Merge pull request #13 from DurhamARC/master

Check that the given schema version is valid on creating the validator
HEPData · Feb 27, 2020 · 7a5b43b · 7a5b43b
2 parents ee168b1 + 3745fe7
commit 7a5b43b
Show file tree

Hide file tree

Showing 5 changed files with 59 additions and 13 deletions.
diff --git a/hepdata_validator/__init__.py b/hepdata_validator/__init__.py
@@ -23,12 +23,14 @@
 # as an Intergovernmental Organization or submit itself to any jurisdiction.
 
 import abc
+import os
 
 from .version import __version__
 
 __all__ = ('__version__', )
 
-LATEST_SCHEMA_VERSION = '1.0.0'
+VALID_SCHEMA_VERSIONS = ['1.0.0', '0.1.0']
+LATEST_SCHEMA_VERSION = VALID_SCHEMA_VERSIONS[0]
 
 RAW_SCHEMAS_URL = 'https://raw.githubusercontent.com/HEPData/hepdata-validator/' \
     + __version__ + '/hepdata_validator/schemas'
@@ -46,6 +48,19 @@ def __init__(self, *args, **kwargs):
         self.default_schema_file = ''
         self.schemas = kwargs.get('schemas', {})
         self.schema_version = kwargs.get('schema_version', LATEST_SCHEMA_VERSION)
+        if self.schema_version not in VALID_SCHEMA_VERSIONS:
+            raise ValueError('Invalid schema version ' + self.schema_version)
+
+    def _get_schema_filepath(self, schema_filename):
+        full_filepath = os.path.join(self.base_path,
+                                     'schemas',
+                                     self.schema_version,
+                                     schema_filename)
+
+        if not os.path.isfile(full_filepath):
+            raise ValueError('Invalid schema file ' + full_filepath)
+
+        return full_filepath
 
     @abc.abstractmethod
     def validate(self, **kwargs):

diff --git a/hepdata_validator/data_file_validator.py b/hepdata_validator/data_file_validator.py
@@ -50,10 +50,7 @@ class DataFileValidator(Validator):
 
     def __init__(self, *args, **kwargs):
         super(DataFileValidator, self).__init__(*args, **kwargs)
-        self.default_schema_file = os.path.join(self.base_path,
-                                                'schemas',
-                                                self.schema_version,
-                                                self.schema_name)
+        self.default_schema_file = self._get_schema_filepath(self.schema_name)
 
     def load_custom_schema(self, type, schema_file_path=None):
         """

diff --git a/hepdata_validator/submission_file_validator.py b/hepdata_validator/submission_file_validator.py
@@ -25,14 +25,8 @@ class SubmissionFileValidator(Validator):
 
     def __init__(self, *args, **kwargs):
         super(SubmissionFileValidator, self).__init__(*args, **kwargs)
-        self.default_schema_file = os.path.join(self.base_path,
-                                                'schemas',
-                                                self.schema_version,
-                                                self.submission_filename)
-        self.additional_info_schema = os.path.join(self.base_path,
-                                                   'schemas',
-                                                   self.schema_version,
-                                                   self.additional_info_filename)
+        self.default_schema_file = self._get_schema_filepath(self.submission_filename)
+        self.additional_info_schema = self._get_schema_filepath(self.additional_info_filename)
 
     def validate(self, **kwargs):
         """

diff --git a/testsuite/test_data_validator.py b/testsuite/test_data_validator.py
@@ -1,5 +1,6 @@
 import os
 import pytest
+from hepdata_validator import VALID_SCHEMA_VERSIONS
 from hepdata_validator.data_file_validator import DataFileValidator
 from hepdata_validator.data_file_validator import UnsupportedDataSchemaException
 
@@ -255,3 +256,22 @@ def test_file_with_inconsistent_values_v1(validator_v1, data_path):
     validator_v1.print_errors(file)
 
     assert is_valid is False
+
+
+def test_invalid_schema_version():
+    with pytest.raises(ValueError) as excinfo:
+        validator = DataFileValidator(schema_version='0.9999.99')
+
+    assert "Invalid schema version 0.9999.99" == str(excinfo.value)
+
+
+def test_invalid_schema_file():
+    # Fudge the schema versions constant so we can check the file check works
+    VALID_SCHEMA_VERSIONS.append('0.9999.9999')
+    try:
+        with pytest.raises(ValueError) as excinfo:
+            validator = DataFileValidator(schema_version='0.9999.9999')
+
+        assert "Invalid schema file" in str(excinfo.value)
+    finally:
+        VALID_SCHEMA_VERSIONS.pop()
diff --git a/testsuite/test_submission_validator.py b/testsuite/test_submission_validator.py
@@ -1,6 +1,7 @@
 import os
 import pytest
 import yaml
+from hepdata_validator import VALID_SCHEMA_VERSIONS
 from hepdata_validator.submission_file_validator import SubmissionFileValidator
 
 # We try to load using the CSafeLoader for speed improvements.
@@ -196,3 +197,22 @@ def test_io_error_submission_yaml_v1(validator_v1, data_path):
     validator_v1.print_errors(file)
 
     assert is_valid is False
+
+
+def test_invalid_schema_version():
+    with pytest.raises(ValueError) as excinfo:
+        validator = SubmissionFileValidator(schema_version='0.9999.99')
+
+    assert "Invalid schema version 0.9999.99" == str(excinfo.value)
+
+
+def test_invalid_schema_file():
+    # Fudge the schema versions constant so we can check the file check works
+    VALID_SCHEMA_VERSIONS.append('0.9999.9999')
+    try:
+        with pytest.raises(ValueError) as excinfo:
+            validator = SubmissionFileValidator(schema_version='0.9999.9999')
+
+        assert "Invalid schema file" in str(excinfo.value)
+    finally:
+        VALID_SCHEMA_VERSIONS.pop()