Merge pull request #451 from biolink/cpath_tests

add test for CPATH

kgx/cli/cli_utils.py

@@ -219,7 +219,7 @@ def validate(
         },  # streaming processing throws the graph data away
         # ... Second, we inject the Inspector into the transform() call,
         # for the underlying Transformer.process() to use...
-        inspector=validator
+        inspector=validator,
     )
 
     if output:

tests/resources/test-transform-rdf-tsv.yaml

@@ -0,0 +1,25 @@
+configuration:
+  output_directory: ../target
+  checkpoint: false
+transform:
+  source:
+    test_graph:
+      name: "Test Graph"
+      input:
+        format: nt
+        filename:
+          - test-transform-rdf.nt
+      output:
+        format: tsv
+        filename:
+          - test-transform-rdf
+    valid_graph:
+      name: "Valid JSON Graph"
+      input:
+        format: json
+        filename:
+          - valid.json
+      output:
+        format: tsv
+        filename:
+          - valid_graph

tests/resources/test-transform-rdf.nt

@@ -0,0 +1,15 @@
+<http://identifiers.org/ensembl/ENSG0000000000001> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://purl.obolibrary.org/obo/SO_0000704> .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/Gene> .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/GenomicEntity> .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/MolecularEntity> .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/name> "Test Gene 123" .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/description> "This is a Test Gene 123" .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/provided_by> "Test Dataset" .
+<http://identifiers.org/ensembl/ENSG0000000000002> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://purl.obolibrary.org/obo/SO_0000704> .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/Gene> .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/GenomicEntity> .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/category> <https://w3id.org/biolink/vocab/MolecularEntity> .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/name> "Test Gene 456" .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/description> "This is a Test Gene 456" .
+<http://identifiers.org/ensembl/ENSG0000000000002> <https://w3id.org/biolink/vocab/provided_by> "Test Dataset" .
+<http://identifiers.org/ensembl/ENSG0000000000001> <https://w3id.org/biolink/vocab/interacts_with> <http://identifiers.org/ensembl/ENSG0000000000002> .

tests/resources/test-transform-tsv-rdf.yaml

@@ -0,0 +1,26 @@
+configuration:
+  output_directory: ../target
+  checkpoint: false
+transform:
+  source:
+    test_graph:
+      name: "Test Graph"
+      input:
+        format: tsv
+        filename:
+          - test2_nodes.tsv
+          - test2_edges.tsv
+      output:
+        format: nt
+        filename:
+          - test-tranform-tsv-rdf.nt
+    valid_graph:
+      name: "Valid JSON Graph"
+      input:
+        format: json
+        filename:
+          - valid.json
+      output:
+        format: tsv
+        filename:
+          - valid_graph

tests/unit/test_cli_utils.py

@@ -866,6 +866,25 @@ def test_transform2():
     assert os.path.exists(os.path.join(RESOURCE_DIR, "graph_edges.tsv"))
 
 
+def test_transform_rdf_to_tsv():
+    """
+    Transform from a test transform YAML.
+    """
+    transform_config = os.path.join(RESOURCE_DIR, "test-transform-rdf-tsv.yaml")
+    transform(inputs=None, transform_config=transform_config)
+    assert os.path.exists(os.path.join(TARGET_DIR, "test-transform-rdf_edges.tsv"))
+    assert os.path.exists(os.path.join(TARGET_DIR, "test-transform-rdf_nodes.tsv"))
+
+
+def test_transform_tsv_to_rdf():
+    """
+    Transform from a test transform YAML.
+    """
+    transform_config = os.path.join(RESOURCE_DIR, "test-transform-tsv-rdf.yaml")
+    transform(inputs=None, transform_config=transform_config)
+    assert os.path.exists(os.path.join(TARGET_DIR, "test-tranform-tsv-rdf.nt"))
+
+
 def test_merge1():
     """
     Transform from test merge YAML.

tests/unit/test_source/test_json_source.py

@@ -40,6 +40,31 @@ def test_read_json1():
     assert e["relation"] == "RO:0004013"
 
 
+def test_read_json_filter():
+    """
+    Read from a JSON using JsonSource.
+    """
+    t = Transformer()
+    s = JsonSource(t)
+    filters = {
+        "category": {"biolink:Disease"}
+    }
+    s.set_node_filters(filters)
+    g = s.parse(os.path.join(RESOURCE_DIR, "valid.json"))
+    nodes = {}
+    edges = {}
+    for rec in g:
+        if rec:
+            if len(rec) == 4:
+                edges[(rec[0], rec[1])] = rec[3]
+            else:
+                nodes[rec[0]] = rec[1]
+
+    for node in nodes:
+        n = nodes[node]
+        assert n["category"] == ["biolink:Disease"]
+
+
 def test_read_json2():
     """
     Read from a JSON using JsonSource.