Create parser.py

app/utils/parser.py

@@ -0,0 +1,60 @@
+# app/utils/parser.py
+import re
+from typing import List, Dict
+
+def parse_questions(raw_text: str) -> List[Dict]:
+    """
+    Parse a model-generated block into list of question dicts:
+    Each block expected like:
+    1. Question ...
+    A) ...
+    B) ...
+    C) ...
+    D) ...
+    Answer: X
+    Explanation: ...
+    """
+    if not raw_text:
+        return []
+
+    # normalize line endings
+    text = raw_text.replace("\r\n", "\n").strip()
+    # split by question number (keep lines starting with digit + dot)
+    parts = re.split(r'\n(?=\d+\.)', text)
+    questions = []
+    for idx, part in enumerate(parts):
+        part = part.strip()
+        if not part:
+            continue
+        # number
+        num_match = re.match(r'^\s*(\d+)\.\s*(.*)', part, re.DOTALL)
+        if num_match:
+            # remove leading "1. " from the part content
+            content = part
+        else:
+            content = part
+
+        # extract question (up to A))
+        q_text = re.split(r'\nA\)', content, maxsplit=1)[0]
+        q_text = re.sub(r'^\s*\d+\.\s*', '', q_text).strip()
+
+        # extract options
+        opts = []
+        for letter in ['A','B','C','D']:
+            m = re.search(rf'{letter}\)\s*(.*?)(?=\n[A-D]\)|\nAnswer:|\nExplanation:|$)', content, re.DOTALL)
+            opts.append(m.group(1).strip() if m else "")
+
+        ans_m = re.search(r'Answer:\s*([A-D])', content, re.IGNORECASE)
+        answer = ans_m.group(1).upper() if ans_m else ""
+
+        expl_m = re.search(r'Explanation:\s*([\s\S]*)', content, re.IGNORECASE)
+        explanation = expl_m.group(1).strip() if expl_m else ""
+
+        questions.append({
+            "number": idx + 1,
+            "question": q_text,
+            "options": opts,
+            "answer": answer,
+            "explanation": explanation
+        })
+    return questions