamrdgGSJElSXzZ5xu1Vu/fTM99ZtXurZxzBkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkglQlYEXEpyLi0Yh4JCKujYiGiBgdEXdGxILSdlQ1apMkSZKknqp4wIqIfYDpQGNmHgLUAqcBM4C5mTkVmFvalyRJkqR+o1pTBOuAIRFRBwwF/gycAMwpHZ8DnFil2iRJkiSpRyoesDLzOeDrwGJgKfBCZv4cmJCZS0vnLAXGV7o2SZIkSdoT1ZgiOIq20ar9gb8EhkXE6d3of05EzIuIeU1NTb1VpiRJkiR1WzWmCP4dsCgzmzKzBfgJcCSwPCImApS2KzrqnJmzM7MxMxvHjRtXsaIlSZIkqSvVCFiLgTdGxNCICOBY4DHgFmBa6ZxpwM1VqE2SJEmSeqyu0jfMzHsj4kbgfqAVeACYDQwHboiIs2kLYadUujZJkiRJ2hMVD1gAmflF4Is7NW+mbTRLkiRJkvqlai3TLkmSJEk9FhGfiohHI+KRiLg2IhoiYnRE3BkRC0rbUZWuy4AlSZIkqV+JiH2A6UBjZh4C1AKnATOAuZk5FZhb2q8oA5YkSZKk/qgOGBIRdcBQ4M+0vQ5qTun4HODEShdlwJIkSZLU14zd9u7b0uec9gcz8zng67QtjrcUeCEzfw5MyMylpXOWAuMrXXhVFrmQJEmSpE6szMzG3R0sPVt1ArA/8DzwXxFxeqWK64wjWJIkSZL6m78DFmVmU2a2AD8BjgSWR8REgNJ2RaULM2BJkiRJ6m8WA2+MiKEREbS97ukx4BZgWumcacDNlS7MKYKSJEmS+pXMvDcibgTuB1qBB4DZwHDghog4m7YQdkqlazNgSZIkSep3MvOLwBd3at5M22hW1ThFUJIkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCVDxgRcRBEfFgu8+LEfHJiBgdEXdGxILSdlSla5MkSZKkPVHxgJWZj2fm4Zl5OHAEsAG4CZgBzM3MqcDc0r4kSZIk9RvVniJ4LPBkZj4DnADMKbXPAU6sWlWSJEmS1APVDlinAdeWvk/IzKUApe34jjpExDkRMS8i5jU1NVWoTEmSJEnqWtkBKyJGlp6f2i8iYk9vHBGDgHcB/9Wdfpk5OzMbM7Nx3Lhxe1qGJEmSJBWmrrODETEC+AjwPmA4sBJoAMZExD3A9zLzf3p47+OA+zNzeWl/eURMzMylETERWNHD60qSJElSVXQ1gnUT0AQcm5lTMvONpcUp9gO+CbwnIs7q4b3fy0vTAwFuAaaVvk8Dbu7hdSVJkiSpKjodwcrMv9tNewL3lj7dFhFDgbcC/9yueSZwQ0ScDSwGTunJtSVJkiSpWjoNWNtExBuBhzNzQ0S8F3gt8O3MfLYnN83MDcCYndpW0baqoCRJkiT1S+UucjEb2BgRrwE+BywHru61qiRJkiSpHyo3YLWWpgWeAHwrM78BjOi9siRJkiSp/ylriiCwPiIuAD4A/G1E1AD1vVeWJEmSJPU/5Y5gvQcI4J9LLwGeBFzaa1VJkiRJUj9UVsDKzD8DP2rXtAK4oVcqkiRJkqR+qtxVBM8CPg7sBfwVbe/B+h7Q4TLukiRJktQfRcQk4DTgKOAvgY3AI8DtwB2ZubWz/uVOEZwOvBF4ESAznwAm9LBmSZIkSepzIuJK4AqgGfh34L3AR4FfAG8H7omIN3d2jXIXudiUmc0Rse3GtT0tWpIkSZL6qG9k5iMdtD8C/CQiBtE2m2+3yh3B+n8R8RmgISKOAa4HbutWqZIkSZLUh3UUriJiVOl9wGRmc2Yu7Owa5QaszwBrgT8B5wJzaXvhsCRJkiQNKBFxd0SMjIjRwEPAlRFR1irq5Qasj2Tmf2TmP2bmiZn5H8DHelqwJEmSJPVhe2Xmi8BJwJWZeQRlLvBXbsA6q4O2s8vsK0mSJEn9SV1ETAROpZuPRnW6yEVEvIe2JQr3j4iftDs0Ani+u1VKkiRJUj/wJeC/gXsy876IOABYUE7HrlYR/D2wCpgEfLdd+1rggR4UKkmSJEl9VmnF9H0z8zXb2jLzKeDd5fTvNGBl5iJgEW3rvkuSJEnSgJaZWyLiXcBlPelf1nuwImItkO361AKbM3NkT24qSZIkSX3YbyLiO7S9nmr9tsbMvL+rjmUFrMwcse17RNTQtprGYd2vU5IkSZL6vCNL2y+3a0vgLV11LCtgtZeZW4EbI+J84PPd7S9JkiRJfVlmHtPTvuVOEXxXu90aoBGInt5UkiRJkvqyiHgn8GqgYVtbZn559z3alDuCdUq7763A08AJ3ahPkiRJkvqFiPg+MBQ4BvghcDJtK6x3qdxnsD7Q4+okSZIkqX85MjNfExEPZ+aXIuIbwE+67EX5UwTHAmcBk9v3ycxzelCsJEmSJPVlG0vbDRHxl7S9G3j/cjqWO0XwZuB3wD3Alm6XJ0mSJEn9x20RsTfwNeB+2lYQ/GE5HcsNWMMy89M9LE6SJEmS+o3M/NfS1x9HxG1AQ2a+UE7fcgPWHRHxtsz8eY8qlCRJkqQ+LiJO6uQYmdnlc1jlBqwPAxdGxAagmbYl2jMzR5fZX5IkSZL6un/o5FhSxkIX5QassWWeJ0mSJEn9UmaeuafX6DRgRcTUzFxA2wu2OvLwnhYgSZIkSX1BRJzX2fHMvLSra3Q1gjUDOBv4bkfXB97c1Q0kSZIkqWilVf5+CBxCWzY5C3gcuJ6210s9DZyamWu6cdkRe1pXpwErM88ubY/a0xtJkiRJUoG+BfwsM0+OiEHAUOBzwNzMnBkRM2gbMLqw3Atm5pf2tKiupgi+q4sCbtnTAiS95JprruG2225j8ODBnHbaabz97W/v9jVaWlqYN28eEcERRxxBfX19L1QqSZJUPRExkrbZdGcAZGYz0BwRJwBHl06bA9xNNwJWu+tPAr4NvIm20bF7gHMzc0lXfbuaIvh/gQeBR7bdq92xBAxYUkF+85vfcP311wOwfv16vve973HQQQex//5dvzS8paWFhQsXstdee3HxxRfz7LPPArD//vvz1a9+lcGDB/dq7ZIkSQUbGxHz2u3PzszZ7fYPAJqAKyPiMGA+cC4wITOXAmTm0ogY38P7Xwn8CDiltH96qe2tXXXsKmC9p/Q5GLgJuC4zF/WwSGlAmjVrFgsXLtzj66xevXqXti996UsMGjSIZcuWATB8+HBGjx5NTU3N9nOam5tZvnw5W7Zs2aX/okWL+MhHPsLw4cPLqmHKlClMnz69hz+BJElSYVZmZmMnx+uA1wGfyMx7I+JbtE0HLMq4zLyy3f5/RsQny+lY09nBzPyvzDwZOBZ4DvhORPw6InwmSypYR6NMgwcPZsWKFdv3161bxwsv7PgS8TVr1nQYrrbZunVrcUVKkiT1DUuAJZl5b2n/RtoC1/KImAhQ2q7YTf+urIyI0yOitvQ5HVhVTsdy34O1HlheKnASUN6fw6WXgSJHfK6++mpuu+02Bg0axHvf+16mTp3KeeftuFroPvvsw8yZM7fvn3vuuSxa1PHA8ogRI7jssssYNWpUYTVKkiRVW2Yui4hnI+KgzHyctgGhP5Y+04CZpe3NPbzFWcB3gMtoezTqN6W2LnW1yMVRwHuBI2l7QOwHmfm7HhYpqQunn346p59++vb9TZs2MWzYMNavX7+97cADD9yhz1FHHbVDwDrggAM4/PDDqamp4e///u8NV5IkaaD6BHBNaQXBp4AzaZuhd0NEnA0s5qVnqLolMxcDnS74tztdjWD9iraXCf+KtuR2akSc2u7Gnb6IS9KeaWho4Pzzz+fiiy+mtbWVI488ktNOO22Hc9797nfT0NDAfffdx6RJkzj11FOpr69n48aNjBkzpkqVS5Ik9a7MfBDo6DmtY/fkuhFxDG3h7aBS02PAdzLz7nL6dxWwzqEtWEmqgg0bNrBy5UpGjhzJ0KFDmTFj12c3I4Ljjz+e448/HoAf//jHXHvttTQ3N/Pa176WCy+8kKFDh1a6dEmSpH4nIt5J29TALwNfom0V9dcBV0TExzPzp11do6sXDf+wiEIldd/mzZu54IILti+5/vzzz9PU1MS4ceN22+e5557jqquuIrPt7yIPPPAAt912G6eeeupu+0iSJGm7C4ATM/Ohdm0PlpaM/zawZwErIv4D+FZm/qmDY0Nom9PYkpnXdqtsqUKKWkK9GtavX09TU9P2/a1bt3Leeed1+kzVhg0btoerbW666SbuueeeXquzu1wKXpIk9WF/sVO4AiAzH46ICeVcoKspgj8E/i0iDqLtWawmoAGYCowF/hP4bncqlipp4cKFPPHI/ew3fPfLmPdVzQwGRu7Q1vrCn9n0QmeBMQhGk+3ewFC3fimb1j/dKzV21+J1tdUuQZIkqTPre3hsu66mCM4HToqIkcAbgInARtpGtR4tt0qpmvYbvoV/acu/+4sAACAASURBVFxX7TK6rXnrev79jw0s3zQIgGF1W7jwlasYPbjzsPjUumZuf24U61pr+P/GruPoCS9WotyyfGWeb3iQJEl92l9FxC0dtAdwQDkXKOs9WJn5IvCLbhQmqRPPrh/E/zSNoC7gzeNf5C+GtOxyzqCa5PxX/pl5q4bTvDVoHL2evQZ1PRJ3wPDNfOKgZb1RtiRJ0kB3QifHvl7OBcp90bCkgizbWM+lf5pIS7ZN47tv9TA+f8gSRtZv3eXcIbXJUePXVrpESZKkl6XM/NWeXqOm61MkFWn+6mHbwxXAxi21PLRmWKd9NrTWMG/VMJ54sYH0xQmSJEl9VrdGsCJicGZu7q1ipJeD4fW7TvP7r8Vj+H8rR/DeV6zkFcOadzi2fFPbiNf61rYFIl43ah1n/VUTL7bU8tCaoYys38Ihe2+gNipSviRJkjpR1ghWRLwhIv4ALCjtHxYR3+7VyqQB6q/HrGNU/Y7PXG0lWLJhMD98cgJbdxqhmrts5PZwBXD/muE8uGYo//rIPly/eCz/+8kJ/MeCslYNlSRJUjdFxKiIKPtP2eWOYM0Cjgf+L0BmPhQRx/SgPqmilixZwvq1tX1u9brdPVW1prmOf50/klpeeh7reQbtct5VT+5NMy+Frj+9OJQvzNubQbQWXWqhnllby7AlS6pdhiRJUoci4gvADZn5p4gYDPwMOAxojYj3ZWaXC/+V+wxWTWY+s1Nb/3uxkNRH1NDxg1Q1bKGGHRe7GMImaHd+Ha1Eh/2dIyhJkrSH3gM8Xvo+rbQdB/wt8G/lXKDcEaxnI+INQEZELfAJ4IluFCpVxaRJk9jUurTPvQdr3qqtXP30eFqzLRQFyV8Oaea9r1jF5OG7Pub41LoNzF89jJH1Wzhq3FpWbKrnm49P3N5/v6GbOf+Va6jp4xnrK/OG0zBpUrXLkCRJ2p3mzO1Liv09cF1mbgEei4iyslO5AesjtE0T3A9YTts7sT7SzWK3i4i9gR8Ch9D2p/mzaEuK1wOTgaeBUzNzTU/vIfVFWxOufnosv181AoB9h27m3ZNWMWVk52vHHDB8Mwe0C16Th29m2v4ruH7xGNa11lFXk7zQUsuoMt6TJUmSpN3aHBGH0JZ5jgHOb3dsaDkXKGuKYGauyMzTMnNs6XNaZq7sfr3bfQv4WWYeTNucxseAGcDczJwKzC3tSwPKIy8M3R6uAJ7dMJjVLT17Hd1tfx7Futa2vk+ta+D6Z8YUUqMkSdLL2CeBG4E/AZdl5iKAiHgH8EA5Fyjrv+wiYj/g47SNLm3vk5knda9eiIiRwJuBM0rXaAaaI+IE4OjSaXOAu4ELu3t9qS9r2rTr/+WaNtV3+zobtwTLN+24+MUz6wf3uC5JkiRBZv4OOLiD9p8CPy3nGuX+6fwW4CrgTtjpCfzuOwBoAq6MiMOA+cC5wITMXAqQmUsjYnxHnSPiHOAcgP32228PS5Eq65C9N3DLc6PZUnp2qobkNaM2lN1/3uphPPHiEPYdupm/aGhmWbuQNXXEpsLrlSRJejmJiPN2akpgJXDPttGsrpQbsJoz89LuFNfFPV8HfCIz742Ib9GN6YCZORuYDdDY2NjxUmxSHzWhoZWT913FTUtG07w1mDikmZF15T039fOle3HLc6NLeyN2WIlwWN0WTt53VS9ULEmS9LIyooO2ycD/ioiLMvO6ri5QbsD6dkT8C/DfwPYn7TPz4TL7t7cEWJKZ95b2b6QtYC2PiIml0auJwIoeXFvaxeJ1fec9WFsImhjDtiXVn9s4mC8/PIFRvNhpv+UbasiGvXZ4anJru2XZ17fW8vWHR/b592AtXlfLgdUuQpIkaTcy80sdtUfEaNoW+issYB0I/BNwHC9NEUzanqXqlsxcFhHPRsRBmfk4cCzwx9JnGjCztL25u9eWdjZlypRql7CDlStXwrodl4xvqRlCw34HddqvecEC6qKm0zdd1U84mIYhQwqosvccSN/730SSJKkrmbk6Isp6IU65AetUYHJmdr6WdPk+AVwTEYOAp4Azafvb/A0RcTawGDiloHvpZWz69OnVLmEHM2bM4I9//OMObePHj2fTpk2MGTOGadOmcfDBuzxXyfTp09mwYQOrV6+mtbVtlKq2tpYtW9qmF06ZMoWvfvWr1NX1bEVCSZIk7V5EvAUo6xVS5f7X2MO0zUcsJGBl5oNAYweHji3i+lJfdfjhh+8QsGpra1m2bBkAy5Yt48tf/jJXXHEFDQ0Nu/QdOnQol1xyCY8++igHHHAANTU1/PrXv6a1tZV58+Zx0kknccghh/CpT32KcePGVexnkiRJGigi4g/Azus8jAb+DHywnGuUG7DGAH+KiHvZ8Rmsbi/TLr2cnXzyyWzatInf/OY3TJgwgZaWlh0C17p163jiiSd4zWte02H/cePGcfTRR2/ff9/73sfHP/5xFi9eDMAjjzzC97//fT7/+c/36s8hSZI0QB2/034CqzJzfbkXKDdgXVx2SZJ2q66ujjPOOIMzzjgDgKuuumqHgFVXV8e+++5b9vU2bty4PVxt88QTTxRSqyRJ0svQqsxc19kJETG8s3PKCliZObe7lUnq2sknn8xTTz3F/fffz7BhwzjrrLMYNWpU2f2HDBnC/vvvz6JFL72W4VWvelVvlCpJkvRycHNEPEjbgnvzt41cRcQBwDG0rU3xv2lbCb1DnQasiPhVZv5tRKxhx7mIAWRmjt5NV0llGDp0KBdddBFr166loaGB+vr6bl/jggsu4Lvf/S5PPvkkr3nNa/jwhz/cC5VKkiQNfJl5bES8A/hn4E0RMQpoBR4HbgemZeayzq7R1QjWMaXt2D0tVtLujRjR0TvtyjNp0iQuueSSAquRJEl6+crMnwI/7Wn/mi6OzyvdZEtHn57eVJIkSZIGoq4CVlkv05IkSZIkdT1FcFxE7PZNrZk5q+B6JEmSJKnf6moEq5a256/G7eYjSZIkSQNKRPyfcto60tUI1tLM/EKPqpLUK1pbW7nlllt49NFHOeiggzjxxBMZNGhQtcuSJEkaSF7dficiaoEjyunYVcDyGSypj7n88su5/fbbAbjvvvt47rnn+NSnPlXlqiRJkvq/iPgs8DlgSES8uK0ZaAZml3ONrgLW23penvTyMGvWLBYuXNir91iwYAEA06dPZ/HixTscu+uuu3jqqaeI2PO/h0yZMoXp03f72KUkSdKAlpmXAJdExCWZ+dmeXKPTZ7Ays6lHlUkq1JAhQxgyZAgAtbW1Oxyrra0tJFxJkiRpu9siYhhARJweEZdGxCvK6djVCJakLlR6xGf+/PnMnDmTzZs3U19fz3nnnceb3vSmitYgSZI0wP0HcFhEHAZ8BrgcuAr42646GrCkfuaII47gyiuvZOHChRxwwAGMHDmy2iVJkiQNNK2ZmRFxAvCtzLw8IqaV07HTgBURa4Ds6BCQmTm6+7VK2lPDhw/n8MMPr3YZkiRJA9Xa0oIXHwCOKq0iWF9Ox65GsMbuaWWSJEmS1M+8B3gfcFZmLouI/YCvldOxq0UutrT/AHsBE9p9JEmSJGlAycxlwDXAXhFxPLApM68qp2+nAWubiHhnRDwBLAHuLW1/2cN6JUmSJKnPiohTgd8DpwCnAvdGxMnl9C13kYuLgTcBP8/M10bEW4F396RYSZIkSerj/hfw+sxcARAR44BfADd21bGsESzaVtFoAmoiIjLzTuB1Pa1WkiRJkvqwmm3hqmQVZWanckewXii9aOse4KqIWAFs7V6NkiRJktQv/Cwi/hu4trT/HuCOcjqWG7BOBDYBnwQ+SNtiF8d3s0hJkiRJ6vMy84KIOAn4G9peUTU7M28qp2+5UwQ/W1pJsCUzL8/MS4HzelivJEmSJPU5ETElIt4EkJk/yczzMvNTwKqI+KtyrlFuwHp7B23vLLOvJEmSJPUH3wTWdtC+oXSsS51OEYyIfwY+DBwYEfe3OzQCmFdmkZIkSZLUH0zOzId3bszMeRExuZwLdPUM1g3AXOASYEa79rU7raohSZIkSf1dQyfHhpRzgU6nCGbmmsxcmJmnlC741tJnXNklSpIkSVL/cF9EfGjnxog4G5hfzgXKWkUwIj4GfAz4v6WmGyLiu5n5vXIrlSRJkqQ+7pPATRHxfl4KVI3AIOAfy7lAucu0/zPwhsxcBxAR/wb8BjBgSZIkSRoQMnM5cGREHAMcUmq+PTN/We41yg1YAbS0228ptUmSJEnSgJKZdwF39aRvV6sI1mVmK/B/gN9FxI9Lh/4RmNOTG0qSJEnSQNXVCNbvgddl5lcj4i7gKNpGrj6cmff1enWSJEmS1I90FbC2TwMsBSpDlSRJkiTtRlcBa1xEnLe7g5l5acH1SJIkSVK/1VXAqgWG44IWkiRJktSlrgLW0sz8ckUqkSRJkqR+rqaL445cSZIkSVKZugpYx1akCkmSJEkaADoNWJm5ulKFSJIkSVJ/19UIliRJkiSpTAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKkhdNW4aEU8Da4EtQGtmNkbEaOB6YDLwNHBqZq6pRn2SJEmS1BPVHME6JjMPz8zG0v4MYG5mTgXmlvYlSZIkqd/oS1METwDmlL7PAU6sYi2SJEmS1G3VClgJ/Dwi5kfEOaW2CZm5FKC0Hd9Rx4g4JyLmRcS8pqamCpUrSZIkSV2ryjNYwJsy888RMR64MyL+VG7HzJwNzAZobGzM3ipQkiRJkrqrKiNYmfnn0nYFcBPwBmB5REwEKG1XVKM2SZIkSeqpigesiBgWESO2fQfeBjwC3AJMK502Dbi50rVJkiRJ0p6oxhTBCcBNEbHt/j/KzJ9FxH3ADRFxNrAYOKUKtUmSJEnqJyKiFpgHPJeZx/eFVz9VPGBl5lPAYR20rwKOrXQ9kiRJkvqtc4HHgJGl/W2vfpoZETNK+xdWsqC+tEy7JEmSJJUlIiYB7wR+2K656q9+MmBJkiRJ6mvGbns1U+lzTgfnfBP4DLC1XVtZr37qTdVapl2SJEmSdmdlZjbu7mBEHA+syMz5EXF05crqmgFLkiRJUn/zJuBdEfEOoAEYGRFXU3r1U2Yurdarn5wiKEmSJKlfyczPZuakzJwMnAb8MjNPpw+8+smAJUmSJGmgmAm8NSIWAG8t7VeUUwQlSZIk9VuZeTdwd+l71V/95AiWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQQxYkiRJklQQA5YkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQaoWsCKiNiIeiIjbSvujI+LOiFhQ2o6qVm2SJEmS1BPVHME6F3is3f4MYG5mTgXmlvYlSZIkqd+oSsCKiEnAO4Eftms+AZhT+j4HOLHSdUmSJEnSnqjWCNY3gc8AW9u1TcjMpQCl7fhqFCZJkiRJPVXxgBURxwMrMnN+D/ufExHzImJeU1NTwdVJkiRJUs9VYwTrTcC7IuJp4DrgLRFxNbA8IiYClLYrOuqcmbMzszEzG8eNG1epmiVJkiSpSxUPWJn52cyclJmTgdOAX2bm6cAtwLTSadOAmytdmyRJkiTtib70HqyZwFsjYgHw1tK+JEmSJPUbddW8eWbeDdxd+r4KOLaa9UiSJEnSnuhLI1iSJEmS1K8ZsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJEmSCmLAkiRJkqSCGLAkSZIkqSAGLEmSJEkqiAFLkiRJkgpiwJIkSZKkghiwJEmSJKkgBixJkiRJKogBS5IkSZIKYsCSJEmSpIIYsCRJkiSpIAYsSZIkSSqIAUuSJElSvxIR+0bEXRHxWEQ8GhHnltpHR8SdEbGgtB1V6doMWJIkSZL6m1bg05n5SuCNwMci4lXADGBuZk4F5pb2K8qAJUmSJKlfycylmXl/6fta4DFgH+AEYE7ptDnAiZWuzYAlSZIkqd+KiMnAa4F7gQmZuRTaQhgwvtL11FX6hpIkSZLUhbERMa/d/uzMnL3zSRExHPgx8MnMfDEiKlbg7hiwJEmSJPU1KzOzsbMTIqKetnB1TWb+pNS8PCImZubSiJgIrOjtQnfmFEFJkiRJ/Uq0DVVdDjyWmZe2O3QLMK30fRpwc6VrcwRLkiRJUn/zJuADwB8i4sFS2+eAmcANEXE2sBg4pdKFGbAkSZIk9SuZeQ+wuweujq1kLTtziqAkSZIkFcSAJUmSJEkFMWBJkiRJUkEMWJIkSZJUEAOWJEmSJBXEgCVJkiRJBTFgSZIkSVJBDFiSJEmSVBADliRJkiQVxIAlSZIkSQUxYEmSJElSQeqqXYAkSeqBi/aq8v1fqO79JamPcgRLkiRJkgpiwJIkSZKkghiwJEmSJKkgFQ9YEdEQEb+PiIci4tGI+FKpfXRE3BkRC0rbUZWuTZIkSZL2RDVGsDYDb8nMw4DDgbdHxBuBGcDczJwKzC3tS5IkSVK/UfGAlW3WlXbrS58ETgDmlNrnACdWujZJkiRJ2hNVeQYrImoj4kFgBXBnZt4LTMjMpQCl7fjd9D0nIuZFxLympqbKFS1JkiRJXahKwMrMLZl5ODAJeENEHNKNvrMzszEz///27j3OyrLe+/jnOzPAICMIBp4Q0Y0p2EHEB6WsLLU8JWW6wzSRsl77eeqhfKKDWy3UUuqVqey0nYFkInnMMKGkrZbxWKAcQgXbEEmSeUAUGRScw2//cV+Da07MMCzWWjPr+3695jX3fd33fa3fva6Bmd+6rvu6jh48ePDuC9LMzMzMzGwnFXUWwYh4FfgdcDLwgqT9ANL3F4sYmpmZmZmZ2U4rxiyCgyXtlbb7AicCTwP3ARPTaROBuYWOzczMzMzMbFdUFeE19wNukVRJluDdGRH3S/ojcKekzwJ/B84uQmxmZmZmZmZdVvAEKyJWAKPbKH8ZOKHQ8ZiZmZmZmeVLMXqwzMzMzKwrpg4o4mtvKt5rm3UjRZ3kwszMzMzMrCdxgmVmZmZmZpYnHiJoZmZmZlaqPCy023EPlpmZmZmZWZ44wTIzMzMzM8sTJ1hmZmZmZmZ54gTLzMzMzMwsT5xgmZmZmZmZ5YkTrG5o8eLFHH/88SxZsqTYoZiZmZmZWQ5P014C6uvrmTFjBg899BADBgxg0qRJvOc972n3/KlTp9LY2Mhll13G/Pnzqaur49577+Wpp57isMMOY/z48SxbtoyXX36ZcePGMWTIkALejZmZmZlZ+XKCVQLmzZvH/PnzAdi6dSvf//73mTlzJgMHDmx17uLFi6mtrQWgtraWJUuWsGjRIn7zm98AsGzZMh544AFeeeUVAGbPns3VV1/NiBEjCnQ3ZmZmZmbly0MES8CqVaua7dfX17NmzZo2z506dWqz/csuu4xHHnmkWVlTcgWwbds27rvvvvwEamZmZmZmO+QEqwSMHDmy2X5VVVW7PU5NvVe5+3vvvfcO629sbNy1AM3MzMzMrFOcYJWA0047jVNOOYXevXszZMgQpkyZ0ubwQICamppW+xdeeCF9+/YFoE+fPgwbNmz78aqqKk4//fTdF7yZmZmZmW3nZ7CKbP369cyYMYNnn32WE088kUmTJtGnTx/++te/ct9999HQ0MCpp57KqFGjgGyI4JQpU7Zff+WVVzJ69GhmzZrF2rVrGT58OL169eLuu+9m9erVfOADH+Dwww8v1u2ZmZmZmZUVJ1hFFBF85zvf4R//+AcA8+fPp6qqivHjx3PxxRezdetWAB599FGuu+46hg0bxtixY6mpqaG2tpaamhrGjBkDwB577ME73vEOABYsWMAdd9wBwNKlS9mwYQNnn312Ee7QzMzMzKy8OMHKs+nTp7c7QUVLdXV125OrJvPmzWPhwoXbkyvIJr245JJLtg8brKjIRnYOGTKEyZMnt6p3/fr1zfZnz57NI488gqSdupcRI0a0Wb+ZmZmZmbXNCVYRVVW1fvsbGhqazQLY8twtW7YA0L9//+2JFmQTWbz88su88cYbntTCzMzMzKxInGDl2c70+NTX13PmmWe2Ko+IZvujRo3i8ssv5/HHH+e73/3u9sRqw4YNHHPMMRx33HE8/PDDLFiwoM3XOfPMM5k4ceJO3IWZmZmZmXWFE6wiqqqqorq6utlwwFwVFRVMmDCBM844gz59+jBnzpxW58ybN4958+a1OfzvkEMOYf/99+ejH/1o3mM3MzMzM7PWyirB2pnnowohItpNriAb9jdnzhx+/vOfs+eee/Laa6/tsK6W1q5dy9q1a1m4cCGDBg2if//+eYk7n/ycl5mZmZn1JGWVYK1Zs4ZlT6ykcY9BxQ4lCQZVQEdzT0TEDpOrzti4cSPPvvQqW+m9S/XkU8XrG4sdgpmZmZlZXpXVQsPZ7Hqte3qKparAsexR0UgV9QV9zR2LVjMempmZmZl1Z2WVYJWafhV1HfZe5ZMENRWllGCZmZmZmfUsZTVEcOjQobywrYqto04vdigADPrLL+ioRy2AtnKwhopeiIDG+p3KklVRydbDS+P+q1fez9Ch+xY7DDMzMzOzvCmrBAuy536qV95f7DAAeFPQp4PsqL0OroqGN9kSvajZyT7IuoYomfvPnsFygmVmZmZmPUdZJVgjRowodgjNbNiwgdra2u37ktqcDbAtkqhRPZWVlTQ0NHTq/OrqaoYNHtxsgeLi2rfk2sTMzMzMbFeUVYJVatOBT5gwodn+ziRYTS666CLmzJnDc8891+45w4cP53vf+x7V1dVditPMzMzMzDqnVLoyytLAgQOb7fft23en6xgyZAhXXXXVDs955plnuOSSS2hsbNzp+s3MzMzMrPOcYBXRpEmT6N07W5eqsrKSU089lX79+rV7fq9evVqV7b///qxbt67D11q9ejUrV67serBmZmZmZtYhJ1hFNHbsWG6++WY+97nPUVNTw1133UVdXR3HHnssQ4cObXZur169GDJkSKs6Nm/evMPhgbmqqspqRKiZmZmZWcH5L+48mz59OmvWrNmpa55//nm2bt0KwJtvvsnixYvZb7/96NWrF3V1dVRUVLDXXnuxadOmZtdVVFQwbdo03nzzzVZ19u/fn82bN29/pqu6upobb7xxp+IaMWJEyT23ZmZmZmZWypxglYC6urpm+42NjVRUVHDAAQdQX5/NFNg0C2B9ff32ZGzw4MFIok+fPgwcOJBNmzYREey5554MHDiQ/v378/rrr1NZWckee+xRjFszMzMzMysrTrDyrCs9PjNnzmTu3Lnb99/97ndz5ZVXtnv+tm3b6N27N1LzVbIaGxuJCCorK3c6BjMzMzMz23VOsErA+eefT79+/Vi+fDkHH3ww55xzzg7P79OnT5vlpbO+lZmZmZlZeXKCVQJ69erFhAkTWq2LZWZmZmZm3Yu7PMzMzMzMzPLECZaZmZmZmVmeOMEyMzMzMzPLEydYZmZmZmZmeeIEy8zMzMzMLE+cYJmZmZmZmeWJEywzMzMzM7M8cYJlZmZmZmaWJ06wzMzMzMzM8sQJlpmZmZmZWZ44wTIzMzMzM8sTJ1hmZmZmZmZ54gTLzMzMzMwsTwqeYEk6UNLDklZJekrSl1L5IEm/lbQ6fR9Y6NjMzMzMzMx2RTF6sOqBr0TESOBY4AuSRgHfAB6MiEOBB9O+mZmZmZlZt1HwBCsi/hkRS9P2ZmAVcAAwHrglnXYL8LFCx2ZmZmZmZrYrivoMlqThwGhgEbBPRPwTsiQMGNLONZ+X9Likx1966aVChWpmZmZmZtahoiVYkmqAe4AvR8Rrnb0uIm6KiKMj4ujBgwfvvgDNzMzMzMx2UlESLEm9yJKr2yLiF6n4BUn7peP7AS8WIzYzMzMzM7OuKsYsggJmAqsi4gc5h+4DJqbticDcQsdmZmZmZmbdg6STJf1F0hpJJTNBXjF6sN4LfBr4kKTl6etUYBpwkqTVwElp38zMzMzMrBlJlcANwCnAKOCcNDN50VUV+gUjYiGgdg6fUMhYzMzMzMysWxoLrImItQCSbieblXxlUaMCFBHFjqHLJL0ErCt2HEXyNmBDsYOwgnKbl59ybvODIqJTMxmV8e+Ccv75KGdu9/JTrm1+GPCXnP2bIuKmph1JZwEnR8SFaf/TwDER8cXChtlawXuw8qmzv3x7IkmPR8TRxY7DCsdtXn7c5p1Trr8L/PNRntzu5cdt3q62RsSVRM9RUdfBMjMzMzMz64L1wIE5+0OB54oUSzNOsMzMzMzMrLt5DDhU0sGSegMTyGYlL7puPUSwzN3U8SnWw7jNy4/b3HbEPx/lye1eftzmbYiIeklfBB4AKoGbI+KpIocFdPNJLszMzMzMzEqJhwiamZmZmZnliRMsMzMzMzOzPHGCVSIkHSjpYUmrJD0l6UupfJCk30panb4PzLnmYklrJP1F0kdyysdIeiIdmy6pvYWdrcgk7SXpbklPp7Yf5zbveSTdLOlFSU/mlOWtnSX1kXRHf7hJGAAAELJJREFUKl8kaXgh78/MzMze4gSrdNQDX4mIkcCxwBckjQK+ATwYEYcCD6Z90rEJwBHAycCNkipTXT8CPg8cmr5OLuSN2E65HvhNRBwOvBtYhdu8J/oprdskn+38WeCViBgBXAt8d7fdiZmZme2QE6wSERH/jIilaXsz2R/aBwDjgVvSabcAH0vb44HbI2JbRPwNWAOMlbQf0D8i/hjZDCY/y7nGSoik/sD7gZkAEfFmRLyK27zHiYhHgI0tivPZzrl13Q2c4F7MnkfSRyT9SNJ9kuambX+Y0oO5zc26J0/TXoLS8J7RwCJgn4j4J2RJmKQh6bQDgD/lXLY+ldWl7ZblVnoOAV4CZkl6N7AE+BJu83KRz3Y+AHg21VUvaROwN7Bh94VvhSTpOuDtZIl108/BUGCypFMi4ktFC852C7e5SRoERES8UuxYbOc4wSoxkmqAe4AvR8RrO/gQuq0DsYNyKz1VwFHA/42IRZKuJw0Ta4fbvDx0pZ39M9DznRoRb29ZKOkO4L/JPpyxnsVtXoYkDQO+B5wAvJoVqT/wEPCNiHimiOFZJ3mIYAmR1IssubotIn6Ril9IQ4NI319M5euBA3MuHwo8l8qHtlFupWc9sD4iFqX9u8kSLrd5echnO2+/RlIVMIDWQxKte9sqaWwb5f8L2FroYKwg3Obl6Q7gXmDfiDg0PVu7H/BL4PaiRmad5gSrRKTnJWYCqyLiBzmH7gMmpu2JwNyc8glp9rCDyR54X5yGHG2WdGyq8/yca6yERMTzwLOSDktFJwArcZuXi3y2c25dZwEPhVeR72kuAP5D0kpJC9LXKuA/0jHreS7AbV6O3hYRd0REQ1NBRDRExO1kQ7+tG5B/B5cGSccBfwCeABpT8b+TPYd1JzAM+DtwdkRsTNdcAnyGbAbCL0fEr1P50WSzlvUFfk02BM0NXYIkHQnMAHoDa4FJZB98uM17EEk/B44H3ga8AHyL7NPIvLSzpGrgVrJnNzcCEyJibaHuzwpH0r5kz9yJrAf8+SKHZLuZ27y8SLqd7P/xW0jP1pKNUJhIlnz9a7Fis85zgmVmZtYNpJ7LsWR/bAfZENHF/jCl53Kblx9JvcmW3hhPTmJNNlJhZkRsK2J41klOsMzMzEqcpA8DNwKrgX+k4qHACOD/RMSCYsVmu4fb3Kz7coJlZmZW4tKzN6e0nEEsPac3Py1Sbz2I27x8SfoI2TqHuT2XcyPiN0UNzDrN07SbmZmVviqar4PW5B9ArwLHYoXhNi9DXv+sZ3CCZWZmVvpuBh5LD8DnPvg+gWwGWut53Oblyeuf9QAeImhmZtYNSBoFnEGLB98jYmVRA7Pdxm1efiStAC6MiMUtyseSTXLxzuJEZjvDCZZ1S5IC+EFEfCXtTwFqImJqF+oaDrwnIubkM8aukjSD7N7a/QUq6QJgQUR4QWEzM7MeQtJRwI+APXlriOCBwGtkk5ssKVZs1nkeImjd1TbgTElXR8SGXaxrOPApoFWCJakqIup3sf5Ok1QZERd24tQLgCfJHnzdrdI0wYqIxg5P3vXXqsxdXNHMMpIGABeTPfg+OBW/SLbY9LSIeLVYsdnu4TYvTxGxFDjG6591bxXFDsCsi+qBm4CLWh6QdJCkByWtSN+HdVDXNOB9kpZLukjSBZLukvQrYIGkmlTPUklPSBqfXme4pFWSfiLpKUkLJPVNxyZLWpliuD2V1UialepYIekTqbxW0hWSFgHjJP0uLSjbdOya9NoPShos6SzgaOC2FHNfSdNyXu/7bbwnUyXdKukhSaslfS7n2FclPZauvbzFvd0ILCX79Cy3vpMlPS1poaTpku7PeZ0pOec9mXoIkXSepMUp5h9Lqmzj/i+VdG/O9SdJ+kUH7WdWDu4EXgGOj4i9I2Jv4IPAq8BdRY3Mdhe3eZlKH2welL4OBA5KZdZNOMGy7uwG4Nz0KV+uHwI/i4h3AbcB0zuo5xvAHyLiyIi4NpWNAyZGxIeArcDHI+Iosl9u1+T8R3cocENEHEH2S+8TOXWOTjH8Wyq7DNgUEe9M5Q+l8n7AkxFxTEQsbBFbP2Bpeu3fA9+KiLuBx4FzI+JIoC/wceCIVO+327nPdwGnpXv7pqT90zorh5ItZHkkMEbS+9P5h6X3cXRErGuqRFI18BPgo8D7gH3bfWffumYk8EngvSnmBuDclvcPXAGMlNT0ae0kYFZH9ZuVgeER8d3cT7Ej4vmImAZ09CGSdU9u8zKUfi+vBqYCp5L93r4cWJ2OWTfgIYLWbUXEa5J+BkwG3sg5NA44M23fCnyvC9X/NiI2pm0BV6XEo5Gsy36fdOxvEbE8bS8hG24IsIKsh+mXwC9T2Ylksz81xf9K2mwA7mknjkbgjrQ9G2irN+c1siRwhqR5wP3t1DU3It4A3pD0MFlSdRzwYWBZOqeGLOH6O7AuIv7URj2Hk933agBJs4HPt/OaTU4AxpDNiAVZUvhiOrb9/iMiJN0KnCdpFllbnt9B3WblYJ2krwG3RMQLAJL2IRsu/OyOLrRuy21enq4HTmxv/TPA6591A+7Bsu7uOuCzZL0g7enKTC5bcrbPJRv/Pib1vrwAVKdj23LOa+CtDy1OI+thGwMskVRFlqi1FcvWnXjuqNX16RmxsWRJyseA9hYibHltpJiuTr13R0bEiIhomv53C+1r7z2tp/n/K03vk8j+SGh6ncNyJiRpef+zgPOAc4C7CvkMnFkJ+ySwN/B7Sa9I2gj8DhgE/GsxA7Pdxm1enrz+WQ/gBMu6tdTLdCdZktXkUd7qKToXaDnsrqXNZLP1tGcA8GJE1En6INmY6HZJqgAOjIiHga8Be5H1DC0Avphz3sAO4oLs3+hZaftTvHUv22OWVAMMiIj5wJfJhvq1Zbykakl7A8cDjwEPAJ9JdSDpAElDOojpaeBgSf+S9s/JOfYMcFSq6yjg4FT+IHBWU92SBklq831MMyM+B1wK/LSDWMzKQurxnkX2f8iBETEoIkZGxNfJPmCxnuftwFURcTjZyIkfAn9NxzwZUM/VtP7Z1yV9Kn19HViE1z/rNpxgWU9wDfC2nP3JwCRla0l8mrQon6QzJF3RxvUrgHpJf5bUatIMsue4jpb0OFnC9nQH8VQCsyU9QTb07to029O3gYFp4oc/kz3P1ZEtwBGSlgAfIntGCbLE4z8lLSdLtO5P9/t72pj4I1kMzAP+BFwZEc9FxAKy2RP/mOK9mx0nm0TEVrIhgfMkLQTW5Ry+BxiU4vrfZIsikqacv5Rs0pAVwG+B/XbwMrcBz3qtF7OMpMlks8d9EXhSabKd5KriRGW72c28NZLgOrL/m6cBr+NnU3usiLia7G8NkQ2Tf0/aPjcds27A62CZlTBJtRFRk4d6pgK1EdFqhsE81H08MCUiTs9jnT8EluUMVzQra+kDkHERUZtm5rwbuDUirpe0LCJGFzVAyztJqyJiZNpumuyo6djyNGTdzEqQe7DMrKSk3rp3kU3qYWaZyoioBUgPvx8PnCLpB2SfblvP86SkSWn7z3pr+Y63A3XFC8t2J0kDlC298rSkl9PXqlS2V7Hjs85xD5aZmVmJk/QQ8P9yZi0lTZ5zM9nQocqiBWe7RVqC5Hqy5TA2kD3f+mz6mhwRfy5ieLabSHqAbBmXW5qm6Fe26PAFwAkRcVIRw7NOcoJlZmZW4iQNBepz10TKOfbeiPj/RQjLCkDSnsAhpNnlmqZst55J0l8i4rCdPWalxQmWmZmZmVkJkLQA+C/aXv/spIg4sYjhWSf5GSwzMzMzs9KQu/7Zxhbrn51dzMCs89yDZWZmZmZW4iRNighP0d8NOMEyMzMzMytxkv4eEcOKHYd1rKrYAZiZmZmZGUha0d4hYJ9CxmJd5wTLzMzMzKw07AN8BHilRbmARwsfjnWFEywzMzMzs9JwP1CTu+ZdE0m/K3w41hV+BsvMzMzMzCxPPE27mZmZmZlZnjjBMjMzMzMzyxMnWGZmZtbtSApJ1+TsT5E0tYt1DZf0qbwFt4skzZA0qoNzLpC0f6FiMrPOc4JlZmZm3dE24ExJb8tDXcOBNhMsSQWdEExSZURcGBErOzj1AqAgCZYyBfmbUVJlIV7HbHdygmVmZmbdUT1wE3BRywOSDpL0oKQV6XtHi7NOA94nabmki1Lv0F2SfgUskFST6lkq6QlJ49PrDJe0StJPJD0laYGkvunYZEkrUwy3p7IaSbNSHSskfSKV10q6QtIiYJyk30k6OufYNem1H5Q0WNJZwNHAbSnmvpKm5bze99t4T6ZKulXSQ5JWS/pczrGvSnosXXt5i3u7EVgKHNiivpMlPS1poaTpku7PeZ0pOec9KWl42j5P0uIU84+bkqkW93+ppHtzrj9J0i86aD+zkuIEy8zMzLqrG4BzJQ1oUf5D4GcR8S7gNmB6B/V8A/hDRBwZEdemsnHAxIj4ELAV+HhEHAV8ELhGktJ5hwI3RMQRwKvAJ3LqHJ1i+LdUdhmwKSLemcofSuX9gCcj4piIWNgitn7A0vTavwe+FRF3A48D50bEkUBf4OPAEaneb7dzn+8CTkv39k1J+0v6cLqHscCRwBhJ70/nH5bex9ERsa6pEknVwE+AjwLvA/Zt951965qRwCeB96aYG4BzW94/cAUwUtLgdGwSMKuj+s1KiRMsMzMz65Yi4jXgZ8DkFofGAXPS9q3AcV2o/rcRsTFtC7hK0grgv4ADyBaEBfhbzppFS8iGGwKsIOthOo+stw3gRLKksCn+psVkG4B72omjEbgjbc9u515eI0sCZ0g6E3i9nbrmRsQbEbEBeJgsqfpw+lpG1lN1OFnCBbAuIv7URj2Hk9336sjW+5ndzuvlOgEYAzwmaXnaPyQd237/qb5bgfMk7UXWlr/uRP1mJcMLDZuZmVl3dh1ZYrCjXo6uLPq5JWf7XGAwMCYi6iQ9A1SnY9tyzmsg602CrKfo/cAZwGWSjiBL1NqKZWtENHQyrlbXR0S9pLFkScsE4IvAhzpxbaSYro6IH+ceSMP6ttC+9t7Tepp/gN/0Pgm4JSIubuOalvc/C/gVWdJ4V0TUt3GNWclyD5aZmZl1W6mX6U7gsznFj5IlGpAlRy2H3bW0GdhzB8cHAC+m5OqDwEE7qixNCHFgRDwMfA3YC6gBFpAlP03nDewgLsj+VjsrbX+Kt+5le8ySaoABETEf+DLZUL+2jJdULWlv4HjgMeAB4DOpDiQdIGlIBzE9DRws6V/S/jk5x54Bjkp1HQUcnMofBM5qqlvSIEltvo8R8RzwHHAp8NMOYjErOe7BMjMzs+7uGnISF7IhgzdL+irwEtlzPEg6Azg6Ir7Z4voVQL2kP5P9Qf9Ki+O3Ab+S9DiwnCzB2JFKYHZ6NkzAtRHxqqRvAzdIepKst+tyoKMJHLYAR0haAmwie46JFOd/SnoDOAWYm56NEm1M/JEsBuYBw4ArmxKZ9HzUH9NjZbXAeSm+NkXEVkmfB+ZJ2kCW9L0jHb4HOD8NA3wM+O90zUpJl5JNGlIB1AFfANa1eoHMbcDgTsymaFZylA11NTMzM7NSI6k2ImryUM9UoDYiWs0wmIe6jwemRMTpeazzh8CyiJiZrzrNCsU9WGZmZmZWMlJv3RbgK8WOxawr3INlZmZmZmaWJ57kwszMzMzMLE+cYJmZmZmZmeWJEywzMzMzM7M8cYJlZmZmZmaWJ06wzMzMzMzM8uR/AEPBrEBy3N2fAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = []\n",
    "for p in jobs['total']['Query'].unique():\n",
    "    data.append([jobs['total'][jobs['total']['Query'] == p]['Time'].sum()/60 * EC2_PRICE['m5dn'][64],\n",
    "                jobs['total'][jobs['total']['Query'] == p]['Time'].sum()/60 * SPOT_PRICE['m5dn'][64],\n",
    "                 p])\n",
    "cost = pandas.DataFrame(data, columns=['Normal', 'SPOT', 'Query'])\n",
    "cost.to_csv(os.path.join(result_dir,'aws_cost_120k.tsv'), sep='\\t', index=None)  \n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12,8), ncols=2)\n",
    "\n",
    "box_plot = sns.boxplot(x=\"Query\", y=\"Time\", data=jobs['total'], ax=ax[0])\n",
    "sns.stripplot(x=\"Query\", y=\"Time\", data=jobs['total'], jitter=True, color=\".3\", ax=ax[0])\n",
    "\n",
    "ax[0].set_ylabel(\"Total Time (Minutes)\")\n",
    "ax[0].set_xlabel(\"No. transcripts per query\")\n",
    "\n",
    "cost.plot(x=\"Query\", kind=\"bar\", ax=ax[1])\n",
    "ax[1].yaxis.tick_right()\n",
    "ax[1].yaxis.set_label_position(\"right\")\n",
    "ax[1].set_ylabel(\"Cost (US Dollars)\")\n",
    "ax[1].set_xlabel(\"No. transcripts per query\")\n",
    "\n",
    "handles, labels = ax[1].get_legend_handles_labels()\n",
    "ax[1].legend(handles, labels)\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Deleting all resources created"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Deregistering job definition m5d_4xlarge\n",
      "Disabling queue m5d_4xlarge_queue\n",
      "Deleting queue definition m5d_4xlarge_queue\n",
      "Disabling compute environment m5d_4xlarge_managed\n",
      "Deleting compute environment m5d_4xlarge_managed\n",
      "Deregistering job definition m5d_8xlarge\n",
      "Disabling queue m5d_8xlarge_queue\n",
      "Deleting queue definition m5d_8xlarge_queue\n",
      "Disabling compute environment m5d_8xlarge_managed\n",
      "Deleting compute environment m5d_8xlarge_managed\n",
      "Deregistering job definition m5d_16xlarge\n",
      "Disabling queue m5d_16xlarge_queue\n",
      "Deleting queue definition m5d_16xlarge_queue\n",
      "Disabling compute environment m5d_16xlarge_managed\n",
      "Deleting compute environment m5d_16xlarge_managed\n",
      "Deregistering job definition m5dn_4xlarge\n",
      "Disabling queue m5dn_4xlarge_queue\n",
      "Deleting queue definition m5dn_4xlarge_queue\n",
      "Disabling compute environment m5dn_4xlarge_managed\n",
      "Deleting compute environment m5dn_4xlarge_managed\n",
      "Deregistering job definition m5dn_8xlarge\n",
      "Disabling queue m5dn_8xlarge_queue\n",
      "Deleting queue definition m5dn_8xlarge_queue\n",
      "Disabling compute environment m5dn_8xlarge_managed\n",
      "Deleting compute environment m5dn_8xlarge_managed\n",
      "Deregistering job definition m5dn_16xlarge\n",
      "Disabling queue m5dn_16xlarge_queue\n",
      "Deleting queue definition m5dn_16xlarge_queue\n",
      "Disabling compute environment m5dn_16xlarge_managed\n",
      "Deleting compute environment m5dn_16xlarge_managed\n"
     ]
    }
   ],
   "source": [
    "for m in MACHINE_TYPES:\n",
    "    for c in CPUs:\n",
    "        job_def_name = '{}_{}'.format(m,c['name'])\n",
    "        print('Deregistering job definition {}'.format(job_def_name))    \n",
    "        !aws batch deregister-job-definition --job-definition {job_def_name}\n",
    "\n",
    "        queue_name = '{}_{}_queue'.format(m,c['name'])\n",
    "        print('Disabling queue {}'.format(queue_name)) \n",
    "        output = !aws batch update-job-queue --job-queue {queue_name} --state DISABLED\n",
    "        if not ''.join(output).endswith('queue does not exist'):\n",
    "            while True:\n",
    "                try:\n",
    "                    time.sleep(15)\n",
    "                    output = !aws batch describe-job-queues --job-queues {queue_name}\n",
    "                    queue = json.loads(''.join(output))\n",
    "                    if 'jobQueues' in queue and len(queue['jobQueues']) == 1 and queue['jobQueues'][0]['state'] == 'DISABLED' and queue['jobQueues'][0]['status'] == 'VALID':\n",
    "                        queue = queue['jobQueues'][0]\n",
    "                        if queue['state'] == 'DISABLED':\n",
    "                            print('Deleting queue definition {}'.format(queue_name)) \n",
    "                            !aws batch delete-job-queue --job-queue {queue_name}\n",
    "                            break\n",
    "                except:\n",
    "                    print('Error: {}'.format(output))\n",
    "                    break\n",
    "\n",
    "        comp_env_name = '{}_{}_managed'.format(m,c['name'])\n",
    "        print('Disabling compute environment {}'.format(comp_env_name)) \n",
    "        output = !aws batch update-compute-environment --compute-environment {comp_env_name} --state DISABLED\n",
    "        if not ''.join(output).endswith('does not exist'):\n",
    "            while True:\n",
    "                try:\n",
    "                    time.sleep(15)\n",
    "                    output = !aws batch describe-compute-environments --compute-environments {comp_env_name}\n",
    "                    comp_env = json.loads(''.join(output))                    \n",
    "                    if 'computeEnvironments' in comp_env and len(comp_env['computeEnvironments']) == 1:\n",
    "                        comp_env = comp_env['computeEnvironments'][0]\n",
    "                        output = !aws batch describe-job-queues --job-queues {queue_name}\n",
    "                        queue = json.loads(''.join(output))\n",
    "                        if 'jobQueues' in queue and len(queue['jobQueues']) == 0 and comp_env['state'] == 'DISABLED' and comp_env['status'] == 'VALID':\n",
    "                            print('Deleting compute environment {}'.format(comp_env_name))\n",
    "                            !aws batch delete-compute-environment --compute-environment {comp_env_name}\n",
    "                            break\n",
    "                except:\n",
    "                    print('Error: {}'.format(output))\n",
    "                    break\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}