simple_fix.py

#!/usr/bin/env python3
"""
Simple script to regenerate embeddings without importing full NoteGraph
"""

import os
import sqlite3
import sqlite_vec
import numpy as np

def main():
    print("🔧 Simple embeddings fix script...")

    # Check if OpenAI API key is available
    if not os.getenv('OPENAI_API_KEY'):
        print("❌ OPENAI_API_KEY environment variable not found!")
        print("Please set your OpenAI API key:")
        print("export OPENAI_API_KEY='your-api-key-here'")
        return False

    try:
        # OpenAI import and setup
        from openai import OpenAI
        client = OpenAI()

        # Connect to database
        conn = sqlite3.connect('note_graph.db')
        conn.enable_load_extension(True)

        # Load the sqlite-vec extension
        extension_path = '/opt/homebrew/lib/python3.14/site-packages/sqlite_vec/vec0.dylib'
        conn.load_extension(extension_path)
        cursor = conn.cursor()

        print("📚 Database connected successfully")

        # Check current state
        cursor.execute("SELECT COUNT(*) FROM notes")
        notes_count = cursor.fetchone()[0]
        print(f"📝 Found {notes_count} notes in database")

        cursor.execute("SELECT COUNT(*) FROM note_embeddings")
        embeddings_count = cursor.fetchone()[0]
        print(f"🔢 Current embeddings: {embeddings_count}")

        # Get all notes
        cursor.execute("SELECT id, title, content FROM notes")
        notes = cursor.fetchall()

        if not notes:
            print("❌ No notes found in database!")
            return False

        print(f"\n🔄 Regenerating embeddings for {len(notes)} notes...")

        # Clear existing embeddings
        cursor.execute("DELETE FROM note_embeddings")
        conn.commit()
        print("🧹 Cleared existing embeddings")

        # Generate embeddings for each note using OpenAI
        success_count = 0
        for i, (note_id, title, content) in enumerate(notes, 1):
            print(f"📊 Processing note {i}/{len(notes)}: {title[:50]}...")

            try:
                # Create full text for embedding
                full_text = f"{title}\n\n{content}"

                # Generate embedding using OpenAI text-embedding-3-small
                response = client.embeddings.create(
                    model="text-embedding-3-small",
                    input=full_text
                )
                embedding = response.data[0].embedding
                embedding_array = np.array(embedding, dtype=np.float32)

                # Store embedding in database
                embedding_blob = embedding_array.tobytes()
                cursor.execute(
                    "INSERT INTO note_embeddings (note_id, embedding) VALUES (?, ?)",
                    (note_id, embedding_blob)
                )

                success_count += 1
                print(f"✅ Successfully generated embedding for note {note_id} ({len(embedding)} dimensions)")

            except Exception as e:
                print(f"❌ Failed to generate embedding for note {note_id}: {e}")

        conn.commit()
        print(f"\n🎉 Successfully generated {success_count}/{len(notes)} embeddings!")

        # Verify the embeddings
        cursor.execute("SELECT COUNT(*) FROM note_embeddings")
        final_count = cursor.fetchone()[0]
        print(f"✅ Final embedding count: {final_count}")

        conn.close()
        return True

    except Exception as e:
        print(f"❌ Error fixing embeddings: {e}")
        import traceback
        traceback.print_exc()
        return False

if __name__ == "__main__":
    success = main()
    exit(0 if success else 1)