@vespermcp/mcp-server 1.1.3 → 1.2.0

package/src/python/test_fusion_engine.py

@@ -0,0 +1,89 @@
+import os
+import tempfile
+import polars as pl
+from fusion_engine import fuse_datasets
+
+
+def run_basic_tests():
+    tmp = tempfile.gettempdir()
+
+    # ----- Test 1: concat -----
+    p1 = os.path.join(tmp, "fuse_test_a.csv")
+    p2 = os.path.join(tmp, "fuse_test_b.csv")
+    out_concat = os.path.join(tmp, "fuse_test_concat.feather")
+
+    df1 = pl.DataFrame({
+        "id": [1, 2, 3],
+        "text": ["a", "b", "c"],
+        "price": [10.0, 20.0, 30.0],
+    })
+    df2 = pl.DataFrame({
+        "id": [4, 5, 3],
+        "text": ["d", "e", "c"],
+        "price": [40.0, 50.0, 30.0],
+        "image_path": ["img1.jpg", "img2.jpg", "img3.jpg"],
+    })
+
+    df1.write_csv(p1)
+    df2.write_csv(p2)
+
+    concat_res = fuse_datasets(
+        sources=[p1, p2],
+        strategy="concat",
+        dedup=True,
+        run_quality_after=False,
+        leakage_check=True,
+        output_path=out_concat,
+        output_format="feather",
+        compression="lz4",
+        preview=True,
+        id_column="id",
+    )
+
+    assert concat_res.get("success") is True, f"Concat failed: {concat_res}"
+    assert os.path.exists(out_concat), "Concat output file missing"
+
+    # ----- Test 2: join with conflicting column names -----
+    p3 = os.path.join(tmp, "fuse_test_c.csv")
+    p4 = os.path.join(tmp, "fuse_test_d.csv")
+    out_join = os.path.join(tmp, "fuse_test_join.parquet")
+
+    left = pl.DataFrame({
+        "id": [1, 2, 3],
+        "price": [100, 200, 300],
+        "text": ["x", "y", "z"],
+    })
+    right = pl.DataFrame({
+        "id": [2, 3, 4],
+        "price": [999, 888, 777],
+        "caption": ["two", "three", "four"],
+    })
+
+    left.write_csv(p3)
+    right.write_csv(p4)
+
+    join_res = fuse_datasets(
+        sources=[p3, p4],
+        strategy="join",
+        join_on="id",
+        how="inner",
+        dedup=True,
+        run_quality_after=False,
+        leakage_check=False,
+        output_path=out_join,
+        output_format="parquet",
+        compression="snappy",
+        preview=True,
+    )
+
+    assert join_res.get("success") is True, f"Join failed: {join_res}"
+    assert os.path.exists(out_join), "Join output file missing"
+    assert len(join_res.get("stats", {}).get("conflict_renames", [])) >= 1, "Expected conflict rename for price column"
+
+    print("✅ Fusion tests passed")
+    print("Concat:", concat_res["stats"])
+    print("Join:", join_res["stats"])
+
+
+if __name__ == "__main__":
+    run_basic_tests()