Skip to content

getting-started.md

Latest commit

 

History

History
233 lines (169 loc) · 4.58 KB

getting-started.md

File metadata and controls

233 lines (169 loc) · 4.58 KB

Getting Started with RLM-REPL

This guide will help you get started with RLM-REPL, from installation to your first query.

Installation

Prerequisites

  • Python 3.9 or higher
  • A language model API endpoint (OpenAI-compatible)

Install from PyPI

pip install rlm-repl

Install from Source

git clone https://github.com/labKnowledge/rlm-repl-sql.git
cd rlm-repl-sql
pip install -e .

Verify Installation

rlm-repl --version

Setting Up Your Model

RLM-REPL works with any OpenAI-compatible API. Here are some common setups:

Option 1: Ollama (Local Models)

  1. Install Ollama from ollama.ai
  2. Pull a model: 
    ollama pull qwen2.5-coder
  3. Start Ollama (usually runs automatically)
  4. Use in RLM-REPL: 
    config = RLMConfig(
    base_url="http://localhost:11434/v1",
    api_key="ollama",
    model="qwen2.5-coder",
)

Option 2: OpenAI

config = RLMConfig(
    base_url="https://api.openai.com/v1",
    api_key="sk-your-api-key-here",
    model="gpt-4",
)

Option 3: LMStudio

  1. Install and start LMStudio
  2. Load a model and start the local server
  3. Use: 
    config = RLMConfig(
    base_url="http://localhost:1234/v1",
    api_key="lm-studio",
    model="your-model-name",
)

Option 4: Environment Variables

Set these environment variables to use defaults:

export RLM_BASE_URL="http://localhost:11434/v1"
export RLM_API_KEY="ollama"
export RLM_MODEL="qwen2.5-coder"

Then create config from environment:

from rlm_repl import RLMConfig

config = RLMConfig.from_env()

Your First Query

Using the CLI

  1. Create a text file (document.txt):

    # Introduction to Python

Python is a high-level programming language known for its simplicity.

## Features

- Easy to learn
- Versatile
- Large standard library

  2. Run RLM-REPL:

    rlm-repl document.txt

  3. In interactive mode, ask questions:

    > What are Python's features?

Using the Python API

from rlm_repl import RLMREPL, RLMConfig

# Create configuration
config = RLMConfig(
    base_url="http://localhost:11434/v1",
    api_key="ollama",
    model="qwen2.5-coder",
)

# Create REPL instance
with RLMREPL(config) as repl:
    # Load document
    repl.load_document("document.txt")
    
    # Ask a question
    result = repl.ask("What are Python's features?")
    
    # Print answer
    print(result.answer)
    print(f"Read {result.total_words} words in {result.elapsed_time:.1f}s")

Loading Text Directly

You can also load text directly without a file:

text = """
# My Document

This is some content.
"""

with RLMREPL(config) as repl:
    repl.load_text(text, "my_document")
    result = repl.ask("What is this document about?")
    print(result.answer)

Understanding the Output

When you ask a question, RLM-REPL will:

  1. Load the document (if not already loaded)
  2. Perform reading iterations:
    • Iteration 1: Overview (read beginning)
    • Iteration 2: Search (find keywords)
    • Iteration 3+: Deep read (extract details)
  3. Synthesize the answer from gathered information

Verbose Mode (Default)

With verbose mode enabled, you'll see:

  • Each iteration's strategy
  • SQL queries being executed
  • Results retrieved
  • Final answer

Quiet Mode

For cleaner output:

config = RLMConfig(
    base_url="http://localhost:11434/v1",
    api_key="ollama",
    model="qwen2.5-coder",
    verbose=False,  # Disable verbose output
)

Or in CLI:

rlm-repl document.txt -q --question "Your question"

Next Steps

Common Issues

"No document loaded"

Make sure to call load_document() or load_text() before asking questions:

repl.load_document("file.txt")  # Must load first!
result = repl.ask("Question?")

Connection Errors

If you see connection errors:

  1. Verify your model API is running
  2. Check the base_url is correct
  3. Test the API endpoint directly (e.g., curl http://localhost:11434/v1/models)

Model Not Found

Ensure:

  • The model name matches exactly (case-sensitive)
  • The model is installed/pulled (for local models)
  • You have access to the model (for cloud APIs)

For more troubleshooting, see Troubleshooting Guide.