Create parsing.py

pipelines/parsing.py

@@ -0,0 +1,41 @@
+import re
+from typing import Dict, List
+
+SECTION_HEADERS = [
+    ("work_experience", ["職歴", "職務経歴", "業務経験", "Work Experience", "Experience"]),
+    ("education", ["学歴", "Education"]),
+    ("certifications", ["資格", "認定", "Certificates", "Certifications"]),
+    ("skills", ["スキル", "Skills"]),
+]
+
+
+def normalize_resume(sections_dict: Dict[str, str]) -> Dict[str, List[Dict]]:
+    # sections_dict は {work_experience, education, certifications, skills, raw_sections?}
+    work_items = []
+    raw_work = sections_dict.get("work_experience", "") or sections_dict.get("work_experience_raw", "")
+    for line in raw_work.splitlines():
+        m = re.search(r"(\d{4}[./年]\d{1,2})\s*[-〜~]\s*(\d{4}[./年]?\d{0,2}|現在|至今)?", line)
+        if m:
+            work_items.append({"period": m.group(0), "text": line.strip()})
+
+    edu_items = [{"text": l.strip()} for l in (sections_dict.get("education", "") or sections_dict.get("education_raw", "")).splitlines() if l.strip()]
+    cert_items = [{"text": l.strip()} for l in (sections_dict.get("certifications", "") or sections_dict.get("certifications_raw", "")).splitlines() if l.strip()]
+
+    # skills はカンマ/改行区切り or list を許容
+    skills_raw = sections_dict.get("skills", "")
+    if isinstance(skills_raw, list):
+        skill_items = skills_raw
+    else:
+        skill_items = [s.strip() for s in re.split(r"[、,\n]\s*", skills_raw) if s.strip()]
+
+    return {
+        "work_experience": work_items,
+        "education": edu_items,
+        "certifications": cert_items,
+        "skills": skill_items,
+        "raw_sections": {
+            "work_experience": raw_work,
+            "education": sections_dict.get("education", "") or sections_dict.get("education_raw", ""),
+            "certifications": sections_dict.get("certifications", "") or sections_dict.get("certifications_raw", ""),
+        },
+    }