AGENTS.md

Project Overview

This project is a Python-based package designed to assist users during fantasy football drafts. It will:

Accept a user-provided PDF sheet as a starting point for rankings.
Parse the PDF to extract player rankings and relevant data, including support for multi-column/tabular formats.
Use LLMs (such as Llama 2 or OpenAI GPT) to help interpret and structure complex or inconsistent PDF data.
Connect to the Sleeper API (and ESPN if possible) to:
- Pull results of completed drafts.
- Analyze the average draft position (ADP) of players.
Provide real-time suggestions for who to draft next based on:
- The user's sheet.
- Live draft results.
- ADP and other relevant data.

Maintain this document as the central source of project requirements and agent instructions.
All code should be written in Python.
Prioritize modular, testable, and well-documented code.
Use popular libraries for PDF parsing (e.g., PyPDF2, pdfplumber) and HTTP requests (e.g., requests).
Integrate LLMs (Llama 2, OpenAI GPT, etc.) for advanced parsing of tabular or complex PDF data.
Keep all API keys and secrets (e.g., OpenAI key) in environment variables, loaded from a
```
.env
```
file, and ensure
```
.env
```
is listed in
```
.gitignore
```
.
Ensure extensibility for supporting additional fantasy platforms (e.g., ESPN).
Provide clear interfaces for inputting the user's sheet and connecting to APIs.
Implement robust error handling and logging.
Update this document as requirements evolve or new features are added.

Set up the Python project structure.
Implement PDF parsing functionality, including LLM-based column/table extraction.
Integrate with the Sleeper API (and ESPN if feasible).
Add OpenAI (or Llama 2) integration for advanced parsing.
Store all API keys in a
```
.env
```
file and keep it out of version control.
Develop logic for real-time draft recommendations.
Add CLI or web interface for user interaction.
Write tests and documentation.

This project is a Python-based package designed to assist users during fantasy football drafts. It will:

Accept a user-provided PDF sheet as a starting point for rankings.
Parse the PDF to extract player rankings and relevant data, including support for multi-column/tabular formats.
Use LLMs (such as Llama 2 or OpenAI GPT) to help interpret and structure complex or inconsistent PDF data.
Connect to the Sleeper API (and ESPN if possible) to:
- Pull results of completed drafts.
- Analyze the average draft position (ADP) of players.
Provide real-time suggestions for who to draft next based on:
- The user's sheet.
- Live draft results.
- ADP and other relevant data.

Maintain this document as the central source of project requirements and agent instructions.
All code should be written in Python.
Prioritize modular, testable, and well-documented code.
Use popular libraries for PDF parsing (e.g., PyPDF2, pdfplumber) and HTTP requests (e.g., requests).
Integrate LLMs (Llama 2, OpenAI GPT, etc.) for advanced parsing of tabular or complex PDF data.
Keep all API keys and secrets (e.g., OpenAI key) in environment variables, loaded from a
```
.env
```
file, and ensure
```
.env
```
is listed in
```
.gitignore
```
.
Ensure extensibility for supporting additional fantasy platforms (e.g., ESPN).
Provide clear interfaces for inputting the user's sheet and connecting to APIs.
Implement robust error handling and logging.
Update this document as requirements evolve or new features are added.

Set up the Python project structure.
Implement PDF parsing functionality, including LLM-based column/table extraction.
Integrate with the Sleeper API (and ESPN if feasible).
Add OpenAI (or Llama 2) integration for advanced parsing.
Store all API keys in a
```
.env
```
file and keep it out of version control.
Develop logic for real-time draft recommendations.
Add CLI or web interface for user interaction.
Write tests and documentation.