Update FIVE_SAFES_PROFILE_VERSION to 0.8.0

[manics]

Testing https://github.com/eScienceLab/rocrate-validator/blob/develop/tests/data/crates/valid/five-safes-crate-result/ro-crate-metadata.json (https://github.com/eScienceLab/rocrate-validator/raw/ddf10285cb1df2effde1032be0939eb11870db5d/tests/data/crates/valid/five-safes-crate-result/ro-crate-metadata.json) as suggested in #102 (comment) with ghcr.io/esciencelab/ro-crate-validation-service-fivesafes-profile:2.0

BASE_URL = "http://<INTERNAL-HOST>:5000"

with open("ro-crate-metadata-5s.json", encoding="utf-8") as f:
    crate_json = f.read()
    reqbody = {"crate_json": crate_json, "profile_name": "five-safes-crate"}
    response = requests.post(
        f"{BASE_URL}/v1/ro_crates/validate_metadata",
        json=reqbody,
    )

    assert response.status_code == 200
    body = response.json()
    display(body)
    assert body["status"] == "valid"

results in these validation errors:

{'created_at': None,
 'detail': {'issues': [{'check': {'description': 'Check if the RO-Crate Metadata File Descriptor has a `conformsTo` property which points to the RO-Crate specification version',
     'identifier': 'ro-crate-1.1_5.3',
     'label': 'MUST 5.3',
     'name': 'Metadata File Descriptor entity: `conformsTo` property',
     'order': 3,
     'requirement': {'description': 'RO-Crate Metadata Descriptor MUST be defined\n    according with the requirements details defined in [RO-Crate Metadata File Descriptor](https://www.researchobject.org/ro-crate/1.1/root-data-entity.html#ro-crate-metadata-file-descriptor)',
      'identifier': 'ro-crate-1.1_5',
      'name': 'RO-Crate Metadata File Descriptor REQUIRED properties',
      'order': 5,
      'profile': {'description': 'RO-Crate Metadata Specification.',
       'identifier': 'ro-crate-1.1',
       'name': 'RO-Crate Metadata Specification 1.1',
       'uri': 'https://w3id.org/ro/crate/1.1'}},
     'severity': 'REQUIRED'},
    'message': 'The RO-Crate metadata file descriptor MUST have a `conformsTo` property with the RO-Crate specification version',
    'severity': 'REQUIRED',
    'violatingEntity': './ro-crate-metadata.json',
    'violatingProperty': 'http://purl.org/dc/terms/conformsTo',
    'violatingPropertyValue': None}],
  'meta': {'version': '0.2'},
  'passed': False,
  'validation_settings': {'abort_on_first': False,
   'enable_profile_inheritance': True,
   'profile_identifier': 'five-safes-crate-0.4',
   'requirement_severity': 'REQUIRED',
   'rocrate_validator_version': '0.11.2'}},
 'profile': 'five-safes-crate',
 'status': 'invalid'}

With this PR the above validation errors go away, but I see a strange UTF-8 error- but maybe there's something weird with my docker build?
podman build --build-arg FIVE_SAFES_PROFILE_ VERSION=five-safes-0.8.0 --build-arg PROFILES_ARCHIVE_URL=https://github.com/eScienceLab/rocrate-va lidator/archive/refs/tags/five-safes-0.8.0.tar.gz --build-arg EXTRA_PROFILES_PATH=/app/extra-profil es -t ro-crate-validation-service .

{'created_at': None,
 'detail': {'issues': [{'check': {'description': 'Check if the file descriptor is UTF-8 encoded',
     'identifier': 'ro-crate-1.2_3.1',
     'label': 'REQUIRED 3.1',
     'name': 'File Descriptor UTF-8 encoding',
     'order': 1,
     'requirement': {'description': 'The file descriptor MUST be UTF-8 encoded',
      'identifier': 'ro-crate-1.2_3',
      'name': 'File Descriptor UTF-8 encoding',
      'order': 3,
      'profile': {'description': 'RO-Crate Metadata Specification.',
       'identifier': 'ro-crate-1.2',
       'name': 'RO-Crate Metadata Specification 1.2',
       'uri': 'https://w3id.org/ro/crate/1.2'}},
     'severity': 'REQUIRED'},
    'message': 'RO-Crate file descriptor "ro-crate-metadata.json" is not UTF-8 encoded',
    'severity': 'REQUIRED',
    'violatingEntity': None,
    'violatingProperty': None,
    'violatingPropertyValue': None}],
  'meta': {'version': '0.2'},
  'passed': False,
  'validation_settings': {'abort_on_first': False,
   'enable_profile_inheritance': True,
   'profile_identifier': 'five-safes-crate-0.4',
   'requirement_severity': 'REQUIRED',
   'rocrate_validator_version': '0.11.2'}},
 'profile': 'five-safes-crate',
 'status': 'invalid'}

[elichad]

Thanks for the PR @manics ! The new error indicates that the RO-Crate metadata file you're testing isn't in UTF-8 encoding format - did the encoding get changed somehow when you downloaded the test files? (haven't tested the docker build locally but I know the original file is UTF-8 encoded)

[manics]

I'm pretty sure the encoding is correct, and it works fine if I run the validator directly
https://gist.github.com/manics/c88952351d3d74138842c2870ce19d57
which is why I'm wondering if there's something weird with my podman/docker build

[elichad]

Strange. I've just replicated this error on my side too:

• Updated images to ghcr.io/esciencelab/ro-crate-validation-service:2.0 in docker-compose.yml
• Ran docker compose up --build
• Ran jq -Rs '{crate_json: ., profile_name: "ro-crate-1.2"}' path/to/ro-crate-metadata.json | curl -X POST http://localhost:5001/v1/ro_crates/validate_metadata -H 'Content -Type: application/json' -d @- (and used the ro-crate-metadata.json directly from my local copy of the rocrate-validator repo)

Output:

{
    "created_at": null,
    "detail": {
        "issues": [
            {
                "check": {
                    "description": "Check if the file descriptor is UTF-8 encoded",
                    "identifier": "ro-crate-1.2_3.1",
                    "label": "REQUIRED 3.1",
                    "name": "File Descriptor UTF-8 encoding",
                    "order": 1,
                    "requirement": {
                        "description": "The file descriptor MUST be UTF-8 encoded",
                        "identifier": "ro-crate-1.2_3",
                        "name": "File Descriptor UTF-8 encoding",
                        "order": 3,
                        "profile": {
                            "description": "RO-Crate Metadata Specification.",
                            "identifier": "ro-crate-1.2",
                            "name": "RO-Crate Metadata Specification 1.2",
                            "uri": "https://w3id.org/ro/crate/1.2"
                        }
                    },
                    "severity": "REQUIRED"
                },
                "message": "RO-Crate file descriptor \"ro-crate-metadata.json\" is not UTF-8 encoded",
                "severity": "REQUIRED",
                "violatingEntity": null,
                "violatingProperty": null,
                "violatingPropertyValue": null
            }
        ],
        "meta": {
            "version": "0.2"
        },
        "passed": false,
        "validation_settings": {
            "abort_on_first": false,
            "enable_profile_inheritance": true,
            "profile_identifier": "ro-crate-1.2",
            "requirement_severity": "REQUIRED",
            "rocrate_validator_version": "0.11.2"
        }
    },
    "profile": "ro-crate-1.2",
    "status": "invalid"
}

[alexhambley]

Hi @manics (and @elichad!), the UTF-8 error is a false positive from an upstream bug in roc-validator 0.11.2 /11.3. I've made an issue here.

The issue is that:

• validate_metadata runs the validator in metadata-only mode, where there is no file on disk.
• In this case, the upstream validator constructs an ROCrateLocalFolder handle whose root directory points at the service process's working dir (/app). This is just a pointer.
• The ro-crate 1.2 profile added a File Descriptor UTF-8 encoding check which, unlike the check above it, doesn't skip in metadata-only mode. For example, in the FileDescriptorExistence class you have:

if context.settings.metadata_only:
            logger.debug(...)
            return True

• It tries to read ro-crate-metadata.json from disk which produces a FileNotFoundError (as I said, there's no such file under /app), and an except Exception reports that as "not UTF-8 encoded".
• The 1.1 profile has no UTF-8 check, hence no issue before.