Processing level - revised

pyQuARC/code/schema_validator.py

@@ -3,7 +3,7 @@
 import re
 
 from io import BytesIO
-from jsonschema import Draft7Validator, draft7_format_checker, RefResolver
+from jsonschema import Draft7Validator, RefResolver
 from lxml import etree
 from urllib.request import pathname2url
 
@@ -91,7 +91,7 @@ def run_json_validator(self, content_to_validate):
         resolver = RefResolver.from_schema(schema, store=schema_store)
 
         validator = Draft7Validator(
-            schema, format_checker=draft7_format_checker, resolver=resolver
+            schema, format_checker=Draft7Validator.FORMAT_CHECKER, resolver=resolver
         )
 
         for error in sorted(
@@ -136,13 +136,14 @@ def _build_errors(error_log, paths):
             # For DIF, because the namespace is specified in the metadata file, lxml library
             # provides field name concatenated with the namespace,
             # the following 3 lines of code removes the namespace
-            namespaces = re.findall("(\{http[^}]*\})", line)
+
+            namespaces = re.findall(r"(\{http[^}]*\})", line)
             for namespace in namespaces:
                 line = line.replace(namespace, "")
-            field_name = re.search("Element\s'(.*)':", line)[1]
+            field_name = re.search(r"Element\s'(.*)':", line)[1]  
             field_paths = [abs_path for abs_path in paths if field_name in abs_path]
             field_name = field_paths[0] if len(field_paths) == 1 else field_name
-            message = re.search("Element\s'.+':\s(\[.*\])?(.*)", line)[2].strip()
+            message = re.search(r"Element\s'.+':\s(\[.*\])?(.*)", line)[2].strip()
             errors.setdefault(field_name, {})["schema"] = {
                 "message": [f"Error: {message}"],
                 "valid": False,

pyQuARC/schemas/check_messages_override.json

@@ -1 +1,18 @@
-{}
+{
+    "processing_level_description_presence_check": {
+       "failure": "The Processing Level Description is missing.",
+       "help": {
+           "message": "Recommend providing a processing level description, using the EOSDIS processing level descriptions as guidance",
+           "url": "https://www.earthdata.nasa.gov/learn/earth-observation-data-basics/data-processing-levels"
+       },
+       "remediation": "Recommend providing a processing level description, using the EOSDIS processing level descriptions as guidance:\nhttps://www.earthdata.nasa.gov/engage/open-data-services-and-software/data-information-policy/data-levels"
+    },
+    "processing_level_description_length_check": {
+        "failure": "The provided description is less than 50 characters and therefore may be lacking in contextual information.",
+        "help": {
+            "message": "Use the EOSDIS Data Processing level description as guidance.",
+            "url": "https://www.earthdata.nasa.gov/engage/open-data-services-and-software/data-information-policy/data-levels"
+        },
+        "remediation": "Recommend providing a more detailed processing level description, using the EOSDIS processing level descriptions as guidance:\nhttps://www.earthdata.nasa.gov/engage/open-data-services-and-software/data-information-policy/data-levels"
+    }
+}

pyQuARC/schemas/rules_override.json

@@ -1 +1,23 @@
-{}
+{
+    "processing_level_description_presence_check": {
+       "rule_name": "Processing Level Description Presence Check",
+       "fields_to_apply": {
+           "echo-c": [
+               {
+                   "fields": [
+                       "Collection/ProcessingLevelDescription"
+                   ]
+               }
+           ],
+           "umm-c": [
+               {
+                   "fields": [
+                       "ProcessingLevel/ProcessingLevelDescription"
+                   ]
+               }
+           ]
+       },
+       "severity": "info",
+       "check_id": "one_item_presence_check"
+   }
+}

tests/test_downloader.py

@@ -9,11 +9,11 @@ class TestDownloader:
     def setup_method(self):
         self.concept_ids = {
             "collection": {
-                "real": "C1339230297-GES_DISC",
+                "real": "C1000000010-CDDIS",
                 "dummy": "C123456-LPDAAC_ECS",
             },
             "granule": {
-                "real": "G1370895082-GES_DISC",
+                "real": "G1001434969-CDDIS",
                 "dummy": "G1000000002-CMR_PROV",
             },
             "invalid": "asdfasdf",