openai_api

Interact with openai APIs using (DALLE2, GPT, Whisper). The code base is primairly in R, but refactored scripts are being developed and tested. For best results, use GPT 4.

Scripts

Below is a summary of each directory. Please see specific scripts for detailed roxygen about their uses with examples.

01 - api_base - Basic API connectivity.

• dalle_api.R - Sends prompt to generate and download images from DALLE.
• fine_tune_model.R - Function to fine-tune a GPT model using the OpenAI API.
• gpt_api.R - Sends a prompt to GPT via the OpenAI API and returns the generated text.
• whisper_api.R - Transcribes audio files.

02 - assistants - Scripts designed to assist in completeing specific tasks.

• gpt_audioquery.R - Record audio for a given duration, transcribe it using the Whisper API, and send the transcription to the GPT.
• gpt_dateparse.R - Parse dates with GPT.
• gpt_classifier.R - Classifies the type of data in each column in a dataset.
• gpt_finetune.R - Wrapper function for fine-tuning models using your own documents. Provides functions to upload documents, check status, and cancel jobs.
• gpt_gridsearch.R - Performs a grid search over the specified GPT models and parameters, using a given prompt.
• gpt_nlp.R - Function that allows the user to convert text into root words, label the part of speech (noun, verb, adverb, etc), or analyze the sentiment of text (-1, 0, 1).
• gpt_persona.R - Illustrates how to vary system messages (change roles/personas) across a dataset to generate multiple responses.
• gpt_random_number_experiment.R - Based on the gpt_gridsearch.R assistant, allows use to vary prompts, temperature, system messages, and add context. Designed to perform an experiment whereby GPT picks random numbers between 0 - 10.
• gpt_read.R - Reads a list of text chunks and a question, and uses the GPT model to generate a response based on the text.
• gpt_sentiment.R - Conducts sentiment analysis on a given dataframe using the GPT.

03 - dashboards - Interactive dashboards integrated with an OpenAI API.

• gpt_sentiment_assistant.R - Use GPT to analyze sentiment via a Shiny dashboard. Populates barplot and searchable table.

04 - refactored - Scripts refactored into other languages.

See a table of scripts and refactored languages below.

Package Name	Description	Python	Ruby
gpt_api	Call the GPT API.	✓	✓
gpt_classifier	Classify/label data using GPT.	✓
gpt_dateparser	Parse dates using GPT.	✓	✓
gpt_read	Read documents and ask questions.	✓
gpt_sentiment	Code sentiment of text using GPT.	✓	✓
whisper_api	Translate or transcribe audio using GPT.	✓