A PHP reporting engine that works especially well with Laravel, built with love at CareSet Systems
The basic idea in Zermelo is to let report authors think exclusively in SQL SELECT statements, and to allow Zermelo to handle the translation of the data that results from queries into complex and rich web interfaces. There are plenty of good tools available to help build SELECT statements, and there are thousands of excellent resources available to learn how to use the SELECT query functions in SQL. And if you know how to use SQL SELECT statements, then with Zermelo, you can automatically create complex and interactive web-based reports.
Generally, this happens using the ability of SQL to have aliases for the output of specific variables. For most of the reporting engines, you can have one or many SQL queries that output into specific aliased columns that Zermelo understands. And then the reporting engine will automatically populate a web-based data view with the data output. For instance, the card-based layout engine allows you to populate rows of data into BootStrap Cards. Almost every portion of the Bootstrap Card can be populated by using column names that correspond to the css classes supported inside the bootstrap card component.
The exception to this approach is the tabular data viewer. Here, you can output anything you want from your SELECT statement and Zermelo will do its best to create a online auto-paging tabular view of your data using the DataTables javascript project.
- Write SQL, automatically get a web-based report
- Decorate each row in the report with links, buttons, JS and other bootstrap-based HTML goodness.
- Control the entire web report using a single PHP file, which contains SQL and web interface decorations.
- GET/POST/JSON/URL parameters are all available as local functions/variables in the single file. This allows the SQL to be heavily modified based on the paramaters passed in to the specific report url.
- Automatically support server side data paging, allows engine to report against very large databases
- Any report data can be downloaded in CSV file(s)
- Report automatically generates JSON data source that can be used as an API
- Supports Laravel's Blade templating engine out of the box (with more effort supports any front end templating engine).
We have a feature roadmap if you want to see where we are going and an extended list of Zermello Reporting features.
Read the Architecture diagram Zermelo is doing the hard work of translating "mere SQL" into something that can consistently and with minimal impact on performance be loaded into a single browser session. Zermelo understands how to push the hard work to the MariaDB/MySQL server, ensuring that the browser gets its data in dribbles. The backend is having to do a huge amount of work in order make that happen.
Some queries will legitimately take hours for the backend to run, even when the resulting data is only a few hundred rows of results. In order to support these heavy loads, Zermelo understands how to cache results. It always caches the results, but for most queries, it always refreshes the cache on every browser call.
You, the user, get to control how this works. Look at the Controlling the Cache documentation to see how.
You will need a modern LAMP server with at least php 7.2 and at least Laravel 5.5 Complete Prerequisites Once the prerequisites are completed you should be able to check URLs on host system's browser at URL: homestead.test
Look in Basic Installation for complete installation instructions and database setup.
For a quick start, assuming your Laravel instance already has access to the DB that it needs
$ composer require careset/zermelo
$ php artisan zermelo:install
This will install and configure zermelo, and create an app/Reports for you to add reports too.
Next, you should test your routes...
$ php artisan route:list | grep Zermelo
| | GET|HEAD | Zermelo/{report_key} |
| | GET|HEAD | ZermeloCard/{report_key} |
| | GET|HEAD | ZermeloGraph/{report_key} |
| | GET|HEAD | api/Zermelo/{report_key}/Download/{parameters?} |
| | GET|HEAD | api/Zermelo/{report_key}/Summary/{parameters?} |
| | GET|HEAD | api/Zermelo/{report_key}/{parameters?} |
| | GET|HEAD | api/ZermeloGraph/{report_key}/Download/{parameters?} |
| | GET|HEAD | api/ZermeloGraph/{report_key}/{parameters?} |
We provide example reports, and the schema and data needed to run those reports. This is a good place to start if you are just exploring the system. Read, Running the Examples
- Edit the file
config/zermelo.php
to change core zermelo setting these values are explained there and in Configuration Documentations - Edit the file
config/zermelobladetabular.php
to change settings specific to zermelo blade tabular view package. - Earlier in the Basic Installation you've already created an app/Reports directory. If desired, you can create a differently named report directory, but you must also change the namespace. Change the REPORT_NAMESPACE setting in config/zermelo.php to something else...
/**
* Namespace of the report where it will attempt to load from
*/
'REPORT_NAMESPACE' =>env("REPORT_NAMESPACE","app\Reports"),
... like "Zermelo" and then create a ~/code/zermelo-demo/app/Zermelo directory to place your example report in. Note: you will also need to change the namespace of Northwind*Reports.php files to "namespace app\Zermelo;" if you change the REPORT_NAMESPACE. 4. To configure middleware, you may add, or edit the MIDDLEWARE config setting in your config/zermelo.php file. This will run the configured middleware on each API request. For example, if you have enabled Laravel's Authentication and wish to protect the Zermelo routes using the auth middleware, you may add the string "auth" to the MIDDLEWARE array in order to exeute the auth middleware on each API request to the Zermelo API. Similarly, for the front-end view packages like zermelobladetabular, you may add the "auth" string to the TABULAR_MIDDLEWARE array in zermelobladetabular.php to enable authentication on that route.
In project home dir:
$ composer update careset/zermelo
$ php artisan zermelo:install
When you install the zermelobladetabular package, Just Say No to 'replace' all those files EXCEPT: 'The [zermelo/tabular.blade.php] view already exists' Y (replace it!) 'The [zermelo/layouts/tabular.blade.php] view already exists.' Y (replace it!)
You can uninstall the composer packages by running 'composer remove' to remove the requirements in composer.json, and to remove the packages from the vendor directory. In project home dir:
$ composer remove careset/zermelo
$ composer clear-cache
-
In order to get your first report, you need to create a report file. The easiest way to create an new report file is to run:
php artisan zermelo:make_tabular [YourNewReportName]
To understand what this does, take a look at the example report model below.
- Edit the new file
/app/Zermelo/[YourNewReportName]
(or, with the defaults mentioned in the instructions,/app/Reports/[YourNewReportName]
) You must fill in a reasonable GetSQL() function that returns either a single SQL text string, or an array of SQL text strings. - Point your browser to https://yourapp.example.com/Zermelo/YourNewReportName
- Enjoy seeing your data in an automatically pagable Datatables display!!
- Various functions and constants in the report file can dramatically change how the report is displayed on the front end. Use them to change the reports (a good first hack is to use the MapRow function to link one report to another report)
GetSQL() This is the core of the report. Implement this function in your report child class, adding your SQL query to populate the report. The column names in your SELECT statement become the headers of your report by default.
GetReportName() Implement this function to return the title of the report.
GetReportDescription() Implement this function, and the returned string will be printed in the description block below title of the report. The string is not escaped, and is passed raw to the report, so it is possible to print HTML (form elements, for example)
GetReportFooter() Implement this function to return a string that will be displayed within the footer tag at the bottom of the report layout. The string is not escaped, and is passed raw to the report, so it is possible to print HTML (form elements, for example)
GetReportFooterClass() Implement this class to add specific class to your footer. Add "fixed" to make your footer fixed to the bottom, and add "centered" to center your footer content. For example, implement this function to return "fixed centered" for your footer to be fixed, and it's content centered.
MapRow(array $row, int $row_number) Implement this method to modify tabular cell content. When displaying to the tabular view, your report child class can chose to modify the content of each row cell.
OverrideHeader(array &$format, array &$tags) Implement this method to verride a default column format or add additional column tag to be sent back to the front end
isCacheEnabled() Turn caching on and off. Default is to return false which means cache is off. If you set this function to return true, this will enable the caching. Creating the cache table always happens, but when the cache is enabled, the cache is used to answer subsequent queries rather than re-running the original query. This can cause confusing results (i.e. changing the underlying data does not change the content of the report when the cache is used) and as a result, it is off by default. But for many large and slow queries, caching is nessecary.
howLongToCacheInSeconds() If cache is enabled, we use this setting to configure how long we wish to retain the cached report, before re-running the original query.
getCacheDatabaseSource() Specify a location where we want our report to point to, overriding the automatically-generated cache table name. With this property, the report developer can specify the source for the cache database. You may want to use this if you're data is ending up in a specific, known location, and you want to point your report at it. You can either implement this function in your report, or set the property $CACHE_DATABASE_SOURCE using this structure:
protected $CACHE_DATABASE_SOURCE = [
'database' => '_zermelo_cache_overload',
'table' => 'northwind_cust_overload'
];
GetReportJS() Implement this function to return a string that will be placed in a script tag before the closing body tag of the report. Do not include script tags. This function should return JS code only. The string is not escaped, and is passed raw to the report.
isSQLPrintEnabled() Will turn on the ability to visit the report using the /ZermeloSQL/ url and instead of running the report, a helpful debugging screen will be presented, showing what SQL would have been returned by GetSQL() in a pretty-printed manner. This is very helpful for debugging inputs. It does require that SQL view be enabled in the /config/zermelo.php file to work, however.
getInput($key = null) Use this function to get the value of of a GET parameter passed to the report. You can use this in your GetSQL() function to affect your query based on additional parameters passed in the request query string.
setInput($key, $new_value) A useful but dangerous function that allows for specific reports to override the input that comes from a user before it is used. Use this carefuly, since this will make changing the setting in the user interface not function properly. TODO (have the UX note that a setting is frozen)
setDefaultInput($key, $new_value) This will set a input variable to starting value.. until the value is reset in the UX. (unlike setInput it will not override user values)
setDefaultSortOrder($sort_array) This is a helper function for setInput() that allows to set a default order in the UI on tabular (and tabular derived) views. The sort_array argument takes the form: $sort_order = [ ['order_count' => 'desc'], ['name' => 'asc'] ];
This would result in listing the rows with the most orders at the top and when several rows had the same number of orders would be listed alphabetically
pushViewVariable($key, $value) Use this function to pass a variable to the view without going through request/response cycle. The key parameter is a string, and will be available on the view template as a php variable. For example, if you have the following in your GetSQL() function:
$this->pushViewVariable('extra_var',true)
then your variable will be available in the view as $extra_var with a value of 'true'.
@if ($extra_var === true)
<p>ExtraVar Is True!</p>
@endif
To see full list of functions and variables, please see the ZermeloReport model - https://github.com/CareSet/Zermelo/blob/master/src/CareSet/Zermelo/Models/ZermeloReport.php
<?php
namespace App\Reports;
use CareSet\Zermelo\Reports\Tabular\AbstractTabularReport;
class ExampleReport extends AbstractTabularReport
{
/*
* Get the Report Name
*/
public function GetReportName(): string {
return("Enter your report name here");
}
/*
* Get the Report Description, bootstrap styled html is OK
*/
public function GetReportDescription(): ?string {
$desc = "This is your report description <b> HTML is fine here </b>";
return($desc);
}
/**
* This is what builds the report. It will accept a SQL statement or an Array of sql statements.
* Can be used in conjunction with Inputs to determine different output based on URI parameters
* Additional URI parameters are passed as
* $this->getCode() - which will give the first url segment after the report name
* $this->getParameters() - which will give an array of every later url segment after the getCode value
* $this->getInput() - which will give _GET and _POST parameters. Should also get values inside JSON that is posted... a unified view of user input
* $this->quote($something_you_got_from_the_user) - This wrapper to the PDO quote function is good for preventing SQL injection
* $this->setInput($key,$new_value) - a way to override _GET parameters (i.e. for initializing a sort for instance)
* For instance $this->setInput('order',[0 => ['order_by_me' => 'asc']]); will order the report, to start by the order_by_me column ASC.
* This replicates what is being passed from the front end data tables to the backend to make sorting work..
**/
public function GetSQL()
{
//replace with your own SQL
$sql = "SELECT * FROM information_schema.TABLES";
return $sql;
}
/**
* Each row content will be passed to MapRow.
* Values and header names can be changed.
* Columns cannot be added or removed
* You can decorate fields with html, with bootstrap css styling
*
*/
public function MapRow(array $row, int $row_number) :array
{
/*
//this logic would ensure that every cell in the TABLE_NAME column, was converted to a link to
//a table drilldown report
$table_name = $row['TABLE_NAME'];
$row['TABLE_NAME'] = "Gotta Love Those Row Decorations: $table_name";
//this will make table name a link to another report
//$row['TABLE_NAME'] = "<a href='/Zermelo/TableDrillDownReport/$table_name/'>$table_name</a>";
//this will do the same thing, but styling the link as a bootstrap button.
//$row['TABLE_NAME'] = "<a class='btn btn-primary btn-sm' href='/Zermelo/TableDrillDownReport/$table_name/'>$table_name</a>";
*/
return $row;
}
/**
* If a column needs to be forced to a certain format (i.e.ear auto-detection gets it wrong), it can be changed here
* Tags can also be applied to each header column
*/
public function OverrideHeader(array &$format, array &$tags): void
{
//$tags['field_to_bold_in_report_display'] = ['BOLD'];
//$tags['field_to_hide_by_default'] = ['HIDDEN'];
//$tags['field_to_italic_in_report_display'] = ['ITALIC'];
//$tags['field_to_right_align_in_report'] = ['RIGHT'];
//How to set the format of the display
//$format['numeric_field'] = 'NUMBER'; // Formats number in table using commas, and right-aligns
//$format['decimal_field'] = 'DECIMAL'; // Formats decimal to 4 places, and right-aligns
//$format['currency_field'] = 'CURRENCY'; // adds $ or Eurosign and right align
//$format['percent_field'] = 'PERCENT'; // adds % in the right place and right align
//$format['url_field'] = 'URL'; // auto-link using <a href='$url_field'>$url_field</a>
//$format['date_field'] = 'DATE'; // future date display
//$format['datetime_field'] = 'DATETIME'; //future date time display
//$format['time_field'] = 'TIME'; // future time display
//$format['row_summary_field'] = ['DETAIL']; // this field will be shown with a + sign in the column.
//pressing the plus will create a new row in the table
//that shows the actual contents of this column.
}
/**
* Header Format 'auto-detection' can be changed per report.
* it is based on seeing the strings below in a field name... it will then assume it should be styled accordingly
* So it a column label is 'very_good_num' or 'this num' will be matched by 'num' but 'number' will not work.
* so it is matched on ignore case on a column name segment, not on substring...
* By default, these are the column formats -
* public $DETAIL = ['Sentence'];
* public $URL = ['URL'];
* public $CURRENCY = ['Amt','Amount','Paid','Cost'];
* public $NUMBER = ['id','#','Num','Sum','Total','Cnt','Count'];
* public $DECIMAL = ['Avg','Average'];
* public $PERCENT = ['Percent','Ratio','Perentage'];
*
* It detects the column by using 'word' matching, separated white spaces or _.
* Example: TABLE_ROWS - ['TABLE','ROWS']
* It will also check the full column name
*/
public $NUMBER = ['ROWS','AVG','LENGTH','DATA_FREE'];
/*
* By Default, any numeric field will have statistical information will be passed on. AVG/STD/MIN/MAX/SUM
* Any Text column will have distinct count information passed on.
* Any Date will have MIN/MAX/AVG
* This field will add a "NO_SUMMARY" field to the column header to suggest the data not be displayed
*/
public $SUGGEST_NO_SUMMARY = ['ID'];
/**
* Want to use your own blade file for the report front-end?
* You can customize the report view based on the report
* When this is set to null, the report will use the view defined in the configuration file.
*
*/
public $REPORT_VIEW = null;
/*
* Get Indexes for cache table
* Because results are saved to a cache table, and then exported from there later searching, using the front end... can be very slow
* This returns an array of Index commands that will be run against the cache table
* because we do not know the name of the cache table in advance, these index commands must use the string '{{_CACHE_TABLE_}}' instead of
* the name of a specific table...
*/
public function GetIndexSQL(): ?array {
$index_sql = [
"ALTER TABLE {{_CACHE_TABLE_}} ADD INDEX(`COLUMN_NAME`);",
"ALTER TABLE {{_CACHE_TABLE_}} ADD INDEX(`TABLE_NAME`);",
"ALTER TABLE {{_CACHE_TABLE_}} ADD INDEX(`database_name`);",
"ALTER TABLE {{_CACHE_TABLE_}} ADD PRIMARY KEY( `database_name`, `TABLE_NAME`, `COLUMN_NAME`); "
];
// you can uncomment this line to enable the default report to automatically index the resulting cached table
// return($index_sql);
//returning null here results in no indexing happening on the cached results table
return(null);
}
/**
* @return null|string
*
* Return the footer of the report. This string will be placed in the footer element
* at the bottom of the page.
*/
public function GetReportFooter(): ?string
{
$footer = <<<FOOTER
<p>Made with love by CareSet</>
FOOTER;
return $footer;
}
/**
* @return null|string
*
* Add a string here to put in the class of the footer element of the report layout
*/
public function GetReportFooterClass(): ?string
{
// Add "fixed centered" to have your footer fixed to the bottom, and/or centered
// This will be put in the class= attribute of the footer
return "";
}
/**
* @return null|string
*
* This will place the enclosed Javascript in a <script> tag just before
* the body of your view. Note, there is no need to include a script tag
* in this string. The content of this string is not HTML encoded, and is passed
* raw to the view.
*/
public function GetReportJS(): ?string
{
$javascript = <<<JS
alert("place javascript code here");
JS;
return $javascript;
}
/**
* If the cache is not enabled, then every time the page is reloaded the entire report is re-processed and put into the cache table
* So if you want to just run the report one time, and then load subsequent data from the cache, set this to return 'true';
*/
public function isCacheEnabled(){
return(false);
}
/**
* This function does nothing if isCacheEnabled is returning false
* But if the cache is enabled, then this will detail how long the report will be reloaded from the cache before the cache is regenerated by re-running the report SQL
*/
public function howLongToCacheInSeconds(){
return(1200); //twenty minutes by default
}
public function isSQLPrintEnabled(): bool{
return(false);
}
}
One of the most advanced features is the use of the "socket/wrench" notion inside reports.
Essentially, each "wrench" that you ask for in a report using:
$this->getSocket('someWrenchName');
You can call this function anywhere you want in your report file, but usually it is called from the GetSQL() function, so that you can use the results to build your sql.
Will ask the user to choose between options that are associated with the 'someWrenchSocket' To use it you have to setup as 'socketwrench' database and populate it with the following tables...
--
-- Table structure for table `socketsource`
--
CREATE TABLE `socketsource` (
`id` int(11) NOT NULL,
`socketsource_name` varchar(255) NOT NULL,
`created_at` datetime NOT NULL,
`updated_at` datetime NOT NULL
) ENGINE=MyISAM DEFAULT CHARSET=utf8mb4;
-- --------------------------------------------------------
--
-- Table structure for table `socket_user`
--
CREATE TABLE `socket_user` (
`id` int(11) NOT NULL,
`user_id` int(11) NOT NULL,
`wrench_id` int(11) NOT NULL,
`current_chosen_socket_id` int(11) NOT NULL,
`created_at` datetime NOT NULL,
`updated_at` datetime NOT NULL
) ENGINE=MyISAM DEFAULT CHARSET=utf8mb4;
-- --------------------------------------------------------
--
-- Table structure for table `socket`
--
CREATE TABLE `socket` (
`id` int(10) UNSIGNED NOT NULL,
`wrench_id` int(11) NOT NULL,
`socket_value` varchar(1024) COLLATE utf8mb4_unicode_ci NOT NULL,
`socket_label` varchar(1024) COLLATE utf8mb4_unicode_ci NOT NULL,
`is_default_socket` tinyint(1) NOT NULL,
`socketsource_id` int(11) NOT NULL,
`created_at` timestamp NULL DEFAULT NULL,
`updated_at` timestamp NULL DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
-- --------------------------------------------------------
--
-- Table structure for table `autosocket`
--
CREATE TABLE `autosocket` (
`id` int(10) UNSIGNED NOT NULL,
`wrench_id` int(11) NOT NULL,
`socket_value` varchar(1024) COLLATE utf8mb4_unicode_ci NOT NULL,
`socket_label` varchar(1024) COLLATE utf8mb4_unicode_ci NOT NULL,
`is_default_socket` tinyint(1) NOT NULL,
`socketsource_id` int(11) NOT NULL,
`created_at` timestamp NULL DEFAULT NULL,
`updated_at` timestamp NULL DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
-- --------------------------------------------------------
--
-- Table structure for table `wrench`
--
CREATE TABLE `wrench` (
`id` int(11) NOT NULL,
`wrench_lookup_string` varchar(200) NOT NULL,
`wrench_label` varchar(200) NOT NULL,
`created_at` datetime NOT NULL,
`updated_at` datetime NOT NULL
) ENGINE=MyISAM DEFAULT CHARSET=utf8mb4;
--
-- Indexes for dumped tables
--
--
-- Indexes for table `socket`
--
ALTER TABLE `socket`
ADD PRIMARY KEY (`id`),
ADD UNIQUE KEY `wrench_value` (`socket_value`,`socket_label`),
ADD KEY `wrench_id` (`wrench_id`);
--
-- Indexes for table `autosocket`
--
ALTER TABLE `autosocket`
ADD PRIMARY KEY (`id`),
ADD UNIQUE KEY `wrench_value` (`socket_value`,`socket_label`),
ADD KEY `wrench_id` (`wrench_id`);
--
-- Indexes for table `socketsource`
--
ALTER TABLE `socketsource`
ADD PRIMARY KEY (`id`);
--
-- Indexes for table `socket_user`
--
ALTER TABLE `socket_user`
ADD PRIMARY KEY (`id`);
--
-- Indexes for table `wrench`
--
ALTER TABLE `wrench`
ADD PRIMARY KEY (`id`),
ADD UNIQUE KEY `wrench_lookup_string` (`wrench_lookup_string`),
ADD UNIQUE KEY `wrench_label` (`wrench_label`);
Then you determine what options are available for each "wrench" by putting the options in the socket database, linking to the right wrench_id. The user can choose which socket they want (they will only see the socket_label text). And then in your code.. you get to have the contents of that choice (socket_value) available to help build your SQL.
Take a look at the NorthwindCustomerSocketReport.php in the examples/reports directory for an idea of the implications and an example use case.
This lets you write one report that operates on a multitude of underlying tables.. and the end user can choose which tables they want the report to target, for instance.
Please refer to the Troubleshooting guide.
Zermelo has been developed by CareSet Systems which provides extensive reporting on CMS, Medicare and Medicaid data. We developed Zermelo to make that task easier. CareSet systems uses Set Theory, SQL and Graph technology to datamine Medicare claims data. We chose the name "CareSet" for our company to highlight our data approach (our logo contains a graph and a 'set' of nodes, which we thought was a good illustration of our analytical approach. In any case, because of our focus on Set-theory approaches to data analytics we thought we should celebrate a famous set theory mathematician with the names of our Open Source projects.
Earnst Zermelo was one of the two independant mathematicians to posit the famous Russell's Paradox, the other being Russell. That paradox is the facinating question "Does a set that contains all sets that are not includes in themselves, contain itself". This paradox was a direct result of Cantor's work on Set Theory. All of which are critical chapters in the work on Foundational Mathematics shortly after the Turn of the 19th century.
So we figured Zermelo did not get enough credit for his independant development of the paradox (and his other work generally) and also, he has a cool name that is not really used much by software projects, with the exception of previous work automating table tennis tournaments or scheduling dutch students, which are both different Zermelo software solutions from this project. But so far, no one has a reporting engine with this name, so we jumped at the opportunity to celebrate Zermelo's contribution to mathematics and data analysis by naming our php reporting engine after him!