Add CLI argument for --profile-path and add profile validation step…

… before bag validation in CLI execution order. See #54. Add unit test and doc entry for above. Fix a typo in a related logging statement.
fair-research · Jul 24, 2023 · 98b716e · prettybits · Jul 25, 2023 · 98b716e
1 parent 8638ca9
commit 98b716e
Show file tree

Hide file tree

Showing 6 changed files with 32 additions and 11 deletions.
diff --git a/bdbag/__init__.py b/bdbag/__init__.py
@@ -26,7 +26,7 @@
 
 logger = logging.getLogger(__name__)
 
-__version__ = "1.7.0dev4"
+__version__ = "1.7.0dev5"
 __bagit_version__ = "1.8.1"
 __bagit_profile_version__ = "1.3.1"
 

diff --git a/bdbag/bdbag_api.py b/bdbag/bdbag_api.py
@@ -446,7 +446,7 @@ def validate_bag_profile(bag_path, profile_path=None):
     profile_url = bag.info.get(BAG_PROFILE_TAG, None)
     if not profile_url:
         raise bdbp.ProfileValidationError("Bag does not contain a BagIt-Profile-Identifier")
-    logger.info("Loading profile: %s" % profile_path if profile_path else profile_url)
+    logger.info("Loading profile: %s" % (profile_path if profile_path else profile_url))
 
     profile = None
     if profile_path:

diff --git a/bdbag/bdbag_cli.py b/bdbag/bdbag_cli.py
@@ -156,6 +156,12 @@ def parse_cli():
         help="Validate a bag against the profile specified by the bag's \"BagIt-Profile-Identifier\" metadata field, "
              "if present. If \"bag-only\" is specified, the bag's serialization method will not be validated.")
 
+    profile_path_arg = "--profile-path"
+    standard_args.add_argument(
+        profile_path_arg, metavar='<file>',
+        help="Optional path to local profile JSON to use for profile validation. If not specified the profile "
+             "referenced in the bag info file will be fetched and used.")
+
     config_file_arg = "--config-file"
     standard_args.add_argument(
         config_file_arg, metavar='<file>',
@@ -374,6 +380,10 @@ def main():
                               config_file=args.config_file,
                               filter_expr=args.fetch_filter)
 
+        if args.validate_profile:
+            if not is_file:
+                profile = bdb.validate_bag_profile(temp_path if temp_path else path, profile_path=args.profile_path)
+
         if args.validate:
             if is_file:
                 temp_path = bdb.extract_bag(path, args.output_path, temp=True if not args.output_path else False)
@@ -392,13 +402,11 @@ def main():
         if archive is None and is_file:
             archive = path
 
-        if args.validate_profile:
-            if is_file:
-                if not temp_path:
-                    temp_path = bdb.extract_bag(path, args.output_path, temp=True if not args.output_path else False)
-            profile = bdb.validate_bag_profile(temp_path if temp_path else path)
-            if not args.validate_profile == "bag-only":
-                bdb.validate_bag_serialization(archive if archive else path, profile)
+        if args.validate_profile == "full" and is_file:
+            if not temp_path:
+                temp_path = bdb.extract_bag(path, args.output_path, temp=args.output_path is None)
+            profile = bdb.validate_bag_profile(temp_path if temp_path else path, profile_path=args.profile_path)
+            bdb.validate_bag_serialization(archive if archive else path, profile)
 
         if args.revert:
             bdb.revert_bag(path)

diff --git a/doc/cli.md b/doc/cli.md
@@ -28,6 +28,7 @@ usage: bdbag
 [--fetch-filter <column><operator><value>]
 [--validate {fast,full,structure,completeness}]
 [--validate-profile {profile-only,full}]
+[--profile-path <file>]
 [--config-file <file>]
 [--keychain-file <file>]
 [--metadata-file <file>]
@@ -185,6 +186,12 @@ Validate a bag against the profile specified by the bag's `BagIt-Profile-Identif
 `bag-only` argument keyword can be used to bypass the otherwise automatic bag serialization validation 
 (implied by the default value, `full`), and therefore is suitable for use on bag directories.
 
+----
+#### `--profile-path <file>`
+Optional path to a `Bagit-Profiles-Specification` JSON file. The file format is described
+[here](https://bagit-profiles.github.io/bagit-profiles-specification/).
+If this argument is not specified, the profile specified by the bag's `BagIt-Profile-Identifier` metadata field will be used, if present.
+
 ----
 #### `--config-file <file>`
 Optional path to a *bdbag* configuration file. The configuration file format is described

diff --git a/setup.py b/setup.py
@@ -75,7 +75,7 @@
         'globus': ["globus_sdk>=1.6.0"],
         'gcs': ["google_cloud_storage"]
     },
-    python_requires='>=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*, <4',
+    python_requires='>=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*, !=3.5.*, <4',
     license='Apache 2.0',
     classifiers=[
         'Intended Audience :: Science/Research',
@@ -85,7 +85,6 @@
         "Operating System :: MacOS :: MacOS X",
         "Operating System :: Microsoft :: Windows",
         'Programming Language :: Python',
-        'Programming Language :: Python :: 3.6',
         'Programming Language :: Python :: 3.7',
         'Programming Language :: Python :: 3.8',
         'Programming Language :: Python :: 3.9',

diff --git a/test/test_cli.py b/test/test_cli.py
@@ -146,6 +146,13 @@ def test_validate_profile_skip_serialization(self):
         self._test_successful_invocation(
             args, ["Bag structure conforms to specified profile"])
 
+    def test_validate_local_profile(self):
+        args = ARGS + [self.test_bag_profile_dir, '--validate-profile', 'bag-only',
+                       '--profile-path', './profiles/bdbag-profile.json']
+        logfile.writelines(self.getTestHeader('validate-local-profile', args))
+        self._test_successful_invocation(
+            args, ["Loading profile: ./profiles/bdbag-profile.json", "Bag structure conforms to specified profile"])
+
 
 class TestCliArgParsing(BaseTest):