Skip to content

README.md

Latest commit

 

History

History
98 lines (65 loc) · 2.65 KB

README.md

File metadata and controls

98 lines (65 loc) · 2.65 KB

FindInFiles

FindInFiles is a smart chatbot that helps you quickly retrieve, search, and summarize information from documents. Simply place your files in a folder (docs), ask your question, search for specific words, or request a summary, and the bot will process the content to provide precise responses. It's an efficient solution for professionals, students, and researchers to streamline document analysis and save time.

Features

  • Intelligent Chat: Engage in a conversational manner to ask questions and get accurate answers based on document content.
  • Quick Search: Rapidly locate specific words or phrases within documents.
  • Document Summarization: Generate concise summaries of document content.
  • Advanced Language Model: Utilizes state-of-the-art natural language processing to understand and process queries.

Installation

  1. Clone the repository:

    git clone https://github.com/yourusername/FindInFiles.git
cd FindInFiles

  2. Install the required dependencies:

    pip install -r requirements.txt

  3. Set up your environment variables:

    • Copy .env-example to .env
    • Fill in your API keys and other required information

Usage

  1. Start the Flask server:

    python app.py

  2. Send POST requests to http://localhost:5000/process with the following JSON structure:

    For chat mode:

    {
  "query": "Your question here",
  "mode": "chat"
}

    For search mode:

    {
  "query": "search word",
  "mode": "search"
}

    For summary mode:

    {
  "query": "Number of lines for summary",
  "mode": "summary"
}

  3. The server will respond with the answer, search results, or summary based on the documents in the docs folder.

Configuration

  • Place your documents in the docs folder.
  • Adjust the max_history_size in chains.py to control the chat history length.
  • Modify the chunk_size and chunk_overlap in retrievers.py to fine-tune document processing.

Evaluation

To evaluate the performance of the chatbot, you can run the following command:

python evals.py

This will perform an experiment and evaluate the results based on the specified evaluators.

Evaluation Config

  • The evaluators are in the eval_utils/evaluators.py file.You can add more evaluators to this file.
  • Dataset is in the eval_utils/datasets.py file. Modify the dataset to add more examples.
  • Experiment is in the eval_utils/experiments.py file.

Evaluation Results

  • The evaluation results will be displayed in the langsmith console.
  • Visit link that appears in the terminal to see the evaluation results.