Project Overview

The ToolBoss application is designed to provide a platform for autonomous compliance and claims verification. It utilizes a combination of natural language processing (NLP) and machine learning algorithms to analyze user input and provide a verdict based on a set of predefined rules and policies.

Key Features

The application features two main components: LEXA AI and Meeting Summarizer. LEXA AI is a digital arbitrator that uses NLP to analyze user input and provide a verdict based on a set of predefined rules and policies. The Meeting Summarizer is a tool that uses speech recognition and NLP to summarize meetings and extract action items.

Architecture

The application is built using a microservices architecture, with each component communicating with the others through RESTful APIs. The LEXA AI component uses a state machine to manage the workflow, while the Meeting Summarizer uses a pipeline approach to process audio files.

Project Structure

The project is structured into several directories, including modules for the LEXA AI and Meeting Summarizer components, pages for the web application, and app.py for the main application entry point.

Tech Stack

The application uses a range of technologies, including Python, Streamlit, and Langchain, as well as several NLP and machine learning libraries.

Main Components

The main components of the application include app.py, which is the main entry point for the application, modules/lexa/engine.py, which manages the LEXA AI workflow, and pages/summarizer_page.py, which handles the Meeting Summarizer functionality.

Installation

To install the application, clone the repository and run pip install -r requirements.txt. Then, run streamlit run app.py to start the application.

Configuration

The application uses environment variables to configure the LEXA AI and Meeting Summarizer components. These variables can be set in a .env file or through the command line.

Usage

To use the application, navigate to the web interface and select either the LEXA AI or Meeting Summarizer component. Follow the prompts to input data and receive a verdict or summary.

Workflow

The workflow for the LEXA AI component involves the following steps:

1. User input: The user inputs a story or claim.
2. Intake: The input is processed and structured into a JSON object.
3. Analysis: The JSON object is analyzed using NLP and machine learning algorithms.
4. Verdict: A verdict is generated based on the analysis.

The workflow for the Meeting Summarizer component involves the following steps:

1. Audio input: The user uploads an audio file.
2. Transcription: The audio file is transcribed using speech recognition.
3. Summary: The transcript is summarized using NLP algorithms.
4. Action items: Action items are extracted from the summary.

Architecture Overview

The ToolBoss application is built using a microservices architecture, with each component communicating with the others through RESTful APIs.

flowchart TD
    A[User Input] -->|input| B[LEXA AI]
    B -->|verdict| C[User Feedback]
    D[Audio Input] -->|input| E[Meeting Summarizer]
    E -->|summary| F[User Feedback]
    G[Database] -->|data| B
    G -->|data| E

Loading

High-Level Design

The application is designed to be scalable and flexible, with each component able to be updated or replaced independently.

Project Structure

The project is structured into several directories, including modules for the LEXA AI and Meeting Summarizer components, pages for the web application, and app.py for the main application entry point.

Execution Flow

The execution flow for the LEXA AI component involves the following steps:

1. User input: The user inputs a story or claim.
2. Intake: The input is processed and structured into a JSON object.
3. Analysis: The JSON object is analyzed using NLP and machine learning algorithms.
4. Verdict: A verdict is generated based on the analysis.

The execution flow for the Meeting Summarizer component involves the following steps:

1. Audio input: The user uploads an audio file.
2. Transcription: The audio file is transcribed using speech recognition.
3. Summary: The transcript is summarized using NLP algorithms.
4. Action items: Action items are extracted from the summary.

Data Flow

The data flow for the application involves the following steps:

1. User input: The user inputs data into the application.
2. Processing: The input data is processed and analyzed using NLP and machine learning algorithms.
3. Storage: The processed data is stored in a database.
4. Retrieval: The stored data is retrieved and used to generate a verdict or summary.

Module Relationships

The modules in the application are related as follows:

• app.py is the main entry point for the application and imports the LEXA AI and Meeting Summarizer components.
• modules/lexa/engine.py manages the LEXA AI workflow and imports the interviewer and auditor components.
• pages/summarizer_page.py handles the Meeting Summarizer functionality and imports the transcription and summary components.

AI / ML Components

The application uses several AI and ML components, including:

• langchain for NLP and machine learning tasks
• streamlit for building the web application
• whisper for speech recognition

Storage and Data Layer

The application uses a database to store processed data and retrieve it for generating verdicts or summaries.

External Services

The application does not use any external services.

Technologies Used

The application uses a range of technologies, including Python, Streamlit, and Langchain, as well as several NLP and machine learning libraries.