update_readme.py UPDATE

Author: gh-develop

.github/scripts/update_readme.py

@@ -1,6 +1,13 @@
 #!/usr/bin/env python3
 # .github/scripts/update_readme.py
-# Debug/verbose version — 폴더/파일/매칭/entries 정보 전부 출력 & 로그파일 작성
+# Debug + Robust + Conflict-resolution version
+# - verbose logging (update_readme_debug.log)
+# - certifi verify for requests
+# - normalize unicode/spaces
+# - fetch from solved.ac & programmers
+# - parse existing README table
+# - detect files bj_/pg_ and prefer file-based platform (remove conflicting entry)
+# - preserve pre/post content and write README
 
 import os, re, time, unicodedata, requests, sys
 from bs4 import BeautifulSoup
@@ -81,7 +88,7 @@ def fetch_programmers_info(pid):
         log("[fetch_programmers_info] exception:", e)
     return {"title":"UNKNOWN","level":"Lv.?","link":url}
 
-# find table block (same robust logic)
+# find table block (robust)
 def find_table_block(lines):
     header_idx = None
     for i, line in enumerate(lines):
@@ -107,13 +114,24 @@ def find_table_block(lines):
     post = lines[rows_end:]
     return pre, header, rows, post
 
+# parse a table row -> normalized platform, number, title, level, status
 def parse_row_line(line):
     cols = [c.strip() for c in line.strip().split("|")[1:-1]]
     status = cols[1] if len(cols) > 1 else "⬜"
-    platform = cols[2] if len(cols) > 2 else ""
+    raw_platform = cols[2] if len(cols) > 2 else ""
+    raw_platform_l = raw_platform.lower()
+    # normalize platform to exact tokens used internally
+    if "백준" in raw_platform or "acmicpc" in raw_platform_l or "baek" in raw_platform_l:
+        platform = "백준"
+    elif "프로그래머스" in raw_platform or "programmers" in raw_platform_l or "pg" in raw_platform_l:
+        platform = "프로그래머스"
+    else:
+        platform = raw_platform.strip() or "UNKNOWN"
+
     number_cell = cols[3] if len(cols) > 3 else ""
     title_cell = cols[4] if len(cols) > 4 else ""
     level = cols[5] if len(cols) > 5 else ""
+
     num = None; num_link = ""
     m = re.search(r"\[(\d+)\]\(([^)]+)\)", number_cell)
     if m:
@@ -122,15 +140,18 @@ def parse_row_line(line):
         m2 = re.search(r"(\d+)", number_cell)
         if m2:
             num = m2.group(1)
+
     title = ""
     m = re.search(r"\[(.*?)\]\(([^)]+)\)", title_cell)
     if m:
         title = m.group(1)
     else:
         title = title_cell
+
     return {"platform": platform, "number": num, "number_link": num_link, "title": title, "level": level, "status": status, "raw": line}
 
 def platform_rank(p): return 0 if p.strip()=="백준" else 1
+
 def level_rank(p, lvl):
     lvl = normalize_spaces(lvl or "")
     if p.strip()=="백준":
@@ -145,7 +166,7 @@ def level_rank(p, lvl):
         except: pass
     return 999
 
-# core: process_directory with verbose logging
+# core: process_directory with verbose logging + conflict resolution
 def process_directory(dir_path):
     readme_path = os.path.join(dir_path, "README.md")
     log(f"[start] folder={dir_path} (readme_exists={os.path.isfile(readme_path)})")
@@ -199,28 +220,38 @@ def process_directory(dir_path):
             }
     log(f"[existing_entries] count={len(entries)}")
 
-    # incorporate matched files
+    # incorporate matched files (file-based additions) and collect found sets
+    found_bj = set()
+    found_pg = set()
     for item in matched:
         fn = item["file"]; name,_ = os.path.splitext(fn)
         if item["bj"]:
             pid = re.search(r"bj_(\d+)", name).group(1)
             status = "❌" if "_FAIL" in name else "✅"
             key = ("백준", pid)
-            if key in entries:
-                entries[key]["status"] = status
-            else:
-                entries[key] = {"platform":"백준","number":pid,"number_link":"","title":"","level":"","status":status}
+            entries[key] = {"platform":"백준","number":pid,"number_link":"","title":"","level":"","status":status}
+            found_bj.add(pid)
         if item["pg"]:
             pid = re.search(r"pg_(\d+)", name).group(1)
             status = "❌" if "_FAIL" in name else "✅"
             key = ("프로그래머스", pid)
-            if key in entries:
-                entries[key]["status"] = status
-            else:
-                entries[key] = {"platform":"프로그래머스","number":pid,"number_link":"","title":"","level":"","status":status}
+            entries[key] = {"platform":"프로그래머스","number":pid,"number_link":"","title":"","level":"","status":status}
+            found_pg.add(pid)
 
     log(f"[after_scan_entries] count={len(entries)} keys={list(entries.keys())}")
 
+    # --- CONFLICT RESOLUTION: prefer file-based platform
+    # If both platforms exist for same number, remove the one without a corresponding file.
+    # i.e., if pg file exists -> remove ('백준', num) entry; if bj file exists -> remove ('프로그래머스', num) entry
+    for num in list(found_pg):
+        if ("백준", num) in entries and ("프로그래머스", num) in entries and num in found_pg:
+            log(f"[conflict] removing 백준 entry for {num} because pg file exists")
+            entries.pop(("백준", num), None)
+    for num in list(found_bj):
+        if ("프로그래머스", num) in entries and ("백준", num) in entries and num in found_bj:
+            log(f"[conflict] removing 프로그래머스 entry for {num} because bj file exists")
+            entries.pop(("프로그래머스", num), None)
+
     # if no entries and no existing readme -> nothing to do
     if not entries and not os.path.isfile(readme_path):
         log(f"[skip] no entries and no readme -> nothing to do for {dir_path}")
@@ -292,22 +323,6 @@ def process_directory(dir_path):
         log(f"[ERROR] writing README {readme_path}: {e}")
         return False
 
-def platform_rank(p): return 0 if p.strip()=="백준" else 1
-
-def level_rank(p, lvl):
-    lvl = normalize_spaces(lvl or "")
-    if p.strip()=="백준":
-        return BOJ_LEVEL_RANK.get(lvl, 999)
-    m = re.search(r"(?:Lv\.?|Level)?\s*\.?\s*(\d+)", lvl, re.IGNORECASE)
-    if m:
-        try: return int(m.group(1))
-        except: pass
-    m2 = re.search(r"(\d+)", lvl)
-    if m2:
-        try: return int(m2.group(1))
-        except: pass
-    return 999
-
 def main():
     # clear log
     try: