Jobs with LLM Analyzer with multiple providers

 

This repository provides 2 new types of jobs for Visual TOM:

• Jobs where the prompt is written directly in the job (internal script field)
• Jobs where the prompt is written in a file and passed as a parameter to the job In both types of jobs, the LLM configuration is possible to choose between Groq, OpenAI and Claude.

Features

• CSV, JSON, Excel, Parquet, PDF analysis
• JSON output
• Text summary
• Temperature and max tokens configuration
• Max samples configuration
• Provider configuration

Disclaimer

No Support and No Warranty are provided by Absyss SAS for this project and related material. The use of this project's files is at your own risk.

Absyss SAS assumes no liability for damage caused by the usage of any of the files offered here via this Github repository.

Consultings days can be requested to help for the implementation.

Prerequisites

• Visual TOM 7.2 or greater
• Python 3.10 or greater
• An API key for chosen provider
• Visual TOM Unix agent

Instructions

Installation

• Install the requirements.txt file
• Create the batch queue in Visual TOM (in GUI and on the Agent):
  • gen-prompt : the prompt is written directly in the job (internal script field)
  • gen-ai : the prompt is written in a file and passed as a parameter to the job
• Import the XML file to create the custom application in Visual TOM

Usage

For both types of jobs, it is required to set the API key for the chosen provider in:

• Visual TOM environment variables using a secret Resource or
• .env file

The names of the environment variables are :

• GROQ_API_KEY
• OPENAI_API_KEY
• ANTHROPIC_API_KEY

Prompt written directly in the job

• Create a job with the gen-prompt batch queue
• In the internal script field, write the prompt
• Pass the following parameters:
  • --data : the path to the data file to analyze (supports: .csv, .xlsx, .xls, .json, .parquet, .pdf)
  • --provider : the provider to use (default: groq)
  • --model : the model to use (default: groq)
  • --max-samples : the maximum number of samples to analyze (default: 50)
  • --temperature : the temperature for the LLM generation (default: 0.3)
  • --max-tokens : the maximum number of tokens to generate (default: 2000)

Prompt written in a file

• Create a job with the gen-ai batch queue
• Fill the parameters:
  • Prompt file : the path to the prompt file
  • Input file : the path to the data file to analyze (supports: .csv, .xlsx, .xls, .json, .parquet, .pdf)
  • Provider : the provider to use (default: groq)
  • Model : the model to use (default: groq)
  • Temperature : the temperature for the LLM generation (default: 0.3)
  • Max tokens : the maximum number of tokens to generate (default: 2000)
  • Max samples : the maximum number of samples to analyze (default: 50)

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details

Code of Conduct

Absyss SAS has adopted the Contributor Covenant as its Code of Conduct, and we expect project participants to adhere to it. Please read the full text so that you can understand what actions will and will not be tolerated.