Update merge_questions.py

merge_questions.py

@@ -30,12 +30,12 @@ def safe_load(path):
         return None
 
 
-def merge_all():
-    os.makedirs(DATA_DIR, exist_ok=True)
+def merge_all(data_dir=DATA_DIR, output_file=OUTPUT_FILE):
+    os.makedirs(data_dir, exist_ok=True)
     merged = []
     next_id = 1000
 
-    p = Path(DATA_DIR)
+    p = Path(data_dir)
     files = sorted(p.glob("spm_*_*.json"))
 
     for file in files:
@@ -78,18 +78,18 @@ def merge_all():
                 "year": int(year)
             }
 
-            # attach correct_answer if scheme provides it (scheme may be {"1": "A"} or {"1": {"correct_answer": "x"}})
             correct_answer = None
             val = scheme.get(str(idx)) if isinstance(scheme, dict) else None
             if isinstance(val, str):
-                # if it's a letter like "A", map to choice text
                 v = val.strip()
                 if len(v) == 1 and merged_q["choices"]:
                     pos = ord(v.upper()) - ord("A")
                     if 0 <= pos < len(merged_q["choices"]):
                         correct_answer = merged_q["choices"][pos]
+                    else:
+                        correct_answer = None
                 else:
-                    correct_answer = v
+                    correct_answer = v if v else None
             elif isinstance(val, dict):
                 correct_answer = val.get("correct_answer")
             else:
@@ -99,11 +99,10 @@ def merge_all():
             merged.append(merged_q)
             next_id += 1
 
-    # overwrite output file safely
-    with open(OUTPUT_FILE, "w", encoding="utf-8") as f:
+    with open(output_file, "w", encoding="utf-8") as f:
         json.dump(merged, f, indent=2, ensure_ascii=False)
 
-    print(f"✅ Merged {len(merged)} questions into {OUTPUT_FILE}")
+    print(f"✅ Merged {len(merged)} questions into {output_file}")
 
 
 if __name__ == "__main__":
@@ -115,3 +114,4 @@ if __name__ == "__main__":
 
 
 
+