backend/debug_score.py

import asyncio
import os
import json
from supabase import create_client
from dotenv import load_dotenv
from src.matching.similarity import calculate_granular_match_score

load_dotenv()

SUPABASE_URL = os.environ.get("SUPABASE_URL")
SUPABASE_KEY = os.environ.get("SUPABASE_SERVICE_ROLE_KEY") or os.environ.get("SUPABASE_KEY")
client = create_client(SUPABASE_URL, SUPABASE_KEY)

async def run_test():
    res = client.table("applications").select("user_id, job_id").limit(1).execute()
    if not res.data:
        print("No apps found")
        return
        
    c_id = res.data[0]["user_id"]
    j_id = res.data[0]["job_id"]
    
    # Raw fetch
    p_emb = client.table("profile_embeddings").select("*").eq("id", c_id).execute().data[0]
    j_emb = client.table("job_embeddings").select("*").eq("job_id", j_id).execute().data[0]
    
    log = []
    log.append(f"Testing {c_id} against {j_id}")
    
    def get_len(v):
        if v is None: return "None"
        if isinstance(v, str): 
            try:
                # Approximate len by comma count
                return v.count(',') + 1
            except: return "StringError"
        return len(v)

    log.append("\n--- Profile Lengths ---")
    for k in ['skills', 'technical_skills', 'experience', 'certifications']:
        log.append(f"{k}: {get_len(p_emb.get(k))}")
        
    log.append("\n--- Job Lengths ---")
    for k in ['skills', 'technical_skills', 'work_experience', 'certifications']:
        log.append(f"{k}: {get_len(j_emb.get(k))}")
        
    result = await calculate_granular_match_score(client, c_id, j_id)
    log.append(f"\nResult: {json.dumps(result)}")
    
    with open("debug_log.txt", "w") as f:
        f.write("\n".join(log))
    print("Logged to debug_log.txt")

if __name__ == "__main__":
    asyncio.run(run_test())