validator: catch exceptions, allow absence of comment, bump version

* Catch ParserError and IOError exceptions when loading YAML files.
* Allow a submission.yaml file without an initial "comment" key.
* Add tests to give 100% coverage.
* Bump version to 0.1.15.

Signed-off-by: Graeme Watt <graeme.watt@durham.ac.uk>

hepdata_validator/data_file_validator.py

@@ -27,6 +27,7 @@
 import os
 import yaml
 from yaml.scanner import ScannerError
+from yaml.parser import ParserError
 
 from hepdata_validator import Validator, ValidationMessage
 from jsonschema import validate as json_validate, ValidationError
@@ -95,14 +96,29 @@ def validate(self, **kwargs):
                     self.add_validation_message(ValidationMessage(file=file_path, message=
                     'There was a problem parsing the file.\n' + str(se)))
                     return False
+                except ParserError as pe:
+                    self.add_validation_message(ValidationMessage(file=file_path, message=
+                    'There was a problem parsing the file.\n' + pe.__str__()))
+                    return False
+                except IOError as ioe:
+                    self.add_validation_message(ValidationMessage(file=file_path, message=
+                    'There was a problem parsing the file.\n' + ioe.__str__()))
+                    return False
             except: #pragma: no cover
                 try:  # pragma: no cover
                     data = yaml.load(open(file_path, 'r'))  # pragma: no cover
                 except ScannerError as se:  # pragma: no cover
-                    self.add_validation_message(
-                        ValidationMessage(file=file_path, message=
+                    self.add_validation_message(ValidationMessage(file=file_path, message=
                     'There was a problem parsing the file.\n' + str(se))) # pragma: no cover
                     return False
+                except ParserError as pe:  # pragma: no cover
+                    self.add_validation_message(ValidationMessage(file=file_path, message=
+                    'There was a problem parsing the file.\n' + pe.__str__()))
+                    return False
+                except IOError as ioe:
+                    self.add_validation_message(ValidationMessage(file=file_path, message=
+                    'There was a problem parsing the file.\n' + ioe.__str__()))
+                    return False
 
         try:
 

hepdata_validator/submission_file_validator.py

@@ -3,6 +3,7 @@
 import os
 import yaml
 from yaml.scanner import ScannerError
+from yaml.parser import ParserError
 from hepdata_validator import Validator, ValidationMessage
 
 __author__ = 'eamonnmaguire'
@@ -48,7 +49,7 @@ def validate(self, **kwargs):
                 if data_item is None:
                     continue
                 try:
-                    if 'comment' in data_item:
+                    if 'name' not in data_item:
                         validate(data_item, additional_file_section_schema)
                     else:
                         validate(data_item, submission_file_schema)
@@ -70,4 +71,12 @@ def validate(self, **kwargs):
                     'This can be because you forgot spaces '
                     'after colons in your YAML file for instance.  '
                     'Diagnostic information follows.\n' + str(se)))
+            return False
+
+        except ParserError as pe:
+            self.add_validation_message(ValidationMessage(file=file_path, message=pe.__str__()))
+            return False
+
+        except IOError as ioe:
+            self.add_validation_message(ValidationMessage(file=file_path, message=ioe.__str__()))
             return False

hepdata_validator/version.py

@@ -27,4 +27,4 @@
 
 from __future__ import absolute_import, print_function
 
-__version__ = "0.1.14"
+__version__ = "0.1.15"

testsuite/test_data/invalid_parser_file.yaml

@@ -0,0 +1,20 @@
+---
+independent_variables:
+  - header: {name: SQRT(S), units: GEV}
+    values:
+      - value: 7000
+dependent_variables:
+  - header: {name: SIG(total), units: FB}
+    qualifiers:
+      - {name: RE, value: P P --> Z0 Z0 X # note missing final "}"
+    values:
+      - value: 6.7
+        errors:
+          - {symerror: 0.45, label: stat}
+          - {asymerror: {plus: 0.4, minus: -0.3}, label: sys}
+          - {symerror: 0.34, label: "sys,lumi"}
+      - value: 5.7
+        errors:
+          - {symerror: 0.4, label: stat}
+          - {asymerror: {plus: 0.42, minus: 0.31}, label: sys}
+          - {symerror: 0.4, label: "sys,lumi"}

testsuite/test_data/invalid_parser_submission.yaml

@@ -0,0 +1,14 @@
+---
+name: "Table 3"
+location: Page 17 of preprint
+description: The measured fiducial cross sections.  The first systematic uncertainty is the combined systematic uncertainty excluding luminosity, the second is the luminosity
+keywords: # used for searching, possibly multiple values for each keyword
+  - { name: reactions, value: [P P --> Z0 Z0 X] # note missing final "}"
+  - { name: observables, value: [SIG]}
+  - { name: energies, value: [7000]}
+data_file: data3.yaml
+additional_resources:
+  - location: "http:github.com/HEPData/hepdata"
+    description: "Full source code for creating this data"
+  - location: "http:github.com/HEPData/hepdata"
+    description: "Full source code for creating this data"

testsuite/validation_test.py

@@ -19,6 +19,7 @@ def setUp(self):
         self.valid_empty_file = 'test_data/valid_submission_empty.yaml'
         self.invalid_file = 'test_data/invalid_submission.yaml'
         self.invalid_syntax_file = 'test_data/invalid_syntax_submission.yaml'
+        self.invalid_parser_file = 'test_data/invalid_parser_submission.yaml'
 
     def test_valid_submission_yaml(self):
         print '___SUBMISSION_FILE_VALIDATION: Testing valid yaml submission___'
@@ -103,6 +104,32 @@ def test_invalid_submission_yaml(self):
 
         self.validator.print_errors(invalid_sub_yaml)
 
+    def test_invalid_parser_submission_yaml(self):
+        print '___SUBMISSION_FILE_VALIDATION: ' \
+              'Testing invalid parser yaml submission___'
+        self.validator = None
+        self.validator = SubmissionFileValidator()
+        invalid_sub_yaml = os.path.join(self.base_dir, self.invalid_parser_file)
+
+        self.assertEqual(self.validator.validate(
+            file_path=invalid_sub_yaml), False
+        )
+
+        self.validator.print_errors(invalid_sub_yaml)
+
+    def test_ioerror_submission_yaml(self):
+        print '___SUBMISSION_FILE_VALIDATION: ' \
+              'Testing ioerror yaml submission___'
+        self.validator = None
+        self.validator = SubmissionFileValidator()
+        invalid_sub_yaml = os.path.join(self.base_dir, self.valid_file[:-1])
+
+        self.assertEqual(self.validator.validate(
+            file_path=invalid_sub_yaml), False
+        )
+
+        self.validator.print_errors(invalid_sub_yaml)
+
 
 class DataValidationTest(unittest.TestCase):
     validator = None
@@ -140,10 +167,21 @@ def setUp(self):
             'test_data/invalid_data_file.yaml'
         )
 
+        self.invalid_parser_file = os.path.join(
+            self.base_dir,
+            'test_data/invalid_parser_file.yaml'
+        )
+
         self.valid_custom_file = os.path.join(
             self.base_dir,
             'test_data/valid_file_custom.yaml')
 
+    def test_no_file_path_supplied(self):
+        try:
+            self.validator.validate(file_path=None)
+        except LookupError as le:
+            assert (le)
+
     def test_valid_yaml_file(self):
         print '___DATA_VALIDATION: Testing valid yaml submission___'
         is_valid = self.validator.validate(file_path=self.valid_file_yaml)
@@ -207,6 +245,20 @@ def test_load_invalid_data_file(self):
         for message in self.validator.get_messages(self.invalid_syntax_data_file):
             self.assertTrue(message.message.index("There was a problem parsing the file.") == 0)
 
+    def test_invalid_parser_yaml_file(self):
+        print '___DATA_VALIDATION: Testing invalid parser yaml submission___'
+        self.assertEqual(self.validator.validate(file_path=self.invalid_parser_file),
+                         False)
+
+        self.validator.print_errors(self.invalid_parser_file)
+
+    def test_ioerror_yaml_file(self):
+        print '___DATA_VALIDATION: Testing ioerror yaml submission___'
+        self.assertEqual(self.validator.validate(file_path=self.valid_file_yaml[:-1]),
+                         False)
+
+        self.validator.print_errors(self.valid_file_yaml[:-1])
+
 
 if __name__ == '__main__':
     unittest.main()