data3/check_relationship.py

#!/usr/bin/env python3
"""
Let's understand the relationship between the datasets by comparing a few records.
"""
import csv
import json

# Read function_dataset_v2.csv and check what each column represents
print("=== function_dataset_v2.csv structure ===")
with open('function_dataset_v2.csv', 'r', encoding='utf-8') as f:
    reader = csv.DictReader(f)
    headers = reader.fieldnames
    print(f"Headers: {headers}")
    
    # Get a row that HAS metadata
    print("\nFinding a row with complete metadata...")
    for row in reader:
        if row['repo_name'] and row['path'] and row['language']:
            print(f"\nSample row WITH metadata:")
            print(f"  original_index: {row['original_index']}")
            print(f"  function_index: {row['function_index']}")
            print(f"  repo_name: {row['repo_name']}")
            print(f"  path: {row['path']}")
            print(f"  language: {row['language']}")
            print(f"  function_name: {row['function_name']}")
            break

# Now check programming_problems.jsonl
print("\n\n=== programming_problems.jsonl structure ===")
with open('programming_problems.jsonl', 'r', encoding='utf-8') as f:
    # Find an entry with row_number that might match
    for line in f:
        data = json.loads(line.strip())
        # Just show first entry
        print(f"First entry:")
        print(f"  row_number: {data.get('row_number')}")
        print(f"  metadata.original_index: {data['metadata']['original_index']}")
        print(f"  metadata.function_name: {data['metadata']['function_name']}")
        print(f"  metadata.repo_name: '{data['metadata']['repo_name']}'")
        print(f"  metadata.path: '{data['metadata']['path']}'")
        print(f"  metadata.language: '{data['metadata']['language']}'")
        break

# The key question: does row_number in JSONL match the row number in CSV?
print("\n\n=== Checking if row_number matches CSV row ===")
with open('programming_problems.jsonl', 'r', encoding='utf-8') as f:
    data = json.loads(f.readline())
    target_row = data.get('row_number')
    print(f"JSONL row_number: {target_row}")

# Get that row from CSV (row_number is probably 1-indexed after header)
with open('function_dataset_v2.csv', 'r', encoding='utf-8') as f:
    reader = csv.DictReader(f)
    for i, row in enumerate(reader):
        if i + 1 == target_row:  # CSV rows are 1-indexed
            print(f"\nCSV row {target_row}:")
            print(f"  original_index: {row['original_index']}")
            print(f"  repo_name: '{row['repo_name']}'")
            print(f"  path: '{row['path']}'")
            print(f"  language: '{row['language']}'") 
            print(f"  function_name: '{row['function_name']}'")
            
            # Check if function names match
            if row['function_name'] == data['metadata']['function_name']:
                print(f"\n✅ Function names match! We should use row_number as the key.")
            else:
                print(f"\n❌ Function names don't match.")
            break