add similarity tests for regular and vector search

Author: esoteric-ephemera

mp_api/client/routes/materials/similarity.py

@@ -117,7 +117,10 @@ def find_similar(
             feature_vector = self.fingerprint_structure(structure_or_mpid)
 
         else:
-            raise ValueError("Please submit a pymatgen Structure or MP ID.")
+            raise MPRestError(
+                "Please submit a pymatgen Structure or MP ID, found ."
+                f"structure_or_mpid = {type(structure_or_mpid)}."
+            )
 
         top = top or MAX_VECTOR_SEARCH_RESULTS
         if not isinstance(top, int) or top < 1:

tests/materials/test_similarity.py

@@ -0,0 +1,84 @@
+
+import os
+
+import numpy as np
+import pytest
+
+from emmet.core.similarity import SimilarityScorer, SimilarityEntry
+from pymatgen.core import Structure
+
+from mp_api.client.core import MPRestError
+from mp_api.client.routes.materials.similarity import SimilarityRester
+
+from core_function import client_search_testing
+
+@pytest.fixture(scope="module")
+def test_struct():
+    poscar = """Al2
+1.0
+   3.3335729972004917    0.0000000000000000    1.9246389981090721
+   1.1111909992801432    3.1429239987499362    1.9246389992542632
+   0.0000000000000000    0.0000000000000000    3.8492780000000000
+Al
+2
+direct
+   0.875    0.875    0.875 Al
+   0.125    0.125    0.125 Al
+"""
+    return Structure.from_str(poscar,fmt="poscar")
+
+@pytest.mark.skipif(os.getenv("MP_API_KEY") is None, reason="No API key found.")
+def test_client():
+    client_search_testing(
+        search_method=SimilarityRester().search,
+        excluded_params=[
+            "num_chunks",
+            "chunk_size",
+            "all_fields",
+            "fields",
+        ],
+        alt_name_dict={
+            "material_ids": "material_id",
+        },
+        custom_field_tests = {
+            "material_ids": ["mp-149","mp-13"],
+            "material_ids": "mp-149"
+        },
+        sub_doc_fields=[],
+    )
+
+@pytest.mark.skipif(os.getenv("MP_API_KEY") is None, reason="No API key found.")
+def test_similarity_vector_search(test_struct):
+
+    rester = SimilarityRester()
+    fv = rester.fingerprint_structure(test_struct)
+    assert isinstance(fv,np.ndarray)
+    assert len(fv) == 122
+    assert isinstance(rester._fingerprinter,SimilarityScorer)
+
+
+    get_top = 5
+    sim_entries = rester.find_similar("mp-149",top=get_top)
+    assert all(
+        isinstance(entry,SimilarityEntry) for entry in sim_entries
+    )
+    assert len(sim_entries) == get_top
+
+    sim_dict_entries = SimilarityRester(use_document_model=False).find_similar("mp-149",top=get_top)
+    assert all(
+        isinstance(entry,dict) and all(
+            k in entry for k in SimilarityEntry.model_fields
+        )
+        for entry in sim_dict_entries
+    )
+
+    with pytest.raises(MPRestError,match="No similarity data available for"):
+        _ = rester.find_similar("mp-0")
+
+    assert all(
+        isinstance(entry,SimilarityEntry) and isinstance(entry.dissimilarity,float)
+        for entry in rester.find_similar(test_struct, top = 2,)
+    )
+
+    with pytest.raises(MPRestError,match="Please submit a pymatgen Structure or MP ID"):
+        _ = rester.find_similar(fv)