matchlab 0.1.0.dev4__tar.gz → 0.1.0.dev5__tar.gz

{matchlab-0.1.0.dev4 → matchlab-0.1.0.dev5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: matchlab
-Version: 0.1.0.dev4
+Version: 0.1.0.dev5
 Summary: A local-first library for building, running and evaluating entity resolution pipelines.
 Author: Department for Business and Trade
 Project-URL: Documentation, https://uktrade.github.io/matchlab/

{matchlab-0.1.0.dev4 → matchlab-0.1.0.dev5}/src/matchlab/models/linkers/splinklinker.py

@@ -217,9 +217,12 @@ class SplinkLinker(Linker):
         self._id_dtype_l = left[self.left_id].dtype
         self._id_dtype_r = right[self.right_id].dtype
 
-        if (set(left.columns) != set(right.columns)) or not left.dtypes == right.dtypes:
+        if set(left.columns) != set(right.columns):
+            raise self._conformancy_error(left, right)
+        if left.columns != right.columns:
+            right = right.select(left.columns)
+        if left.dtypes != right.dtypes:
             raise self._conformancy_error(left, right)
-
         # Convert to pandas for Splink compatibility
         left_pd = left.with_columns(pl.col(self.left_id).cast(pl.String)).to_pandas()
         right_pd = right.with_columns(pl.col(self.right_id).cast(pl.String)).to_pandas()

{matchlab-0.1.0.dev4 → matchlab-0.1.0.dev5}/src/matchlab.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: matchlab
-Version: 0.1.0.dev4
+Version: 0.1.0.dev5
 Summary: A local-first library for building, running and evaluating entity resolution pipelines.
 Author: Department for Business and Trade
 Project-URL: Documentation, https://uktrade.github.io/matchlab/

matchlab-0.1.0.dev5/test/models/test_linkers_splink.py

@@ -0,0 +1,65 @@
+"""Direct unit tests for `SplinkLinker` behavior.
+
+These tests cover Splink-specific schema handling and diagnostics, rather than the
+shared match/no-match behavior exercised in the other linker suites.
+"""
+
+from datetime import date
+from unittest.mock import Mock, patch
+
+import polars as pl
+from splink import SettingsCreator
+from splink import blocking_rule_library as brl
+from splink import comparison_library as cl
+
+from matchlab.models.linkers.splinklinker import SplinkLinker
+
+
+def test_splink_schema_order() -> None:
+    """Splink accepts matching schemas even when the incoming column order differs."""
+    linker = SplinkLinker(
+        left_id="id",
+        right_id="id",
+        linker_training_functions=[],
+        linker_settings=SettingsCreator(
+            link_type="link_only",
+            blocking_rules_to_generate_predictions=[brl.block_on("company_name")],
+            comparisons=[cl.ExactMatch("company_name")],
+        ),
+        threshold=None,
+    )
+    left = pl.DataFrame(
+        {
+            "id": [1],
+            "postcode": ["SW1A 1AA"],
+            "duns_number": ["123456789"],
+            "company_name": ["left"],
+            "incorporation_date": [date(2024, 1, 1)],
+        }
+    )
+    right = pl.DataFrame(
+        {
+            "id": [2],
+            "company_name": ["right"],
+            "postcode": ["SW1A 1AA"],
+            "incorporation_date": [date(2024, 1, 1)],
+            "duns_number": ["123456789"],
+        }
+    )
+
+    fake_linker = Mock()
+    with patch(
+        "matchlab.models.linkers.splinklinker.SplinkLibLinkerClass",
+        return_value=fake_linker,
+    ) as splink_linker:
+        linker.prepare(left, right)
+
+    left_pd, right_pd = splink_linker.call_args.kwargs["input_table_or_tables"]
+    assert list(left_pd.columns) == list(right_pd.columns)
+    assert list(left_pd.columns) == [
+        "id",
+        "postcode",
+        "duns_number",
+        "company_name",
+        "incorporation_date",
+    ]

matchlab-0.1.0.dev4/test/models/test_linkers_splink.py

@@ -1,42 +0,0 @@
-"""Direct unit tests for `SplinkLinker` behavior.
-
-These tests cover Splink-specific schema handling and diagnostics, rather than the
-shared match/no-match behavior exercised in the other linker suites.
-"""
-
-import polars as pl
-import pytest
-from splink import SettingsCreator
-from splink import blocking_rule_library as brl
-from splink import comparison_library as cl
-
-from matchlab.models.linkers.splinklinker import SplinkLinker
-
-
-def test_splink_conformancy_error() -> None:
-    """Splink surfaces the concrete schema mismatch from the real prepare path."""
-    linker = SplinkLinker(
-        left_id="id",
-        right_id="id",
-        linker_training_functions=[],
-        linker_settings=SettingsCreator(
-            link_type="link_only",
-            blocking_rules_to_generate_predictions=[brl.block_on("id")],
-            comparisons=[cl.ExactMatch("id")],
-        ),
-        threshold=None,
-    )
-    left = pl.DataFrame({"id": [1], "company_name": ["left"]})
-    right = pl.DataFrame({"id": [2], "postcode": ["SW1A 1AA"]})
-
-    with pytest.raises(
-        ValueError,
-        match="SplinkLinker requires input data to be conformant",
-    ) as exc_info:
-        linker.prepare(left, right)
-
-    message = str(exc_info.value)
-    assert "Left schema: id=Int64, company_name=String" in message
-    assert "Right schema: id=Int64, postcode=String" in message
-    assert "Columns only on left: ['company_name']" in message
-    assert "Columns only on right: ['postcode']" in message