README.md

LLM Document Manager

This application enables users to upload company documents and generate relevant metadata. It supports various document formats including PDF, DOC, DOCX, TXT, CSV, and XLS.

Key Features

1. Instructor Integration: Utilizes the Instructor library to enhance the OpenAI client, ensuring reliability and consistency in the LLM's data structures.

• For more information, visit: https://jxnl.github.io/instructor/

2. Pydantic Data Models: Leverages Pydantic to define data models, including optional fields and various field validation requirements, ensuring robust and clear data handling.

3. Llama Index Parsing: Employs Llama Index for parsing documents, enabling efficient and effective extraction of information from supported formats.

• For more information, visit: https://www.llamaindex.ai/, https://llamahub.ai/

Installation

1. Clone the repository:

   git clone https://github.com/duncandevs/llm-document-manager.git
   cd llm-document-manager

2. Install the backend dependencies: from root cd into back-end

   cd back-end

   pip install -r requirements.txt

   pip install -e .

3. Set the OpenAI key:

   replace OPEN_AI_KEY in back-end .env file with your openai key

4. Install the front end dependencies from root cd into front-end

   cd front-end

   run install command with yarn

   yarn install

Starting the App

1. From root Start the pocketbase db instance, run the following command:

   ./pocketbase serve

   you may encounter a permission error at this step. If this happens open the root folder and double click the pocketbase exec file to grant permission, then retry this step.

2. From back-end dir, run the following command:

   start-app

3. From front-end, run the following command:

   yarn dev

Usage:

1. View local app url

   localhost:3000

2. View local database url

   http://127.0.0.1:8090/_/

   enter the following DB credentials email: [email protected] password: password1234

Resources:

• useful materials to continue building on this demo

1. Jason Liu guide to prompt Engineering:

• https://www.oreilly.com/radar/what-we-learned-from-a-year-of-building-with-llms-part-i/

2. Instructor For Structured LLMs:

• https://www.oreilly.com/radar/what-we-learned-from-a-year-of-building-with-llms-part-i/

3. LLama Index Document Summary Index:

• https://docs.llamaindex.ai/en/stable/examples/index_structs/doc_summary/DocSummary/

4. LLama Index Pydantic Metadata Extractor

• https://docs.llamaindex.ai/en/stable/examples/metadata_extraction/PydanticExtractor/