Skip to content

scripts.md

Latest commit

 

History

History
158 lines (128 loc) · 4.6 KB

scripts.md

File metadata and controls

158 lines (128 loc) · 4.6 KB
Raw

PineScript Expert Agent: Scripts Overview

This document provides a comprehensive overview of all the scripts in the PineScript Expert Agent project, explaining their purpose, functionality, and usage order.

Core Components

1. agent.py

Purpose: Main implementation of the PineScript Expert Agent using Pydantic AI. Features:

  • Defines the agent with OpenAI/OpenRouter model configuration
  • Implements the retrieval tool for searching Pine Script documentation
  • Contains the dependency injection mechanism for database and API connections
  • Handles API key management and validation

2. db_schema.py

Purpose: Defines the database schema for storing Pine Script documentation. Features:

  • Creates tables with pgvector support for vector embeddings
  • Includes validation functions to verify the database setup
  • Defines the schema for the pinescript_docs table

Setup and Initialization Scripts

3. setup.py

Purpose: Initial project setup script to prepare the environment. Features:

  • Creates a virtual environment
  • Installs dependencies from requirements.txt
  • Sets up the .env file from the template
  • Makes the shell script executable Usage: Run this first when setting up a new installation.

4. init_db.py

Purpose: Initializes the database with the required schema. Features:

  • Creates necessary tables and extensions
  • Verifies that pgvector is installed
  • Checks and reports on database connection status Usage: Run after setup.py to prepare the database.

Data Management Scripts

5. clear_database.py

Purpose: Utility to clear the database for a fresh start. Features:

  • Removes all data from the pinescript_docs table
  • Keeps schema intact
  • Provides confirmation prompt for safety

6. pinescript_crawler.py

Purpose: Crawls the TradingView Pine Script documentation. Features:

  • Uses crawl4ai to extract documentation from the TradingView website
  • Processes and splits documentation into sections
  • Generates embeddings and stores in the database Usage: Run after init_db.py to populate the database with documentation.

7. db_inspect.py

Purpose: Utility to inspect and query the database. Features:

  • Counts entries in the database
  • Lists document titles and URLs
  • Tests search functionality
  • Verifies vector quality

User Interface Scripts

8. run.py

Purpose: Main entry point that provides multiple command options. Features:

  • interactive: Launches interactive shell
  • query: Processes a single query
  • check: Verifies database setup Usage: The recommended way to interact with the agent.

9. interactive.py

Purpose: Provides an interactive command-line interface. Features:

  • Command-line conversation with the agent
  • History management
  • Example queries Usage: Run directly or through run.py interactive.

10. streamlit_ui.py

Purpose: Streamlit UI with persistent conversation history. Features:

  • Saves chat history to disk
  • Maintains conversation context between sessions
  • Same features as the regular Streamlit UI
  • Example queries
  • Status monitoring for database and API keys

Usage: Run with streamlit run streamlit_ui.py

Debugging and Testing Scripts

11. api_debug.py

Purpose: Debugging tool for API connections. Features:

  • Tests OpenAI API connection
  • Diagnoses API key issues
  • Tests various client configurations Usage: Use when troubleshooting API connection problems.

Setup Process Workflow

For a new installation, follow these steps in order:

  1. Initial Setup:

    python setup.py

    This prepares your environment, installing dependencies and setting up configuration files.

  2. Database Initialization:

    python init_db.py

    This creates the necessary database schema with pgvector support.

  3. Populate the Database (Crawl documentation):

    python pinescript_crawler.py

  4. Verify Setup:

    python run.py check

    This confirms that everything is set up correctly.

  5. Run the Agent (choose one interface):

    # Command-line interface
python run.py interactive

# Web interface
streamlit run streamlit_ui.py

Maintenance Tasks

  • Clear Database: If you need to start fresh with documentation:

    python clear_database.py

  • Inspect Database: To check the contents of the database:

    python db_inspect.py count
python db_inspect.py list
python db_inspect.py search "your query"

  • Debug API Issues: If you encounter API problems:

    python api_debug.py