fix bug in ItemMatcher colation (#70)

* fix bug in ItemMatcher colation

* deep checkout for workflow actions

.github/workflows/docs_build.yml

@@ -25,6 +25,8 @@ jobs:
     steps:
       - name: Checkout 🛒
         uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
       - name: ⚙️ Setup - Python 🐍
         uses: actions/setup-python@v5

.github/workflows/validate.yml

@@ -58,6 +58,8 @@ jobs:
     steps:
       - name: 🛒 Checkout
         uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
       - name: ⚙️ Setup - Python 🐍
         uses: actions/setup-python@v5

docs/release-history.rst

@@ -31,6 +31,14 @@ Release History
 The format is based on `Keep a Changelog <https://keepachangelog.com/en>`_,
 and this project adheres to `Semantic Versioning <https://semver.org/spec/v2.0.0.html>`_
 
+0.9.1
+=====
+
+Fixed
+-----
+
+* Bug in :py:meth:`.ItemMatcher.match` where operations always returned the last item in the given list of ``results``
+
 
 0.9.0
 =====

musify/processors/match.py

@@ -304,6 +304,7 @@ def match[T: MusifyObject](
             else:
                 match_on_filtered.add(match_field)
 
+        min_score = limit_value(min_score, floor=0.01, ceil=1.0)
         max_score = limit_value(max_score, floor=0.01, ceil=1.0)
         self._log_algorithm(source=source, extra=[f"max_score={max_score}"])
 
@@ -316,34 +317,20 @@ def match[T: MusifyObject](
                 allow_karaoke=allow_karaoke
             )
 
-        result = None
-        best_score = 0
-
-        def sum_nested_scores(futures: list[list[Future[float]]]) -> float:
-            """Sum the scores from a given list of nested Futures"""
-            scores_summed = [sum(score.result() for score in nested) / len(nested) for nested in futures]
-            return sum(scores_summed) / len(scores_summed)
-
-        for result, result_scores in scores:
-            result_scores = [
-                sum_nested_scores(score) if isinstance(score, list) else score.result()
-                for score in result_scores.values()
-            ]
-            best_score = sum(result_scores) / len(result_scores)
-            if best_score > max_score:
-                break
+        result, score = self._get_match_from_scores(scores, max_score=max_score)
 
-        min_score = limit_value(min_score, floor=0.01, ceil=1.0)
-        if best_score > min_score:
+        if result is not None and score > min_score:
             extra = [
-                f"best score: {'%.2f' % round(best_score, 2)} > {'%.2f' % round(min_score, 2)}"
-                if best_score < max_score else
-                f"max score reached: {'%.2f' % round(best_score, 2)} > {'%.2f' % round(max_score, 2)}"
+                f"best score: {score:.2f} > {min_score:.2f}"
+                if score < max_score else
+                f"max score reached: {score:.2f} > {max_score:.2f}"
             ]
             self._log_match(source=source, result=result, extra=extra)
             return result
         else:
-            self._log_test(source=source, result=result, test=best_score, extra=[f"NO MATCH: {best_score}<{min_score}"])
+            self._log_test(
+                source=source, result=result, test=score, extra=[f"NO MATCH: {score:.2f}<{min_score:.2f}"]
+            )
 
     def _score[T: MusifyObject](
             self,
@@ -432,6 +419,34 @@ def _get_scores[T: MusifyObject](
 
         return scores
 
+    def _get_match_from_scores[T: MusifyObject](
+            self,
+            scores: Iterable[tuple[T, dict[TagField, Future[float] | list[list[Future[float]]]]]],
+            max_score: float,
+    ) -> tuple[T | None, float]:
+        best_result = None
+        best_score = 0
+
+        def sum_nested_scores(futures: list[list[Future[float]]]) -> float:
+            """Sum the scores from a given list of nested Futures"""
+            scores_summed = [sum(score.result() for score in nested) / len(nested) for nested in futures]
+            return sum(scores_summed) / len(scores_summed)
+
+        for result, result_scores in scores:
+            result_scores = [
+                sum_nested_scores(score) if isinstance(score, list) else score.result()
+                for score in result_scores.values()
+            ]
+            score = sum(result_scores) / len(result_scores)
+            if score > best_score:
+                best_score = score
+                best_result = result
+
+            if best_score > max_score:
+                break
+
+        return best_result, best_score
+
     def as_dict(self) -> dict[str, Any]:
         return {
             "clean_tags": {

tests/processors/test_match.py

@@ -184,7 +184,7 @@ def test_match_all(self, matcher: ItemMatcher, track1: LocalTrack, track2: Local
         track3.artist = f"artist{sep}nope{sep}other"
         track3.year = 2015
         assert matcher.match(track1, [track2, track3], min_score=0.2, max_score=0.8) == track3
-        assert matcher(track1, [track2, track3], min_score=0.2, max_score=0.8) == track3
+        assert matcher(track1, [track3, track2], min_score=0.2, max_score=0.8) == track3
 
         # track4 score is above max_score causing an early stop
         track4 = random_track()
@@ -193,7 +193,7 @@ def test_match_all(self, matcher: ItemMatcher, track1: LocalTrack, track2: Local
         track4.album = "album"
         track4._reader.file.info.length = 100
         track4.year = 2015
-        assert matcher.match(track1, [track2, track4, track3], min_score=0.2, max_score=0.8) == track4
+        assert matcher.match(track1, [track4, track2, track3], min_score=0.2, max_score=0.8) == track4
         assert matcher(track1, [track2, track4, track3], min_score=0.2, max_score=0.8) == track4
 
     def test_allows_karaoke(self, matcher: ItemMatcher, track1: LocalTrack, track2: LocalTrack):