style: auto-format with black + isort + prettier [skip ci-format-check]

apps/worker/video_processing/dedup.py

@@ -179,11 +179,11 @@ class VideoDeduplicator:
     """
 
     # ── Issue #1658: 校准后的常量 ──
-    PHASH_THRESHOLD = 8                # 从 10 收紧到 8
-    MATCH_RATIO_THRESHOLD = 0.7        # 至少 70% 帧匹配
-    DUPLICATE_THRESHOLD = 0.70         # 融合后相似度阈值
-    PHASH_WEIGHT = 0.7                 # pHash 权重
-    HISTOGRAM_WEIGHT = 0.3             # 直方图权重
+    PHASH_THRESHOLD = 8  # 从 10 收紧到 8
+    MATCH_RATIO_THRESHOLD = 0.7  # 至少 70% 帧匹配
+    DUPLICATE_THRESHOLD = 0.70  # 融合后相似度阈值
+    PHASH_WEIGHT = 0.7  # pHash 权重
+    HISTOGRAM_WEIGHT = 0.3  # 直方图权重
 
     def compute_fingerprint(self, video_path: str) -> VideoFingerprint:
         """Compute video fingerprint using MD5, pHash, and color histogram.
@@ -291,9 +291,7 @@ class VideoDeduplicator:
         a = hist_a[:min_len]
         b = hist_b[:min_len]
         # max(0, ...) 防御上游异常负值导致 sqrt domain error
-        return float(
-            sum(max(0.0, ai * bi) ** 0.5 for ai, bi in zip(a, b, strict=True))
-        )
+        return float(sum(max(0.0, ai * bi) ** 0.5 for ai, bi in zip(a, b, strict=True)))
 
     @staticmethod
     def _compute_histogram_similarity(
@@ -376,9 +374,7 @@ class VideoDeduplicator:
 
         # 直方图相似度：任一方无数据时统一返回 0.0（无法判定），避免不对称
         if existing_histograms and fingerprint.color_histograms:
-            hist_similarity = self._compute_histogram_similarity(
-                fingerprint.color_histograms, existing_histograms
-            )
+            hist_similarity = self._compute_histogram_similarity(fingerprint.color_histograms, existing_histograms)
         else:
             hist_similarity = 0.0
         combined_score = self.PHASH_WEIGHT * phash_similarity + self.HISTOGRAM_WEIGHT * hist_similarity
@@ -471,8 +467,7 @@ class VideoDeduplicator:
             if chunk_data:
                 existing_phashes = [c["phash_binary"] for c in chunk_data]
                 existing_histograms = [
-                    (ch if ch is not None else [])
-                    for ch in (c.get("color_histogram") for c in chunk_data)
+                    (ch if ch is not None else []) for ch in (c.get("color_histogram") for c in chunk_data)
                 ]
             else:
                 # 回退：从 JSON 字段读取（存量旧视频）
@@ -539,8 +534,7 @@ class VideoDeduplicator:
             if chunk_data:
                 existing_phashes = [c["phash_binary"] for c in chunk_data]
                 existing_histograms = [
-                    (ch if ch is not None else [])
-                    for ch in (c.get("color_histogram") for c in chunk_data)
+                    (ch if ch is not None else []) for ch in (c.get("color_histogram") for c in chunk_data)
                 ]
             else:
                 existing_phashes = ef.get("keyframe_phashes", [])
@@ -632,8 +626,7 @@ class VideoDeduplicator:
             if chunk_data:
                 existing_phashes = [c["phash_binary"] for c in chunk_data]
                 existing_histograms = [
-                    (ch if ch is not None else [])
-                    for ch in (c.get("color_histogram") for c in chunk_data)
+                    (ch if ch is not None else []) for ch in (c.get("color_histogram") for c in chunk_data)
                 ]
             else:
                 existing_phashes = ef.get("keyframe_phashes", [])

tests/unit/test_dedup_v2.py

@@ -96,13 +96,14 @@ sys.modules["packages.adapters.sqlalchemy_impl.models"] = _mock_module(
 sys.modules["packages.shared.config"] = _mock_module(get_shared_settings=MagicMock(return_value=MagicMock()))
 sys.modules["packages.shared.storage"] = _mock_module()
 
+import video_processing.dedup as _dedup_module  # noqa: E402
+
 # Import while mocks active
 from video_processing.dedup import (  # noqa: E402
     VideoDeduplicator,
     VideoFingerprint,
     hamming_distance,
 )
-import video_processing.dedup as _dedup_module  # noqa: E402
 
 # Restore sys.modules
 for _key in list(sys.modules.keys()):
@@ -179,6 +180,7 @@ class TestMedianVsMean:
     def test_median_resists_outlier(self):
         """距离 [3, 3, 3, 3, 30]：均值=8.4，中位数=3。"""
         import statistics
+
         distances = [3, 3, 3, 3, 30]
         mean_val = sum(distances) / len(distances)
         median_val = statistics.median(distances)
@@ -620,7 +622,14 @@ class TestComputeDuplicateRate:
         with patch.object(_dedup_module, "SQLAlchemyGeneratedVideoRepository", return_value=mock_repo):
             with patch.object(d, "_get_existing_chunks", return_value=[]):
                 # Patch the import inside the method
-                with patch.dict(sys.modules, {"packages.adapters.sqlalchemy_impl.models": _mock_module(GeneratedVideoModel=mock_generated_model)}):
+                with patch.dict(
+                    sys.modules,
+                    {
+                        "packages.adapters.sqlalchemy_impl.models": _mock_module(
+                            GeneratedVideoModel=mock_generated_model
+                        )
+                    },
+                ):
                     # Need to patch session.query to return mock_query
                     mock_session.query.return_value = mock_query
                     result = d.compute_duplicate_rate(fp, "proj_1", None, mock_session)
@@ -658,7 +667,14 @@ class TestComputeDuplicateRate:
         mock_generated_model = MagicMock()
         with patch.object(_dedup_module, "SQLAlchemyGeneratedVideoRepository", return_value=mock_repo):
             with patch.object(d, "_get_existing_chunks", return_value=[]):
-                with patch.dict(sys.modules, {"packages.adapters.sqlalchemy_impl.models": _mock_module(GeneratedVideoModel=mock_generated_model)}):
+                with patch.dict(
+                    sys.modules,
+                    {
+                        "packages.adapters.sqlalchemy_impl.models": _mock_module(
+                            GeneratedVideoModel=mock_generated_model
+                        )
+                    },
+                ):
                     result = d.compute_duplicate_rate(fp, "proj_1", None, mock_session)
 
         # 完全相同 → similarity≈1.0 → rate≈100
@@ -698,7 +714,14 @@ class TestComputeDuplicateRate:
         mock_generated_model = MagicMock()
         with patch.object(_dedup_module, "SQLAlchemyGeneratedVideoRepository", return_value=mock_repo):
             with patch.object(d, "_get_existing_chunks", return_value=[]):
-                with patch.dict(sys.modules, {"packages.adapters.sqlalchemy_impl.models": _mock_module(GeneratedVideoModel=mock_generated_model)}):
+                with patch.dict(
+                    sys.modules,
+                    {
+                        "packages.adapters.sqlalchemy_impl.models": _mock_module(
+                            GeneratedVideoModel=mock_generated_model
+                        )
+                    },
+                ):
                     result = d.compute_duplicate_rate(fp, "proj_1", None, mock_session)
 
         # phash 完全不同(0.0) + 直方图完全相同(1.0) → 0.7*0 + 0.3*1 = 0.3 → 30.0
@@ -720,7 +743,10 @@ class TestComputeDuplicateRate:
 
         mock_generated_model = MagicMock()
         with patch.object(_dedup_module, "SQLAlchemyGeneratedVideoRepository", return_value=mock_repo):
-            with patch.dict(sys.modules, {"packages.adapters.sqlalchemy_impl.models": _mock_module(GeneratedVideoModel=mock_generated_model)}):
+            with patch.dict(
+                sys.modules,
+                {"packages.adapters.sqlalchemy_impl.models": _mock_module(GeneratedVideoModel=mock_generated_model)},
+            ):
                 result = d.compute_duplicate_rate(fp, "proj_1", None, mock_session)
 
         assert result == 0.0