More readable output for validate

setup.cfg

@@ -151,6 +151,7 @@ ignore =
     S318 # don't worry about unsafe xml
     S310 # TODO remove this later and switch to using requests
     B018 # This is 'useless' statements which are new atm.
+    B009 # Prevents me from using get_attr after checking has_attr
 exclude =
     sssom/sssom_datamodel.py
     sssom/cliquesummary.py

sssom/io.py

@@ -28,6 +28,7 @@
 from .typehints import Metadata
 from .util import (
     MappingSetDataFrame,
+    SssomMalformedYamlError,
     are_params_slots,
     augment_metadata,
     is_curie,
@@ -114,18 +115,24 @@ def parse_file(
 
 
 def validate_file(
-    input_path: str, validation_types: List[SchemaValidationType]
+    input_path: str, validation_types: List[SchemaValidationType], verbose: bool = False
 ) -> None:
     """Validate the incoming SSSOM TSV according to the SSSOM specification.
 
     :param input_path: The path to the input file in one of the legal formats, eg obographs, aligmentapi-xml
     :param validation_types: A list of validation types to run.
+    :param verbose: If true, print detailed error message.
     """
     # Two things to check:
     # 1. All prefixes in the DataFrame are define in prefix_map
     # 2. All columns in the DataFrame abide by sssom-schema.
-    msdf = parse_sssom_table(file_path=input_path)
-    validate(msdf=msdf, validation_types=validation_types)
+    try:
+        msdf = parse_sssom_table(file_path=input_path)
+        validate(msdf=msdf, validation_types=validation_types)
+    except SssomMalformedYamlError as e:
+        print("Your SSSOM mapping set metadata is malformed YAML.")
+        if verbose:
+            print(e)
 
 
 def split_file(input_path: str, output_directory: Union[str, Path]) -> None:

sssom/parsers.py

@@ -58,9 +58,12 @@
     URI_SSSOM_MAPPINGS,
     MappingSetDataFrame,
     NoCURIEException,
+    SssomMalformedYamlError,
     curie_from_uri,
     get_file_extension,
+    get_sssom_error_message,
     is_multivalued_slot,
+    purl_to_source_id,
     raise_for_bad_path,
     read_pandas,
     to_mapping_set_dataframe,
@@ -534,9 +537,13 @@ def from_alignment_minidom(
                 elif node_name == "onto2":
                     ms[OBJECT_SOURCE_ID] = e.firstChild.nodeValue
                 elif node_name == "uri1":
-                    ms[SUBJECT_SOURCE] = e.firstChild.nodeValue
+                    ms[SUBJECT_SOURCE] = purl_to_source_id(
+                        uri=e.firstChild.nodeValue, prefix_map=prefix_map
+                    )
                 elif node_name == "uri2":
-                    ms[OBJECT_SOURCE] = e.firstChild.nodeValue
+                    ms[OBJECT_SOURCE] = purl_to_source_id(
+                        uri=e.firstChild.nodeValue, prefix_map=prefix_map
+                    )
 
     ms.mappings = mlist  # type: ignore
     _set_metadata_in_mapping_set(mapping_set=ms, metadata=meta)
@@ -745,9 +752,12 @@ def _read_metadata_from_table(path: Union[str, Path]) -> Dict[str, Any]:
                 break
 
     if yamlstr:
-        meta = yaml.safe_load(yamlstr)
-        logging.info(f"Meta={meta}")
-        return meta
+        try:
+            meta = yaml.safe_load(yamlstr)
+            logging.info(f"Meta={meta}")
+            return meta
+        except Exception as e:
+            raise SssomMalformedYamlError(get_sssom_error_message(e))
     return {}
 
 

sssom/util.py

@@ -941,10 +941,24 @@ def get_dict_from_mapping(map_obj: Union[Any, Dict[Any, Any], SSSOM_Mapping]) ->
     return map_dict
 
 
+def get_sssom_error_message(e: Exception):
+    """Get a human readble error message from an exception."""
+    msg = ""
+    if hasattr(e, "message"):
+        msg = getattr(e, "message")
+    else:
+        str(e)
+    return msg
+
+
 class NoCURIEException(ValueError):
     """An exception raised when a CURIE can not be parsed with a given prefix map."""
 
 
+class SssomMalformedYamlError(ValueError):
+    """An exception raised when the YAML header is malformed."""
+
+
 CURIE_RE = re.compile(r"[A-Za-z0-9_.]+[:][A-Za-z0-9_]")
 
 
@@ -1012,6 +1026,19 @@ def get_prefixes_used_in_table(df: pd.DataFrame) -> List[str]:
     return list(set(prefixes))
 
 
+def purl_to_source_id(uri: str, prefix_map: Mapping[str, str]) -> str:
+    """Parse a CURIE from an PURL to serve as a generic source reference (by splitting of file extension).
+
+    :param uri: The URI to parse. If this is already a CURIE, return directly.
+    :param prefix_map: The prefix map against which the IRI is checked
+    :return: A CURIE
+    """
+    curie = curie_from_uri(uri=uri, prefix_map=prefix_map)
+    if curie.endswith(".owl"):
+        curie = curie[:-4]
+    return curie
+
+
 def filter_out_prefixes(
     df: pd.DataFrame, filter_prefixes: List[str], features: list = KEY_FEATURES
 ) -> pd.DataFrame:
@@ -1245,7 +1272,7 @@ def sort_df_rows_columns(
             col for col in SCHEMA_DICT["slots"].keys() if col in df.columns
         ]
         df = df.reindex(column_sequence, axis=1)
-    if by_rows and len(df) > 0:
+    if by_rows and len(df) > 0 and len(df.columns) > 0:
         df = df.sort_values(by=df.columns[0], ignore_index=True)
     return df
 

sssom/validators.py

@@ -10,26 +10,34 @@
 
 from sssom.context import add_built_in_prefixes_to_prefix_map
 from sssom.parsers import to_mapping_set_document
-from sssom.util import MappingSetDataFrame, get_all_prefixes
+from sssom.util import MappingSetDataFrame, SssomMalformedYamlError, get_all_prefixes
 
 from .constants import SCHEMA_YAML, SchemaValidationType
 
 
 def validate(
-    msdf: MappingSetDataFrame, validation_types: List[SchemaValidationType]
+    msdf: MappingSetDataFrame,
+    validation_types: List[SchemaValidationType],
+    verbose=False,
 ) -> None:
     """Validate SSSOM files against `sssom-schema` using linkML's validator function.
 
     :param msdf: MappingSetDataFrame.
     :param validation_types: SchemaValidationType
+    :param verbose: If true, print detailed error message.
     """
     validation_methods = {
         SchemaValidationType.JsonSchema: validate_json_schema,
         SchemaValidationType.Shacl: validate_shacl,
         SchemaValidationType.PrefixMapCompleteness: check_all_prefixes_in_curie_map,
     }
     for vt in validation_types:
-        validation_methods[vt](msdf)
+        try:
+            validation_methods[vt](msdf)
+        except SssomMalformedYamlError as e:
+            print("Your SSSOM mapping set metadata is malformed YAML.")
+            if verbose:
+                print(e)
 
 
 def validate_json_schema(msdf: MappingSetDataFrame) -> None: