Download test_semantic_search.py from StillWorking/sentinel-face-recognition: direct link, hf CLI and curl.
- Browser
- Download file 5.33 kB
-
https://huggingface.co/StillWorking/sentinel-face-recognition/resolve/main/test_semantic_search.py
- Command line
-
hf download hf://StillWorking/sentinel-face-recognition/test_semantic_search.py
-
curl -L -o test_semantic_search.py https://huggingface.co/StillWorking/sentinel-face-recognition/resolve/main/test_semantic_search.py
5.33 kB
| """ | |
| End-to-end smoke test for the Sentinel-AI API. | |
| Make sure the backend is running before executing this script. | |
| Usage: | |
| python test_semantic_search.py | |
| """ | |
| import json | |
| import os | |
| import sys | |
| from typing import Any | |
| import requests | |
| API_BASE = os.getenv("SENTINEL_API_BASE_URL", "http://localhost:8000").rstrip("/") | |
| # Optional shared-deployment token (start the backend with SENTINEL_API_KEY). | |
| _API_KEY = os.getenv("SENTINEL_API_KEY", "").strip() | |
| _HEADERS = {"Authorization": f"Bearer {_API_KEY}"} if _API_KEY else {} | |
| def _print_header(title: str) -> None: | |
| print("\n" + "=" * 70) | |
| print(title) | |
| print("=" * 70) | |
| def test_health() -> bool: | |
| """Test if the backend is running.""" | |
| _print_header("TEST: Health Check") | |
| try: | |
| response = requests.get(f"{API_BASE}/health", headers=_HEADERS, timeout=5) | |
| response.raise_for_status() | |
| data = response.json() | |
| print("Backend is running") | |
| print(f" Status: {data['status']}") | |
| print(f" Index ready: {data['index_ready']}") | |
| print(f" Model: {data['model']}") | |
| return bool(data.get("status") == "ok") | |
| except requests.exceptions.ConnectionError: | |
| print(f"Cannot connect to {API_BASE}") | |
| print(" Start the backend with:") | |
| print(" cd backend && python -m uvicorn main:app --host 127.0.0.1 --port 8000 --reload") | |
| return False | |
| except Exception as exc: | |
| print(f"Health check failed: {exc}") | |
| return False | |
| def test_semantic_search(query: str, top_k: int = 5) -> dict[str, Any] | None: | |
| """Run one semantic-search query and return the response (or None).""" | |
| _print_header("TEST: Semantic Search") | |
| print(f"Query: '{query}'") | |
| print(f"Top K: {top_k}") | |
| try: | |
| response = requests.post( | |
| f"{API_BASE}/api/v1/semantic-search", | |
| json={"query": query, "top_k": top_k}, | |
| headers=_HEADERS, | |
| timeout=30, | |
| ) | |
| response.raise_for_status() | |
| data = response.json() | |
| attrs = data.get("extracted_attrs") or [] | |
| excluded = data.get("excluded_attrs") or [] | |
| matches = data.get("matches") or [] | |
| print("Request successful") | |
| print(f" Status: {data.get('status')}") | |
| print(f" Extracted attributes: {', '.join(attrs) if attrs else 'None'}") | |
| print(f" Excluded attributes: {', '.join(excluded) if excluded else 'None'}") | |
| print(f" Matches found: {len(matches)}") | |
| metrics = data.get("cascadeflow_metrics") | |
| if metrics: | |
| print(f" Latency: {metrics.get('latency_ms', 0):.1f} ms") | |
| print(f" Cost: ${metrics.get('cost_usd', 0):.6f}") | |
| if matches: | |
| print(f"\nTop {min(3, len(matches))} matches:") | |
| for i, match in enumerate(matches[:3], 1): | |
| shown = ", ".join(match["attributes"][:5]) | |
| print(f" {i}. Image ID: {match['image_id']} (faiss_id={match['faiss_id']})") | |
| print(f" Attributes: {shown}...") | |
| return data | |
| except requests.exceptions.Timeout: | |
| print("Request timed out (>30 seconds)") | |
| return None | |
| except requests.exceptions.HTTPError as exc: | |
| print(f"HTTP Error: {exc}") | |
| try: | |
| print(f" Detail: {json.dumps(response.json(), indent=2)}") | |
| except Exception: | |
| print(f" Response: {response.text}") | |
| return None | |
| except Exception as exc: | |
| print(f"Request failed: {exc}") | |
| return None | |
| def main() -> int: | |
| """Run all tests and return a process exit code.""" | |
| _print_header("SENTINEL-AI SEMANTIC SEARCH TEST SUITE") | |
| if not test_health(): | |
| print("\nBackend is not running. Please start it first.") | |
| return 1 | |
| results: list[tuple[str, bool]] = [] | |
| # A response is only a pass if the request succeeded *and* the extracted | |
| # attributes are sensible for the query. | |
| data = test_semantic_search("smiling woman with blond hair", top_k=5) | |
| ok = bool(data) and "Male" in data.get("excluded_attrs", []) | |
| if ok: | |
| ok = not any("Male" in m.get("attributes", []) for m in data.get("matches", [])) | |
| results.append(("Smiling woman query (excludes Male from results)", ok)) | |
| data = test_semantic_search("young male with wavy hair and eyeglasses", top_k=5) | |
| ok = bool(data) and {"Male", "Young"} <= set(data.get("extracted_attrs", [])) | |
| results.append(("Young male with glasses query", ok)) | |
| data = test_semantic_search("person with beard and mustache", top_k=5) | |
| ok = bool(data) and "No_Beard" not in data.get("extracted_attrs", []) | |
| results.append(("Beard/mustache query (no No_Beard inversion)", ok)) | |
| data = test_semantic_search("alien with three eyes", top_k=5) | |
| ok = bool(data) and not data.get("matches") | |
| results.append(("Edge-case query returns no matches", ok)) | |
| _print_header("SUMMARY") | |
| for label, passed in results: | |
| print(f" [{'PASS' if passed else 'FAIL'}] {label}") | |
| passed_count = sum(1 for _, passed in results if passed) | |
| print(f"\nTests passed: {passed_count}/{len(results)}") | |
| if passed_count == len(results): | |
| print("\nALL TESTS PASSED") | |
| return 0 | |
| print(f"\n{len(results) - passed_count} test(s) failed.") | |
| return 1 | |
| if __name__ == "__main__": | |
| sys.exit(main()) | |