Spaces:
Running
Running
File size: 18,275 Bytes
e9fe93a 900c34e e9fe93a 900c34e e9fe93a 900c34e e9fe93a 900c34e e9fe93a 900c34e e9fe93a 900c34e 25bfed8 900c34e e9fe93a 900c34e e9fe93a 900c34e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 | # ruff: noqa: E501
"""Full pipeline end-to-end test with real Ollama LLM.
Tests every agent node in the LangGraph search pipeline."""
from __future__ import annotations
import asyncio
import logging
import sys
from pathlib import Path
# Ensure project root is on path
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(name)s: %(message)s")
logger = logging.getLogger("e2e-test")
errors: list[str] = []
successes: list[str] = []
def check(step: str, ok: bool, detail: str = ""):
if ok:
successes.append(step)
logger.info(f" PASS [{step}] {detail}")
else:
errors.append(step)
logger.error(f" FAIL [{step}] {detail}")
import traceback # noqa: E402
async def run_tests():
from src.core.models import MatchResult, MatchScores
def _make_match_result(pid: str, name: str, title: str) -> MatchResult:
"""Build a minimal MatchResult for testing."""
return MatchResult(
query_id="e2e-test",
rank=0,
profile_id=pid,
name=name,
current_title=title,
current_company="",
matched_skills=[],
missing_skills=[],
scores=MatchScores(
semantic_similarity=0.5,
keyword_match=0.5,
skill_match=0.5,
experience_match=0.5,
location_match=0.5,
overall=0.5,
confidence=0.5,
),
)
# ==========================================================
# STEP 1: Config & LLM connectivity
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 1: Config & LLM Connectivity")
logger.info("=" * 60)
from src.core.config import get_llm_client, get_settings
settings = get_settings()
check("config-loads", settings.llm_provider == "ollama", f"provider={settings.llm_provider}")
client = get_llm_client()
try:
resp = await client.ainvoke("Say exactly: HELLO OLLAMA")
text = resp.content if hasattr(resp, 'content') else str(resp)
check("ollama-connect", "HELLO OLLAMA" in text.upper(), f"response: {text[:100]}")
except Exception as e:
check("ollama-connect", False, str(e))
# ==========================================================
# STEP 2: Profile Store & Index Loading
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 2: Profile Store & Index Loading")
logger.info("=" * 60)
from src.core.config import DATA_DIR
from src.core.profile_store import ProfileStore
from src.matching.scorer import DEFAULT_SLIDER_WEIGHTS, CandidateScorer
store = ProfileStore()
check("profile-store-loaded", len(store) > 0, f"profiles={len(store)}")
# Fetch a sample profile using get_all_sample
samples = store.get_all_sample()
check("sample-profiles-loaded", len(samples) > 0, f"count={len(samples)}")
sample = next(iter(samples.values())) if samples else None
check("profile-fetch", sample is not None, f"first profile id={sample.profile_id if sample else 'N/A'}")
# Load indexes
from src.language.multilingual import MultilingualEmbedder
from src.search.bm25_search import BM25Search
from src.search.hybrid import HybridSearch
from src.search.reranker import CrossEncoderReranker
from src.search.vector_search import VectorSearch
index_dir = DATA_DIR / "indexes"
faiss_path = index_dir / "faiss_index.bin"
bm25_path = index_dir / "bm25_index.pkl"
id_map_path = index_dir / "faiss_id_map.json"
check("faiss-exists", faiss_path.exists(), str(faiss_path))
check("bm25-exists", bm25_path.exists(), str(bm25_path))
vector_search = VectorSearch(dimension=384)
try:
vector_search.load(faiss_path, id_map_path)
check("faiss-load", vector_search.size > 0, f"vectors={vector_search.size}")
except Exception as e:
check("faiss-load", False, str(e))
bm25 = BM25Search()
try:
bm25.load(bm25_path)
check("bm25-load", bm25.size > 0, f"docs={bm25.size}")
except Exception as e:
check("bm25-load", False, str(e))
logger.info("Loading embedding model (may take ~30s on first run)...")
embedder = MultilingualEmbedder()
hybrid = HybridSearch(vector_search, bm25, embedder)
reranker = CrossEncoderReranker(timeout_ms=5000)
# Test search
try:
results = hybrid.search("python developer", top_k=5)
check("hybrid-search", len(results) > 0, f"results={len(results)}")
except Exception as e:
check("hybrid-search", False, str(e))
# Test empty search
try:
empty_results = hybrid.search("zzzzzzzzzznonexistent", top_k=5)
check("hybrid-empty-search", len(empty_results) > 0, "empty query still returns results as fallback")
except Exception as e:
check("hybrid-empty-search", False, str(e))
# ==========================================================
# STEP 3: Create executor with dependencies
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 3: Create agent dependencies")
logger.info("=" * 60)
scorer = CandidateScorer()
from src.agents.executor import ExecutorAgent
executor = ExecutorAgent(hybrid, reranker, scorer, store)
from src.agents.planner import PlannerAgent
planner = PlannerAgent()
from src.agents.reflector import ReflectorAgent
reflector = ReflectorAgent()
# ==========================================================
# STEP 4: Planner Agent (async)
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 4: Planner Agent")
logger.info("=" * 60)
try:
query = "senior DevOps engineer with 5+ years AWS experience in Bangalore"
plan = await planner.plan(query)
check("planner-plan", plan is not None, f"plan type={type(plan).__name__ if plan else 'None'}")
if plan:
req_skills = plan.required_skills
check("planner-has-required-skills", len(req_skills) > 0,
f"required_skills={[str(s) for s in req_skills[:5]]}")
check("planner-has-experience", plan.experience is not None,
f"exp={plan.experience}")
except Exception as e:
check("planner-plan", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 5: Executor Agent (async)
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 5: Executor Agent")
logger.info("=" * 60)
try:
exec_results = await executor.execute(plan, top_k=5)
check("executor-execute", exec_results is not None, f"type={type(exec_results).__name__}")
if exec_results and isinstance(exec_results, list):
check("executor-has-results", len(exec_results) > 0, f"count={len(exec_results)}")
if exec_results:
r = exec_results[0]
check("executor-result-has-score", hasattr(r, 'scores') and r.scores is not None,
f"scores={r.scores}")
except Exception as e:
check("executor-execute", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 6: Scorer with slider weights
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 6: Scorer with Slider Weights")
logger.info("=" * 60)
slider_weights = {
"skill_match": 80,
"experience_match": 70,
"education_match": 50,
"assessment_score": 40,
"behavioral_signals": 30,
"cultural_fit": 20,
}
try:
scores = scorer.compute_overall(
{
"semantic_similarity": 0.85,
"keyword_match": 0.75,
"skill_match": 0.90,
"experience_match": 0.80,
"location_match": 0.60,
"education_match": 0.70,
"cross_encoder_score": 0.50,
"behavioral_signals": 0.65,
"cultural_fit": 0.55,
},
slider_weights=slider_weights,
)
check("scorer-overall", scores.overall > 0, f"overall={scores.overall:.3f}")
check("scorer-confidence", scores.confidence > 0, f"confidence={scores.confidence:.3f}")
except Exception as e:
check("scorer-overall", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 7: Reflector Agent
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 7: Reflector Agent")
logger.info("=" * 60)
# Build a MatchResult list for the reflector
match_results = []
if exec_results and isinstance(exec_results, list):
for mr in exec_results:
match_results.append(mr)
try:
eval_results = await reflector.reflect(plan, match_results)
check("reflector-evaluate", eval_results is not None,
f"type={type(eval_results).__name__ if eval_results else 'None'}")
if eval_results and isinstance(eval_results, dict):
check("reflector-has-feedback", "feedback" in eval_results or "evaluations" in eval_results,
f"keys={list(eval_results.keys())}")
except Exception as e:
check("reflector-evaluate", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 8: Rationale Generator
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 8: Rationale Generator")
logger.info("=" * 60)
from src.rationale.generator import RationaleGenerator
gen = RationaleGenerator()
try:
# Use template rationale with a sample match result
if exec_results and isinstance(exec_results, list) and len(exec_results) > 0:
mr = exec_results[0]
profile = store.get(mr.profile_id)
if profile:
rationale = gen._template_rationale(mr, profile)
check("rationale-template-fallback", rationale is not None and len(rationale.summary) > 0,
f"summary_len={len(rationale.summary) if rationale else 0}")
else:
check("rationale-template-fallback", True, "skipped (no profile)")
else:
check("rationale-template-fallback", True, "skipped (no results)")
except Exception as e:
check("rationale-template-fallback", False, str(e))
traceback.print_exc()
# ==========================================================
# STEP 9: Full Orchestrator
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 9: Full Orchestrator (LangGraph)")
logger.info("=" * 60)
from src.agents.orchestrator import Orchestrator
orchestrator = Orchestrator(planner, executor, reflector)
try:
result = await orchestrator.run(
raw_query="senior software engineer Python AWS",
slider_weights=DEFAULT_SLIDER_WEIGHTS,
)
check("orchestrator-run", result is not None, f"type={type(result).__name__ if result else 'None'}")
if result:
from src.core.models import SearchResponse
if isinstance(result, SearchResponse):
results_arr = result.results
check("orchestrator-has-results", len(results_arr) > 0, f"count={len(results_arr)}")
search_metadata = result.search_metadata
check("orchestrator-has-metadata", search_metadata is not None,
f"metadata={search_metadata}")
if search_metadata:
check("orchestrator-processing-time", search_metadata.total_time_ms > 0,
f"time={search_metadata.total_time_ms}ms")
check("orchestrator-listwise",
search_metadata.listwise_ranked is not None,
f"listwise_ranked={search_metadata.listwise_ranked}")
except Exception as e:
check("orchestrator-run", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 10: Fairness Components
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 10: Fairness Components")
logger.info("=" * 60)
from src.fairness.anonymizer import anonymize_text_for_bias, style_anonymize
from src.fairness.bias_detector import BiasDetector
try:
pii_text = "John Doe worked at Google in Bangalore. Email: johndoe@email.com"
anon_text = anonymize_text_for_bias(pii_text)
check("pii-anonymize-text", "[NAME]" in anon_text, f"result: {anon_text[:80]}")
except Exception as e:
check("pii-anonymize-text", False, str(e))
try:
styled = style_anonymize("I spearheaded the devops transformation and leveraged cutting-edge tech")
check("style-anonymize", "spearheaded" not in styled and "leveraged" not in styled,
f"result: {styled[:80]}")
except Exception as e:
check("style-anonymize", False, str(e))
detector = BiasDetector()
try:
if samples:
sample = next(iter(samples.values()))
mr = _make_match_result(
pid=sample.profile_id,
name=sample.personal.name if sample.personal else sample.profile_id,
title=sample.professional.current_title if sample.professional else "",
)
profiles_dict = {sample.profile_id: sample}
bias_result = detector.detect_bias(
matches=[mr], profiles=profiles_dict, bias_type="name",
)
check("bias-detector", isinstance(bias_result, dict), f"result={bias_result}")
except Exception as e:
check("bias-detector", False, str(e))
# ==========================================================
# STEP 11: Listwise Ranking
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 11: Plackett-Luce Listwise Ranking")
logger.info("=" * 60)
from src.ranking.listwise_ranker import PlackettLuceRanker
ranker = PlackettLuceRanker()
try:
# Build MatchResult objects from sample profiles
results_for_rank: list[MatchResult] = []
if samples:
for pid, prof in list(samples.items())[:10]:
mr = _make_match_result(
pid=pid,
name=prof.personal.name if prof.personal else pid,
title=prof.professional.current_title if prof.professional else "",
)
results_for_rank.append(mr)
if results_for_rank:
ranked = await ranker.arank(results_for_rank)
check("listwise-rank", ranked is not None and len(ranked) > 0,
f"count={len(ranked) if ranked else 0}")
except Exception as e:
check("listwise-rank", False, f"Exception: {e}")
traceback.print_exc()
# ==========================================================
# STEP 12: Evaluation Metrics
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 12: Evaluation Metrics")
logger.info("=" * 60)
from src.evaluation.metrics import mean_reciprocal_rank, ndcg_at_k, precision_at_k, recall_at_k
try:
retrieved = ["a", "b", "c", "d", "e"]
relevant = {"a", "c", "f"}
p_at_3 = precision_at_k(retrieved, relevant, 3)
r_at_5 = recall_at_k(retrieved, relevant, 5)
mrr = mean_reciprocal_rank(retrieved, relevant)
ndcg = ndcg_at_k(retrieved, relevant, 5)
check("eval-precision", 0.0 <= p_at_3 <= 1.0, f"p@3={p_at_3:.3f}")
check("eval-recall", 0.0 <= r_at_5 <= 1.0, f"r@5={r_at_5:.3f}")
check("eval-mrr", 0.0 <= mrr <= 1.0, f"mrr={mrr:.3f}")
check("eval-ndcg", 0.0 <= ndcg <= 1.0, f"ndcg={ndcg:.3f}")
except Exception as e:
check("eval-metrics", False, str(e))
# ==========================================================
# STEP 13: Hinglish query
# ==========================================================
logger.info("=" * 60)
logger.info("STEP 13: Hinglish Query")
logger.info("=" * 60)
from src.language.code_mixed import CodeMixedProcessor
cm_processor = CodeMixedProcessor()
try:
hinglish_query = "Mujhe ek senior Python developer chahiye with 5 years experience"
lang_result = cm_processor.detect_code_mixed(hinglish_query)
check("hinglish-detect", lang_result is not None, f"result={lang_result}")
except Exception as e:
check("hinglish-detect", False, str(e))
try:
hinglish_plan = await planner.plan(hinglish_query)
check("hinglish-plan", hinglish_plan is not None,
f"plan type={type(hinglish_plan).__name__ if hinglish_plan else 'None'}")
except Exception as e:
check("hinglish-plan", False, str(e))
# ==========================================================
# SUMMARY
# ==========================================================
logger.info("=" * 60)
logger.info("SUMMARY")
logger.info("=" * 60)
logger.info(f"Total steps: {len(successes) + len(errors)}")
logger.info(f"Passed: {len(successes)}")
logger.info(f"Failed: {len(errors)}")
if errors:
logger.error("FAILED STEPS:")
for e in errors:
logger.error(f" - {e}")
return 1
else:
logger.info("ALL STEPS PASSED!")
return 0
if __name__ == "__main__":
exit_code = asyncio.run(run_tests())
sys.exit(exit_code)
|