fix: compare INTEGER answers by value, not string

Numerical answers (INTEGER/INTEGER_2) were scored with exact string
equality, so equal-but-differently-formatted values were marked wrong
(e.g. '9.8' != '9.80', '5' != '5.00'). Add integer_answers_match() which
falls back to a float comparison with a tight tolerance, used in both the
JEE Main and JEE Advanced INTEGER scoring paths. Range answers are
unaffected (they already use is_within_range/float).

Adds 4 tests covering the trailing-zero cases plus a false-positive guard
(9.79 != 9.80).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/evaluation.py

@@ -1,4 +1,5 @@
 import logging
+import math
 import re
 from typing import List, Optional, Union, Dict, Any
 
@@ -33,6 +34,24 @@ def is_within_range(predicted_value_str: str, lower_bound_str: str, upper_bound_
     return lower_bound <= predicted_value <= upper_bound
 
 
+def integer_answers_match(predicted_value_str: str, truth_str: str) -> bool:
+    """
+    Checks whether a predicted numerical answer equals an exact ground-truth value.
+
+    Falls back to a numeric comparison so that differently-formatted but equal
+    values match (e.g. '9.8' == '9.80', '5' == '5.00'). Exact string equality is
+    tried first to handle any non-numeric tokens gracefully.
+    """
+    if predicted_value_str == truth_str:
+        return True
+    try:
+        # abs_tol is tiny: real JEE numerical keys differ by >= 0.01, so this only
+        # collapses formatting differences (trailing zeros), never distinct values.
+        return math.isclose(float(predicted_value_str), float(truth_str), rel_tol=0.0, abs_tol=1e-9)
+    except (ValueError, TypeError):
+        return False
+
+
 def calculate_single_question_score_details(result_item: Dict[str, Any]) -> Dict[str, Any]:
     """
     Calculates marks_awarded and evaluation_status for a single question result.
@@ -121,7 +140,7 @@ def calculate_single_question_score_details(result_item: Dict[str, Any]) -> Dict
             is_correct = False
             if len(pred_set) == 1:
                 predicted_answer_str = list(pred_set)[0]
-                if predicted_answer_str in truth_processed: # Check against single string truths
+                if any(integer_answers_match(predicted_answer_str, t) for t in truth_processed): # Check against single string truths
                     is_correct = True
             
             if is_correct:
@@ -144,7 +163,7 @@ def calculate_single_question_score_details(result_item: Dict[str, Any]) -> Dict
                                 is_correct = True
                                 break # Found a matching range, no need to check others
                         elif isinstance(gt_entry, str): # This is an exact integer match
-                            if predicted_answer_str == gt_entry: # gt_entry is already uppercase
+                            if integer_answers_match(predicted_answer_str, gt_entry): # gt_entry is already uppercase
                                 is_correct = True
                                 break # Found an exact match, no need to check others
                 

tests/test_evaluation.py

@@ -182,6 +182,49 @@ class TestSingleQuestionScoring:
         assert result["marks_awarded"] == 0
         assert result["evaluation_status"] == "incorrect"
 
+    def test_jee_advanced_integer_trailing_zero_equivalent(self):
+        """'9.8' must match the key '9.80' — same numeric value, different formatting."""
+        result = calculate_single_question_score_details({
+            "question_id": "JA_FMT", "exam_name": "JEE_ADVANCED",
+            "question_type": "INTEGER",
+            "ground_truth": ["9.80"], "predicted_answer": ["9.8"],
+            "api_call_successful": True,
+        })
+        assert result["marks_awarded"] == 4
+        assert result["evaluation_status"] == "correct"
+
+    def test_jee_advanced_integer_half_equivalent(self):
+        """'0.5' must match the key '0.50'."""
+        result = calculate_single_question_score_details({
+            "question_id": "JA_FMT2", "exam_name": "JEE_ADVANCED",
+            "question_type": "INTEGER",
+            "ground_truth": ["0.50"], "predicted_answer": ["0.5"],
+            "api_call_successful": True,
+        })
+        assert result["marks_awarded"] == 4
+        assert result["evaluation_status"] == "correct"
+
+    def test_jee_advanced_integer_numeric_mismatch_still_wrong(self):
+        """A genuinely different value must still score 0 (no false positives)."""
+        result = calculate_single_question_score_details({
+            "question_id": "JA_FMT3", "exam_name": "JEE_ADVANCED",
+            "question_type": "INTEGER",
+            "ground_truth": ["9.80"], "predicted_answer": ["9.79"],
+            "api_call_successful": True,
+        })
+        assert result["marks_awarded"] == 0
+        assert result["evaluation_status"] == "incorrect"
+
+    def test_jee_main_integer_trailing_zero_equivalent(self):
+        """Numeric-equivalence fix also applies to JEE Main INTEGER scoring."""
+        result = calculate_single_question_score_details({
+            "question_id": "JM_FMT", "exam_name": "JEE_MAIN",
+            "question_type": "INTEGER",
+            "ground_truth": ["5.00"], "predicted_answer": ["5"],
+            "api_call_successful": True,
+        })
+        assert result["marks_awarded"] == 4
+
     def test_unexpected_prediction_type_no_penalty(self):
         """Unexpected pred type is a code/data bug, not a deliberate wrong answer — no penalty."""
         result = calculate_single_question_score_details({