partial work toward aggregating errors

Author: nsheff

eido/cli.py

@@ -33,6 +33,26 @@ def _parse_filter_args_str(input):
         else lst
     )
 
+def print_error_summary(errors_by_type):
+        """ Print a summary of errors, organized by error type """
+        n_error_types = len(errors_by_type)
+        print(f"Found {n_error_types} types of error:")
+        for type in errors_by_type:
+            n = len(errors_by_type[type])
+            msg = f"  - {type}: ({n} samples) "
+            if n < 50:
+                msg += ", ".join([x["sample_name"] for x in errors_by_type[type]])
+            print(msg)
+
+        if len(errors_by_type) > 1:
+            final_msg = f"Validation unsuccessful. {len(errors_by_type)} error types found."
+        else:
+            final_msg = f"Validation unsuccessful. {len(errors_by_type)} error type found."
+        
+        print(final_msg)
+        return final_msg
+
+
 
 def main():
     """Primary workflow"""
@@ -128,7 +148,7 @@ def main():
         try:
             validator(*arguments)
         except EidoValidationError as e:
-            print(e)
+            print_error_summary(e.errors_by_type)
             return False
         _LOGGER.info("Validation successful")
         sys.exit(0)

eido/exceptions.py

@@ -40,6 +40,6 @@ def __init__(self, key):
 class EidoValidationError(EidoException):
     """Object was not validated successfully according to schema."""
 
-    def __init__(self, message, errors):
+    def __init__(self, message, errors_by_type):
         super().__init__(message)
-        self.errors = errors
+        self.errors_by_type = errors_by_type

eido/output_formatters.py

@@ -102,13 +102,14 @@ def _convert_sample_to_row(
             Representation of sample as a CSV row.
         """
         sample_row = []
+
         for attribute in sample_attributes:
             if MultilineOutputFormatter._sample_attribute_is_list(
                 sample, attribute
-            ) and getattr(sample, attribute):
-                value = getattr(sample, attribute)[sample_index]
+            ) and sample[attribute]:
+                value = sample[attribute][sample_index]
             else:
-                value = getattr(sample, attribute, "")
+                value = sample[attribute]
 
             sample_row.append(value or "")
 

eido/validation.py

@@ -23,7 +23,7 @@
 _LOGGER = getLogger(__name__)
 
 
-def _validate_object(object, schema, exclude_case=False):
+def _validate_object(object, schema, exclude_case=False, sample_name_colname=False):
     """
     Generic function to validate object against a schema
 
@@ -34,37 +34,30 @@ def _validate_object(object, schema, exclude_case=False):
     """
 
     validator = Draft7Validator(schema)
+    print(object,schema)
     if not validator.is_valid(object):
         errors = sorted(validator.iter_errors(object), key=lambda e: e.path)
         errors_by_type = {}
+
         # Accumulate and restructure error objects by error type
         for error in errors:
             if not error.message in errors_by_type:
                 errors_by_type[error.message] = []
+            
+            try:
+                instance_name = error.instance[sample_name_colname]
+            except KeyError:
+                instance_name = "unnamed"
             errors_by_type[error.message].append(
                 {
                     "type": error.message,
-                    "message": f"{error.message} on instance {error.instance['sample_name']}",
-                    "sample_name": error.instance['sample_name']
+                    "message": f"{error.message} on instance {instance_name}",
+                    "sample_name": instance_name
                 })
-
-        # Print a summary of errors, organized by error type
-        n_error_types = len(errors_by_type)
-        print(f"Found {n_error_types} types of error:")
-        for type in errors_by_type:
-            n = len(errors_by_type[type])
-            msg = f"  - {type}: ({n} samples) "
-            if n < 50:
-                msg += ", ".join([x["sample_name"] for x in errors_by_type[type]])
-            print(msg)
-
-        if len(errors) > 1:
-            final_msg = f"Validation unsuccessful. {len(errors)} errors found."
-        else:
-            final_msg = f"Validation unsuccessful. {len(errors)} error found."
-
-        raise EidoValidationError(final_msg, errors)
-
+            
+        raise EidoValidationError("Validation failed", errors_by_type)
+    else:
+        _LOGGER.debug("Validation was successful...")
 
 def validate_project(project, schema, exclude_case=False):
     """
@@ -75,6 +68,7 @@ def validate_project(project, schema, exclude_case=False):
     :param bool exclude_case: whether to exclude validated objects
     from the error. Useful when used ith large projects
     """
+    sample_name_colname = project.sample_name_colname
     schema_dicts = read_schema(schema=schema)
     for schema_dict in schema_dicts:
         project_dict = project.to_dict()

requirements/requirements-all.txt

@@ -1,5 +1,5 @@
 jsonschema>=3.0.1
 logmuse>=0.2.5
 pandas
-peppy>=0.32.0
+peppy>=0.40.0
 ubiquerg>=0.5.2

tests/conftest.py

@@ -90,7 +90,12 @@ def output_pep_with_fasta_column(path_pep_with_fasta_column):
 
 @pytest.fixture
 def taxprofiler_csv_multiline_output(path_to_taxprofiler_csv_multiline_output):
-    return pd.read_csv(path_to_taxprofiler_csv_multiline_output).to_csv(
+    with open(path_to_taxprofiler_csv_multiline_output, 'r') as file:
+        data = file.read()
+    return data
+    # This is broken unless I add na_filter=False. But it's a bad idea anyway, since
+    # we're just using this for string comparison anyway...
+    return pd.read_csv(path_to_taxprofiler_csv_multiline_output, na_filter=False).to_csv(
         path_or_buf=None, index=None
     )
 

tests/data/schemas/test_schema_invalid_with_type.yaml

@@ -0,0 +1,25 @@
+description: test PEP schema
+type: object
+properties:
+  dcc:
+    type: object
+    properties:
+      compute_packages:
+        type: object
+  samples:
+    type: array
+    items:
+      type: object
+      properties:
+        sample_name:
+          type: string
+        protocol:
+          type: string
+        genome:
+          type: string
+  invalid:
+    type: string
+
+required:
+  - samples
+  - invalid