# Log Parser CLI

## Overview

This project provides a simple command‑line tool that parses raw HTTP log lines into typed API events using **Pydantic v2** and **LangChain** structured output. The script accepts a string or a file containing log lines, parses each line into one of two event types (`HttpOkEvent` or `HttpErrorEvent`), and prints a summary table.

## Installation

```bash
# Create a virtual environment (recommended)
python -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt
```

The `requirements.txt` contains the following packages:

- `langchain-core`
- `langchain-openai`
- `pydantic`
- `python-dotenv`

## Environment Variables

The tool uses OpenAI (or any compatible LLM) for structured parsing. Set the following variables in a `.env` file or your environment:

```env
OPENAI_API_KEY=your_api_key
OPENAI_BASE_URL=https://api.openai.com/v1
OPENAI_MODEL=gpt-3.5-turbo
```

If you are using a local LLM (e.g., Ollama or LM Studio), set:

```env
OPENAI_BASE_URL=http://localhost:11434/v1
OPENAI_API_KEY=ollama
OPENAI_MODEL=llama3
```

## Usage

```bash
# Parse a file containing log lines
python agent.py --input logs.txt

# Or provide raw log string directly
python agent.py --input "200 /api/v1/home\n404 /api/v1/notfound error: Not Found"

# If no input is provided, a sample log is used
python agent.py
```

## Sample Log

The repository contains a sample log you can use for testing:

```
200 /api/v1/users
404 /api/v1/items error: Not Found
500 /api/v1/orders error: Internal Server Error
```

## Output

The script prints a table with columns: `Kind`, `Status`, `Path`, and `Error` (if applicable). Example:

```
Kind  Status Path                Error                          
----- ------ -------------------- --------------------------------
ok    200    /api/v1/users         
error 404    /api/v1/items         Not Found                     
error 500    /api/v1/orders        Internal Server Error         
```

## Extending

You can extend the Pydantic models or the parsing logic by editing `agent.py`. The LLM prompt is generated automatically using the `PydanticOutputParser` schema instructions.

---

Feel free to open issues or pull requests if you encounter bugs or have suggestions.