setup.py

import os
import sys
import subprocess
import requests
from pathlib import Path

# Fix python path to allow importing from src
sys.path.append(str(Path(__file__).resolve().parent))

def check_dependencies():
    """Verify system-level dependencies are accessible."""
    print("[Setup] Checking system binaries...")
    
    # Check Tesseract OCR
    from config.config import TESSERACT_CMD
    if not os.path.exists(TESSERACT_CMD):
        print(f"[Warning] Tesseract binary not found at '{TESSERACT_CMD}'. OCR capabilities for scanned documents will be unavailable.")
        print("Install it via Homebrew: 'brew install tesseract'")
    else:
        print("  -> Tesseract OCR: FOUND")
        
    # Check Ollama connection
    from config.config import OLLAMA_BASE_URL
    try:
        res = requests.get(f"{OLLAMA_BASE_URL}/api/tags", timeout=2.0)
        if res.status_code == 200:
            print("  -> Ollama LLM Service: FOUND & RUNNING")
        else:
            print(f"[Warning] Ollama returned status {res.status_code}. Make sure Ollama server is running.")
    except Exception:
        print("[Warning] Ollama service not running. Please launch Ollama and run 'ollama serve'.")

def pull_ollama_model():
    """Pull default model if it's missing on Ollama server."""
    from config.config import OLLAMA_BASE_URL, DEFAULT_LLM_MODEL
    try:
        print(f"[Setup] Checking if default model '{DEFAULT_LLM_MODEL}' is pulled on Ollama...")
        res = requests.get(f"{OLLAMA_BASE_URL}/api/tags", timeout=2.0)
        models = [m["name"] for m in res.json().get("models", [])]
        
        # Check for phi3 or model variants
        if not any(DEFAULT_LLM_MODEL in m for m in models):
            print(f"[Setup] Default model '{DEFAULT_LLM_MODEL}' not found. Attempting to pull model via Ollama...")
            # We run a subprocess call to pull model
            subprocess.run(["ollama", "pull", DEFAULT_LLM_MODEL], check=True)
            print(f"  -> Pulled '{DEFAULT_LLM_MODEL}' successfully.")
        else:
            print(f"  -> Model '{DEFAULT_LLM_MODEL}': ALREADY PULLED")
    except Exception as e:
        print(f"[Warning] Could not automatically pull model '{DEFAULT_LLM_MODEL}': {e}.")
        print(f"Please run 'ollama pull {DEFAULT_LLM_MODEL}' in a terminal.")

def pre_index_documents():
    """Submit existing PDFs in data/raw/ to the auto-ingestion folder and wait for indexing to complete."""
    from src.ingestion.ingestion_queue import IngestionTaskQueue
    from config.config import RAW_DIR, UPLOADED_DOCS_DIR
    import shutil
    import time
    
    # 1. Copy active PDFs to uploaded_documents
    if RAW_DIR.exists():
        pdf_files = list(RAW_DIR.glob("*.pdf"))
        print(f"[Setup] Submitting {len(pdf_files)} active PDFs to Auto-Ingestion folder...")
        for file_path in pdf_files:
            dest_file = UPLOADED_DOCS_DIR / file_path.name
            if not dest_file.exists():
                shutil.copy2(str(file_path), str(dest_file))
                print(f"  -> Copied '{file_path.name}' to uploaded_documents/")
                
    # 2. Wait for queue worker to finish processing
    queue_mgr = IngestionTaskQueue()
    print("[Setup] Waiting for background indexing to complete...")
    while True:
        with queue_mgr.lock:
            all_done = True
            for task in queue_mgr.manifest.values():
                if task["status"] in ["queued", "processing"]:
                    all_done = False
                    break
        if all_done:
            break
        print("  -> Indexing in progress... waiting 2s")
        time.sleep(2)
    print("[Setup] Indexing complete. Manifest and Qdrant DB are in sync.")
    queue_mgr.close()


def main():
    print("="*60)
    print("  LOCAL RAG PIPELINE ENVIRONMENT SETUP  ")
    print("="*60)
    
    # 1. System checks
    check_dependencies()
    
    # 2. Model pull
    pull_ollama_model()
    
    # 3. Compile & Index samples
    pre_index_documents()
    
    # 4. Launch Web application
    print("\n[Setup] Launching Streamlit Web App...")
    cmd = ["python3.12", "-m", "streamlit", "run", "app.py"]
    print(f"Running command: {' '.join(cmd)}")
    
    try:
        subprocess.run(cmd)
    except KeyboardInterrupt:
        print("\n[Setup] Stopped Streamlit server.")

if __name__ == "__main__":
    main()