Add datapackage_reports_path option.

`datapackage_reports_path` defines the path used to write the `path`
property of a report in the datapackage.json. This maybe different from
the `reports_path`, which is where reports are actually written. If
`datapackage_reports_path` is missing, no `reports` property is added to
the datapackage.

README.md

@@ -28,18 +28,20 @@ Add the following to the pipeline-spec.yml configuration to validate each resour
   - run: goodtables.validate
     parameters:
         fail_on_error: True,
-        reports_path: 'reports',
+        reports_path: 'path/to/datapackage/reports',  # where reports will be written
+        datapackage_reports_path: 'reports',  # relative to datapackage.json
         write_report: True,
         goodtables:
             <key>: <value>  # options passed to goodtables.validate()
 ```
 
 - `fail_on_error`: An optional boolean to determine whether the pipeline should fail on validation error (default `True`).
 - `reports_path`: An optional string to define where Goodtables reports should be written (default is `reports`).
+- `datapackage_reports_path`: An optional string to define the path to the report, relative to the datapackage.json (see note below).
 - `write_report`: An optional boolean to determine whether a goodtables validation report should be written to `reports_path` (default is `True`).
 - `goodtables`: An optional object passed to `goodtables.validate()` to customise its behaviour. See [`goodtables.validate()`](https://github.com/frictionlessdata/goodtables-py/#validatesource-options) for available options.
 
-If reports are written, a `reports` property will be added to the datapackage, detailing the path to the report for each resource:
+If reports are written, and `datapackage_reports_path` is defined, a `reports` property will be added to the datapackage, detailing the path to the report for each resource:
 
 ```json
 ...
@@ -51,3 +53,5 @@ If reports are written, a `reports` property will be added to the datapackage, d
     }
 ]
 ```
+
+It is recommended that `datapackage_reports_path` is used to define a relative path, from the datapackage.json file, that represents where the report was written. `datapackage_reports_path` does not define where the reports will be written, but helps ensure a correct path is defined in the `reports` property in datapackage.json. This is useful when the pipeline concludes with a `dump_to.path` processor.

datapackage_pipelines_goodtables/processors/validate.py

@@ -16,6 +16,7 @@
 write_report = parameters.get('write_report', True)
 goodtables_options = parameters.get('goodtables', {})
 reports_path = parameters.get('reports_path', 'reports')
+datapackage_reports_path = parameters.get('datapackage_reports_path', None)
 
 
 def process_resources(res_iter_, datapackage, goodtables_options):
@@ -57,14 +58,15 @@ def _validate_resource(res, dp_res):
         yield _validate_resource(res, dp_res)
 
 
-# add report info to datapackage
-if write_report:
+# Add report info to datapackage
+if write_report and datapackage_reports_path is not None:
     reports = datapackage.get('reports', [])
     for dp_res in datapackage['resources']:
         reports.append({
             'resource': dp_res['name'],
             'reportType': 'goodtables',
-            'path': '{}/{}.json'.format(reports_path, dp_res['name'])
+            'path': '{}/{}.json'.format(datapackage_reports_path,
+                                        dp_res['name'])
         })
     datapackage['reports'] = reports
 

tests/test_validate.py

@@ -50,7 +50,9 @@ def test_validate_processor_no_resources_no_params(self):
             'project': 'my-project',
             'resources': []
         }
-        params = {}
+        params = {
+            'datapackage_reports_path': 'reports'
+        }
 
         # Path to the processor we want to test
         processor_dir = os.path.dirname(
@@ -79,7 +81,8 @@ def test_validate_processor_no_resources_with_params(self):
             'resources': []
         }
         params = {
-            'fail_on_error': True
+            'fail_on_error': True,
+            'datapackage_reports_path': 'reports'
         }
 
         # Path to the processor we want to test
@@ -120,7 +123,8 @@ def test_validate_processor_valid_resource(self):
         report_dir = '{}'.format(self.get_base_path())
 
         params = {
-            'reports_path': report_dir
+            'reports_path': report_dir,
+            'datapackage_reports_path': 'reports',
         }
 
         def row_yielder():
@@ -150,7 +154,7 @@ def row_yielder():
         assert len(reports) == 1
         assert reports[0]['resource'] == 'my-resource'
         assert reports[0]['reportType'] == 'goodtables'
-        assert reports[0]['path'] == '{}/my-resource.json'.format(report_dir)
+        assert reports[0]['path'] == 'reports/my-resource.json'
 
         with io.open('{}/my-resource.json'.format(report_dir), 'r') as f:
             report_json = json.loads(f.read())
@@ -336,6 +340,7 @@ def test_validate_processor_invalid_resource_dont_fail(self):
 
         params = {
             'reports_path': report_dir,
+            'datapackage_reports_path': 'reports',
             'fail_on_error': False
         }
 
@@ -366,8 +371,7 @@ def row_yielder():
         assert len(reports) == 1
         assert reports[0]['resource'] == 'resource-dont-fail-write'
         assert reports[0]['reportType'] == 'goodtables'
-        assert reports[0]['path'] == \
-            '{}/resource-dont-fail-write.json'.format(report_dir)
+        assert reports[0]['path'] == 'reports/resource-dont-fail-write.json'
 
         with io.open('{}/resource-dont-fail-write.json'.format(report_dir),
                      'r') as f: