feat: add submitted collection, /similar and /submitted endpoints (Stage 4)

Made-with: Cursor

tests/test_vector_store_submitted.py

@@ -0,0 +1,212 @@
+"""
+Тесты для submitted-коллекции VectorStore.
+"""
+
+import numpy as np
+import pytest
+
+from app.storage.vector_store import VectorStore
+
+
+@pytest.fixture
+def vector_store(tmp_path):
+    """VectorStore с временным путём для submitted."""
+    return VectorStore(
+        vector_dim=4,
+        max_examples=10,
+        max_submitted=5,
+        storage_path=None,
+        submitted_path=str(tmp_path / "submitted.npz"),
+    )
+
+
+@pytest.fixture
+def sample_vector():
+    """Нормализованный вектор для тестов."""
+    v = np.array([1.0, 0.0, 0.0, 0.0], dtype=np.float32)
+    return v / np.linalg.norm(v)
+
+
+def test_add_submitted(vector_store, sample_vector):
+    """Добавление submitted-поста."""
+    added = vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="abc123",
+        created_at=1000,
+        post_id=42,
+        text="Test post",
+        rag_score=0.85,
+    )
+    assert added is True
+    assert vector_store.submitted_count == 1
+
+
+def test_add_submitted_duplicate(vector_store, sample_vector):
+    """Дубликат по хешу не добавляется."""
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="same_hash",
+        created_at=1000,
+        text="First",
+    )
+    added = vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="same_hash",
+        created_at=2000,
+        text="Second",
+    )
+    assert added is False
+    assert vector_store.submitted_count == 1
+
+
+def test_add_submitted_fifo(vector_store, sample_vector):
+    """При превышении max_submitted удаляется самый старый (FIFO)."""
+    for i in range(7):
+        v = np.array(
+            [float(i + 1), 0.0, 0.0, 0.0], dtype=np.float32
+        )  # i+1 чтобы избежать нулевого вектора
+        v = v / np.linalg.norm(v)
+        vector_store.add_submitted(
+            vector=v,
+            text_hash=f"hash_{i}",
+            created_at=1000 + i,
+            post_id=i,
+            text=f"Post {i}",
+        )
+    assert vector_store.submitted_count == 5  # max_submitted
+    # Должны остаться посты 2, 3, 4, 5, 6 (удалены 0, 1)
+    post_ids = vector_store._submitted_post_ids
+    assert 0 not in post_ids
+    assert 1 not in post_ids
+    assert 2 in post_ids
+
+
+def test_find_similar_submitted_empty(vector_store, sample_vector):
+    """Поиск в пустой коллекции возвращает пустой список."""
+    result = vector_store.find_similar_submitted(
+        vector=sample_vector,
+        threshold=0.5,
+        hours=24,
+    )
+    assert result == []
+
+
+def test_find_similar_submitted(vector_store, sample_vector):
+    """Поиск похожих постов с фильтром по времени и threshold."""
+    import time
+
+    now = int(time.time())
+    # Похожий вектор
+    similar_v = np.array([0.99, 0.01, 0.0, 0.0], dtype=np.float32)
+    similar_v = similar_v / np.linalg.norm(similar_v)
+    # Непохожий вектор
+    different_v = np.array([0.0, 1.0, 0.0, 0.0], dtype=np.float32)
+    different_v = different_v / np.linalg.norm(different_v)
+
+    vector_store.add_submitted(
+        vector=similar_v,
+        text_hash="similar",
+        created_at=now - 3600,  # 1 час назад
+        post_id=1,
+        text="Similar post",
+        rag_score=0.9,
+    )
+    vector_store.add_submitted(
+        vector=different_v,
+        text_hash="different",
+        created_at=now - 3600,
+        post_id=2,
+        text="Different post",
+        rag_score=0.5,
+    )
+
+    result = vector_store.find_similar_submitted(
+        vector=sample_vector,
+        threshold=0.9,
+        hours=24,
+    )
+    assert len(result) == 1
+    assert result[0]["post_id"] == 1
+    assert result[0]["text"] == "Similar post"
+    assert result[0]["similarity"] >= 0.9
+
+
+def test_find_similar_submitted_time_filter(vector_store, sample_vector):
+    """Фильтр по hours исключает старые посты."""
+    import time
+
+    now = int(time.time())
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="old",
+        created_at=now - 48 * 3600,  # 48 часов назад
+        post_id=1,
+        text="Old post",
+    )
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="recent",
+        created_at=now - 3600,  # 1 час назад
+        post_id=2,
+        text="Recent post",
+    )
+
+    result = vector_store.find_similar_submitted(
+        vector=sample_vector,
+        threshold=0.5,
+        hours=24,
+    )
+    assert len(result) == 1
+    assert result[0]["post_id"] == 2
+
+
+def test_submitted_persistence(vector_store, sample_vector, tmp_path):
+    """Сохранение и загрузка submitted-коллекции."""
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="persist",
+        created_at=12345,
+        post_id=999,
+        text="Persisted post",
+        rag_score=0.77,
+    )
+    vector_store.save_submitted_to_disk()
+
+    # Новый store загружает данные
+    store2 = VectorStore(
+        vector_dim=4,
+        max_submitted=5,
+        storage_path=None,
+        submitted_path=str(tmp_path / "submitted.npz"),
+    )
+    assert store2.submitted_count == 1
+    assert store2._submitted_post_ids[0] == 999
+    assert store2._submitted_texts[0] == "Persisted post"
+    assert store2._submitted_rag_scores[0] == 0.77
+
+
+def test_get_stats_includes_submitted(vector_store, sample_vector):
+    """get_stats включает submitted_count и max_submitted."""
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="stat",
+        created_at=1000,
+        text="For stats",
+    )
+    stats = vector_store.get_stats()
+    assert "submitted_count" in stats
+    assert stats["submitted_count"] == 1
+    assert "max_submitted" in stats
+    assert stats["max_submitted"] == 5
+
+
+def test_clear_submitted(vector_store, sample_vector):
+    """clear() очищает submitted-коллекцию."""
+    vector_store.add_submitted(
+        vector=sample_vector,
+        text_hash="clear",
+        created_at=1000,
+        text="To clear",
+    )
+    vector_store.clear()
+    assert vector_store.submitted_count == 0