Task 5.23: Deduplication / Merge — ContactMergeHistory model, dedup service, API routes (duplicates/merge/merge-history), DedupDialog frontend, i18n, 5 tests

2026-07-23 23:58:45 +02:00
parent a914280a4e
commit d54a87cf84
11 changed files with 1051 additions and 3 deletions
@@ -0,0 +1,216 @@
+"""Deduplication / merge tests — Task 5.23."""
+
+from __future__ import annotations
+
+import pytest
+from httpx import AsyncClient
+
+from tests.conftest import ORIGIN_HEADER, login_client, seed_tenant_and_users
+
+
+async def _get_csrf_token(client: AsyncClient) -> str:
+    """Extract CSRF token from login response."""
+    # The login already happened, but we need the token from the response.
+    # Re-login to get a fresh token (or extract from the last response).
+    resp = await client.post(
+        "/api/v1/auth/login",
+        json={"email": "admin@tenanta.com", "password": "TestPass123!"},
+        headers=ORIGIN_HEADER,
+    )
+    return resp.json().get("csrf_token", "")
+
+
+def _csrf_headers(token: str) -> dict:
+    """Return headers with CSRF token."""
+    headers = dict(ORIGIN_HEADER)
+    headers["X-CSRF-Token"] = token
+    return headers
+
+
+@pytest.mark.asyncio
+class TestFindDuplicates:
+    """POST /api/v1/contacts/duplicates — find duplicate contacts."""
+
+    async def test_find_duplicates_by_email(self, client: AsyncClient, db_session):
+        """Two contacts with same email should be detected as duplicates."""
+        await seed_tenant_and_users(db_session)
+        await login_client(client, "admin@tenanta.com")
+        csrf = await _get_csrf_token(client)
+        hdrs = _csrf_headers(csrf)
+
+        # Create two contacts with same email
+        resp1 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "John", "surname": "Doe", "email_1": "john@example.com"},
+            headers=hdrs,
+        )
+        assert resp1.status_code == 201, f"Create 1 failed: {resp1.text}"
+
+        resp2 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Johnny", "surname": "Doe", "email_1": "john@example.com"},
+            headers=hdrs,
+        )
+        assert resp2.status_code == 201, f"Create 2 failed: {resp2.text}"
+
+        # Find duplicates
+        resp = await client.post(
+            "/api/v1/contacts/duplicates",
+            json={"threshold": 0.5, "limit": 50},
+            headers=hdrs,
+        )
+        assert resp.status_code == 200
+        data = resp.json()
+        assert isinstance(data, list)
+        assert len(data) >= 1
+        pair = data[0]
+        assert "source_contact" in pair
+        assert "target_contact" in pair
+        assert "similarity_score" in pair
+        assert "match_reasons" in pair
+        assert "email_match" in pair["match_reasons"]
+
+    async def test_find_duplicates_empty(self, client: AsyncClient, db_session):
+        """No duplicates when contacts are unique."""
+        await seed_tenant_and_users(db_session)
+        await login_client(client, "admin@tenanta.com")
+        csrf = await _get_csrf_token(client)
+        hdrs = _csrf_headers(csrf)
+
+        resp = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Unique", "surname": "Person", "email_1": "unique@example.com"},
+            headers=hdrs,
+        )
+        assert resp.status_code == 201, f"Create failed: {resp.text}"
+
+        resp = await client.post(
+            "/api/v1/contacts/duplicates",
+            json={"threshold": 0.7, "limit": 50},
+            headers=hdrs,
+        )
+        assert resp.status_code == 200
+        data = resp.json()
+        assert isinstance(data, list)
+        assert len(data) == 0
+
+
+@pytest.mark.asyncio
+class TestMergeContacts:
+    """POST /api/v1/contacts/merge — merge two contacts."""
+
+    async def test_merge_contacts_success(self, client: AsyncClient, db_session):
+        """Merge source into target, source should be soft-deleted."""
+        await seed_tenant_and_users(db_session)
+        await login_client(client, "admin@tenanta.com")
+        csrf = await _get_csrf_token(client)
+        hdrs = _csrf_headers(csrf)
+
+        # Create source contact with phone
+        resp1 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Source", "surname": "Test", "email_1": "source@example.com", "phone_1": "+49123456789"},
+            headers=hdrs,
+        )
+        assert resp1.status_code == 201, f"Create source failed: {resp1.text}"
+        source_id = resp1.json()["id"]
+
+        # Create target contact without phone
+        resp2 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Target", "surname": "Test", "email_1": "target@example.com"},
+            headers=hdrs,
+        )
+        assert resp2.status_code == 201, f"Create target failed: {resp2.text}"
+        target_id = resp2.json()["id"]
+
+        # Merge
+        resp = await client.post(
+            "/api/v1/contacts/merge",
+            json={"source_contact_id": source_id, "target_contact_id": target_id},
+            headers=hdrs,
+        )
+        assert resp.status_code == 200, f"Merge failed: {resp.text}"
+        data = resp.json()
+        assert data["source_contact_id"] == source_id
+        assert data["target_contact_id"] == target_id
+        assert "merge_id" in data
+        assert "merged_fields" in data
+        # Phone should have been auto-merged
+        assert "phone_1" in data["merged_fields"]
+
+        # Source should be soft-deleted (not in list)
+        list_resp = await client.get("/api/v1/contacts", headers=ORIGIN_HEADER)
+        contact_ids = [c["id"] for c in list_resp.json()["items"]]
+        assert source_id not in contact_ids
+        assert target_id in contact_ids
+
+    async def test_merge_same_contact_fails(self, client: AsyncClient, db_session):
+        """Merging a contact with itself should fail."""
+        await seed_tenant_and_users(db_session)
+        await login_client(client, "admin@tenanta.com")
+        csrf = await _get_csrf_token(client)
+        hdrs = _csrf_headers(csrf)
+
+        resp = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Same", "surname": "Contact", "email_1": "same@example.com"},
+            headers=hdrs,
+        )
+        assert resp.status_code == 201, f"Create failed: {resp.text}"
+        contact_id = resp.json()["id"]
+
+        resp = await client.post(
+            "/api/v1/contacts/merge",
+            json={"source_contact_id": contact_id, "target_contact_id": contact_id},
+            headers=hdrs,
+        )
+        assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+class TestMergeHistory:
+    """GET /api/v1/contacts/merge-history — merge history."""
+
+    async def test_merge_history_returns_records(self, client: AsyncClient, db_session):
+        """After a merge, history should contain the record."""
+        await seed_tenant_and_users(db_session)
+        await login_client(client, "admin@tenanta.com")
+        csrf = await _get_csrf_token(client)
+        hdrs = _csrf_headers(csrf)
+
+        # Create and merge two contacts
+        resp1 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Hist", "surname": "Source", "email_1": "hist-source@example.com"},
+            headers=hdrs,
+        )
+        assert resp1.status_code == 201
+        source_id = resp1.json()["id"]
+
+        resp2 = await client.post(
+            "/api/v1/contacts",
+            json={"type": "person", "firstname": "Hist", "surname": "Target", "email_1": "hist-target@example.com"},
+            headers=hdrs,
+        )
+        assert resp2.status_code == 201
+        target_id = resp2.json()["id"]
+
+        merge_resp = await client.post(
+            "/api/v1/contacts/merge",
+            json={"source_contact_id": source_id, "target_contact_id": target_id, "note": "Test merge"},
+            headers=hdrs,
+        )
+        assert merge_resp.status_code == 200
+
+        # Get history (GET doesn't need CSRF)
+        resp = await client.get("/api/v1/contacts/merge-history", headers=ORIGIN_HEADER)
+        assert resp.status_code == 200
+        data = resp.json()
+        assert "items" in data
+        assert "total" in data
+        assert data["total"] >= 1
+        item = data["items"][0]
+        assert item["source_contact_id"] == source_id
+        assert item["target_contact_id"] == target_id
+        assert item["note"] == "Test merge"