feat: Detect empty packages. Add Format enumeration.

docs/changelog.rst

@@ -1,6 +1,19 @@
 Changelog
 =========
 
+1.1.10 (2024-04-15)
+-------------------
+
+Added
+~~~~~
+
+-  :class:`ocdskit.util.Format` enumeration
+
+Changed
+~~~~~~~
+
+-  :meth:`ocdskit.util.detect_format`: Detect empty packages that set metadata fields but not a ``releases`` or ``records`` field.
+
 1.1.9 (2024-01-05)
 ------------------
 

docs/conf.py

@@ -26,7 +26,7 @@
 author = "Open Contracting Partnership"
 
 # The short X.Y version
-version = "1.1.9"
+version = "1.1.10"
 # The full version, including alpha/beta/rc tags
 release = version
 

ocdskit/util.py

@@ -216,12 +216,14 @@ def detect_format(path, root_path='', reader=open):
         releases_prefix = f'{prefix}releases'
         ocid_prefix = f'{prefix}ocid'
         tag_item_prefix = f'{prefix}tag.item'
+        metadata_prefixes = {f'{prefix}{field}' for field in ('publishedDate', 'publisher.name', 'uri', 'version')}
 
         has_records = False
         has_releases = False
         has_ocid = False
         has_tag = False
         is_compiled = False
+        metadata_count = 0
         is_array = event == 'start_array'
 
         for prefix, event, value in events:
@@ -235,25 +237,47 @@ def detect_format(path, root_path='', reader=open):
                 has_tag = True
                 if value == 'compiled':
                     is_compiled = True
+            elif prefix in metadata_prefixes:
+                metadata_count += 1
             if not prefix and event not in ('end_array', 'end_map', 'map_key'):
-                return _detect_format_result(True, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled)
+                return _detect_format_result(
+                    True, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled, metadata_count
+                )
 
-        return _detect_format_result(False, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled)
+        return _detect_format_result(
+            False, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled, metadata_count
+        )
 
 
-def _detect_format_result(is_concatenated, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled):
+from enum import Enum
+
+class Format(str, Enum):
+    compiled_release = 'compiled release'
+    empty_package = 'empty package'
+    record = 'record'
+    record_package = 'record package'
+    release = 'release'
+    release_package = 'release package'
+    versioned_release = 'versioned release'
+
+
+def _detect_format_result(
+    is_concatenated, is_array, has_records, has_releases, has_ocid, has_tag, is_compiled, metadata_count
+):
     if has_records:
-        detected_format = 'record package'
+        detected_format = Format.record_package
     elif has_releases and has_ocid:
-        detected_format = 'record'
+        detected_format = Format.record
     elif has_releases:
-        detected_format = 'release package'
+        detected_format = Format.release_package
     elif is_compiled:
-        detected_format = 'compiled release'
+        detected_format = Format.compiled_release
     elif has_tag:
-        detected_format = 'release'
+        detected_format = Format.release
     elif has_ocid:
-        detected_format = 'versioned release'
+        detected_format = Format.versioned_release
+    elif metadata_count == 4:
+        detected_format = Format.empty_package
     else:
         if is_array:
             infix = 'array'

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = ocdskit
-version = 1.1.9
+version = 1.1.10
 author = Open Contracting Partnership
 author_email = data@open-contracting.org
 license = BSD

tests/commands/test_detect_format.py

@@ -17,6 +17,7 @@
     ('release-packages.jsonl', 'concatenated JSON, starting with a JSON array of release packages'),
     ('detect-format_mixed.json', 'concatenated JSON, starting with release'),
     ('detect-format_whitespace.json', 'release'),
+    ('detect-format_empty.json', ('empty package')),
 ])
 def test_command(filename, result, capsys, monkeypatch):
     expected = f'tests{os.sep}fixtures{os.sep}{filename}: {result}\n'

tests/fixtures/detect-format_empty.json

@@ -0,0 +1 @@
+{"uri":"https://staging.opentender.eu//protected/master_tender/id/02a99dc5-0d65-4379-b2b7-708d75bc1d1b?format=ocds","version":"1.1","publishedDate":"2024-04-10T08:17:05.643Z","publisher":{"name":"staging.opentender.eu"},"extensions":["https://raw.githubusercontent.com/open-contracting/ocds_lots_extension/v1.1.1/extension.json","https://raw.githubusercontent.com/open-contracting/ocds_bid_extension/v1.1.1/extension.json","https://raw.githubusercontent.com/open-contracting/ocds_requirements_extension/master/extension.json"],"metaData":{"lastModified":"2022-09-10T03:59:29.973141487"}}

tests/test_util.py

@@ -148,6 +148,7 @@ def test_json_dump(data, expected, tmpdir):
     ('release-packages.jsonl', ('release package', True, True)),
     ('detect-format_mixed.json', ('release', True, False)),
     ('detect-format_whitespace.json', ('release', False, False)),
+    ('detect-format_empty.json', ('empty package', False, False)),
 ])
 def test_detect_format(filename, expected):
     result = detect_format(path(filename))