Eng-Musa commited on
Commit
ad0333f
Β·
1 Parent(s): c1c2a54

improve calculator

Browse files
Files changed (1) hide show
  1. services/job_matcher.py +17 -53
services/job_matcher.py CHANGED
@@ -11,15 +11,7 @@ from services.taxonomy_client import (
11
  DEFAULT_SKILL_ALIASES,
12
  DEFAULT_CATEGORY_SKILLS,
13
  SENIORITY_ORDER,
14
- _extract_skills,
15
- _infer_category,
16
- _extract_years,
17
- _detect_seniority,
18
- refine_seniority_with_years,
19
- _extract_cv_title,
20
- compute_years_from_experience,
21
  )
22
- from services.cv_chunker import _parse_cv_sections, _parse_experience
23
 
24
 
25
  # ---------------------------------------------------------------------------
@@ -215,25 +207,26 @@ class JobMatcher:
215
  self,
216
  job_json: Dict[str, Any],
217
  cv_markdown: Optional[str] = None,
218
- section_embeddings: Optional[Dict[str, Any]] = None,
219
  cv_title: Optional[str] = None,
220
  cv_years: Optional[int] = None,
221
  cv_seniority: Optional[str] = None,
222
- cv_skills: Optional[List[str]] = None,
 
 
223
  cv_category: Optional[str] = None,
224
  ) -> Dict[str, Any]:
225
  """
226
  Predict the match between one job and one CV.
227
- Pass ``section_embeddings`` from ``/process-cv`` to skip re-encoding.
228
  """
229
  cv_data = self._parse_cv(
230
- cv_markdown=cv_markdown,
231
- section_embeddings=section_embeddings,
232
- cv_title=cv_title,
233
- cv_years=cv_years,
234
- cv_seniority=cv_seniority,
235
- cv_skills=cv_skills,
236
- cv_category=cv_category,
 
237
  )
238
  return self._predict_single(job_json, cv_data)
239
 
@@ -366,7 +359,7 @@ class JobMatcher:
366
  semantic_matches = self._semantic_skill_matches(
367
  missing_technical,
368
  individual_skills_vecs,
369
- threshold=0.92
370
  )
371
  if semantic_matches:
372
  matched_technical = sorted(set(matched_technical) | semantic_matches)
@@ -377,13 +370,13 @@ class JobMatcher:
377
 
378
  extra = sorted(cv_skills - job_technical - job_soft)[:8]
379
 
380
- tech_denom = min(len(job_technical), 15) if job_technical else 1
381
- soft_denom = min(len(job_soft), 4) if job_soft else 1
382
-
383
  tech_score = min(1.0, len(matched_technical) / tech_denom) if job_technical else 0.50
384
  soft_score = min(1.0, len(matched_soft) / soft_denom) if job_soft else 0.50
385
 
386
- skill = min(1.0, (tech_score * 0.70) + (soft_score * 0.30))
387
 
388
  # ── Component scores ─────────────────────────────────────────────
389
  cv_cat_lower = cv_category.strip().lower() if cv_category else ""
@@ -399,12 +392,9 @@ class JobMatcher:
399
  experience = _experience_fit(cv_years, job.min_years_experience)
400
 
401
  # ── Weighted aggregation ─────────────────────────────────────────
402
- # Discount semantic score by category score (prevents broad semantic clustering across domains)
403
- semantic_weighted = semantic * cat_score
404
-
405
  base_final = (
406
  0.42 * skill
407
- + 0.23 * semantic_weighted
408
  + 0.12 * title
409
  + 0.11 * cat_score
410
  + 0.07 * seniority
@@ -540,29 +530,3 @@ class JobMatcher:
540
  cv_vec = self._embed_title(cv_title)
541
  job_vec = self._embed_title(job_title)
542
  return max(0.2, min(1.0, _cosine_similarity(cv_vec, job_vec)))
543
-
544
- def _category_distance_modifier(self, cv_category: str, job_category: str) -> float:
545
- """
546
- Calculate a score modifier between -10 and +10 based on how far apart
547
- the CV category and Job category are semantically.
548
- """
549
- if not cv_category or not job_category:
550
- return 0.0
551
-
552
- cv_cat_clean = cv_category.strip().lower()
553
- job_cat_clean = job_category.strip().lower()
554
-
555
- if cv_cat_clean == job_cat_clean:
556
- return 10.0
557
-
558
- cv_vec = self._embed(f"industry sector: {cv_cat_clean}")
559
- job_vec = self._embed(f"industry sector: {job_cat_clean}")
560
-
561
- similarity = _cosine_similarity(cv_vec, job_vec)
562
-
563
- # Map similarity [0, 1] to [-10, 10]
564
- # Assuming similarity < 0.5 is distant (-10 to 0)
565
- # Assuming similarity > 0.5 is close (0 to 10)
566
- modifier = (similarity - 0.5) * 20.0
567
-
568
- return max(-10.0, min(10.0, modifier))
 
11
  DEFAULT_SKILL_ALIASES,
12
  DEFAULT_CATEGORY_SKILLS,
13
  SENIORITY_ORDER,
 
 
 
 
 
 
 
14
  )
 
15
 
16
 
17
  # ---------------------------------------------------------------------------
 
207
  self,
208
  job_json: Dict[str, Any],
209
  cv_markdown: Optional[str] = None,
 
210
  cv_title: Optional[str] = None,
211
  cv_years: Optional[int] = None,
212
  cv_seniority: Optional[str] = None,
213
+ cv_skills_canonical: Optional[List[str]] = None,
214
+ cv_skills_technical: Optional[List[str]] = None,
215
+ cv_skills_soft: Optional[List[str]] = None,
216
  cv_category: Optional[str] = None,
217
  ) -> Dict[str, Any]:
218
  """
219
  Predict the match between one job and one CV.
 
220
  """
221
  cv_data = self._parse_cv(
222
+ cv_markdown=cv_markdown or "",
223
+ cv_title=cv_title or "Unknown",
224
+ cv_years=cv_years or 0,
225
+ cv_seniority=cv_seniority or "mid",
226
+ cv_skills_canonical=cv_skills_canonical or [],
227
+ cv_skills_technical=cv_skills_technical or [],
228
+ cv_skills_soft=cv_skills_soft or [],
229
+ cv_category=cv_category or "Other",
230
  )
231
  return self._predict_single(job_json, cv_data)
232
 
 
359
  semantic_matches = self._semantic_skill_matches(
360
  missing_technical,
361
  individual_skills_vecs,
362
+ threshold=0.88
363
  )
364
  if semantic_matches:
365
  matched_technical = sorted(set(matched_technical) | semantic_matches)
 
370
 
371
  extra = sorted(cv_skills - job_technical - job_soft)[:8]
372
 
373
+ tech_denom = min(len(job_technical), 12) if job_technical else 1
374
+ soft_denom = min(len(job_soft), 3) if job_soft else 1
375
+
376
  tech_score = min(1.0, len(matched_technical) / tech_denom) if job_technical else 0.50
377
  soft_score = min(1.0, len(matched_soft) / soft_denom) if job_soft else 0.50
378
 
379
+ skill = min(1.0, (tech_score * 0.80) + (soft_score * 0.20))
380
 
381
  # ── Component scores ─────────────────────────────────────────────
382
  cv_cat_lower = cv_category.strip().lower() if cv_category else ""
 
392
  experience = _experience_fit(cv_years, job.min_years_experience)
393
 
394
  # ── Weighted aggregation ─────────────────────────────────────────
 
 
 
395
  base_final = (
396
  0.42 * skill
397
+ + 0.23 * semantic
398
  + 0.12 * title
399
  + 0.11 * cat_score
400
  + 0.07 * seniority
 
530
  cv_vec = self._embed_title(cv_title)
531
  job_vec = self._embed_title(job_title)
532
  return max(0.2, min(1.0, _cosine_similarity(cv_vec, job_vec)))