Beispiel #1
0
def g_repositories(app, db, g_repositories_data):
    """Fixture for GitHub Repository objects."""
    for rd in g_repositories_data:
        repository = Repository(**rd)
        # We don't call GitHub hence the hook is not important, yet it should
        # be not null
        repository.hook = 12345
        db.session.add(repository)
        db.session.commit()
        rd['id'] = repository.id
    return g_repositories_data
Beispiel #2
0
def g_repositories(app, db, g_repositories_data):
    """Fixture for GitHub Repository objects."""
    for rd in g_repositories_data:
        repository = Repository(**rd)
        # We don't call GitHub hence the hook is not important, yet it should
        # be not null
        repository.hook = 12345
        db.session.add(repository)
        db.session.commit()
        rd['id'] = repository.id
    return g_repositories_data
def test_extract_metadata(app, db, tester_id, remote_token, github_api):

    Repository.enable(tester_id, github_id=2, name='repo-2', hook=1234)
    event = Event(
        receiver_id='github',
        user_id=tester_id,
        payload=fixtures.PAYLOAD('auser', 'repo-2', 2, tag='v1.0'),
    )
    release = Release.create(event)
    gh = GitHubRelease(release)
    metadata = gh.metadata

    assert metadata['upload_type'] == 'dataset'
    assert metadata['license'] == 'mit-license'
    assert len(metadata['creators']) == 2
Beispiel #4
0
def test_extract_metadata(app, db, tester_id, remote_token, github_api):

    Repository.enable(tester_id, github_id=2, name='repo-2', hook=1234)
    event = Event(
        receiver_id='github',
        user_id=tester_id,
        payload=fixtures.PAYLOAD('auser', 'repo-2', 2, tag='v1.0'),
    )
    release = Release.create(event)
    gh = GitHubRelease(release)
    metadata = gh.metadata

    assert metadata['upload_type'] == 'dataset'
    assert metadata['license'] == 'mit-license'
    assert len(metadata['creators']) == 2
Beispiel #5
0
def load_github_releases(releases_file):
    """Load GitHub releases information.

    Updates the missing releases and the submission dates.
    """
    from invenio_github.models import Release, Repository, ReleaseStatus
    from sqlalchemy.orm.exc import NoResultFound
    import arrow
    import json
    releases_db = json.load(releases_file)
    with click.progressbar(releases_db) as releases:
        for release in releases:
            repo_name, new_repo_name, gh_repo_id, ra_id, user_id, dep = release
            try:
                repo = Repository.query.filter_by(github_id=gh_repo_id).one()
            except NoResultFound:
                repo = Repository.create(user_id=user_id, github_id=gh_repo_id,
                                         name=new_repo_name)
            pid = PersistentIdentifier.get(pid_type='recid',
                                           pid_value=str(dep['record_id']))
            rel = Release.query.filter_by(
                repository_id=repo.id,
                record_id=pid.get_assigned_object()).first()
            created = arrow.get(dep['submitted']).datetime.replace(tzinfo=None)
            if rel:
                rel.created = created
            else:
                rel = Release(tag=dep['github_ref'],
                              errors=dep['errors'],
                              record_id=pid.get_assigned_object(),
                              repository_id=repo.id,
                              status=ReleaseStatus.PUBLISHED,
                              created=created)
                db.session.add(rel)
        db.session.commit()
Beispiel #6
0
def repository_model(app, db, tester_id):
    """Github repository fixture."""
    repository = Repository(
        github_id=1, name='testuser/testrepo', user_id=tester_id)
    db.session.add(repository)
    db.session.commit()
    return repository
def test_webhook_post(app, db, tester_id, location, remote_token, github_api):
    """Test payload parsing on webhook."""
    from . import fixtures
    with patch('invenio_deposit.api.Deposit.indexer'):
        # Enable repository webhook.
        Repository.enable(tester_id, github_id=3, name='arepo', hook=1234)
        db.session.commit()

        # JSON payload parsing.
        payload = json.dumps(fixtures.PAYLOAD('auser', 'arepo', 3, 'v1.0'))
        headers = [('Content-Type', 'application/json')]
        with app.test_request_context(headers=headers, data=payload):
            event = Event.create(receiver_id='github', user_id=tester_id)
            db.session.commit()
            event.process()
            db.session.commit()

        from invenio_records.models import RecordMetadata
        assert RecordMetadata.query.count() == 2
Beispiel #8
0
def test_webhook_post(app, db, tester_id, location, remote_token, github_api):
    """Test payload parsing on webhook."""
    from . import fixtures
    with patch('invenio_deposit.api.Deposit.indexer'):
        # Enable repository webhook.
        Repository.enable(tester_id, github_id=3, name='arepo', hook=1234)
        db.session.commit()

        # JSON payload parsing.
        payload = json.dumps(fixtures.PAYLOAD('auser', 'arepo', 3, 'v1.0'))
        headers = [('Content-Type', 'application/json')]
        with app.test_request_context(headers=headers, data=payload):
            event = Event.create(receiver_id='github', user_id=tester_id)
            db.session.commit()
            event.process()
            db.session.commit()

        from invenio_records.models import RecordMetadata
        assert RecordMetadata.query.count() == 2
def test_handle_payload(app, db, location, tester_id, remote_token,
                        github_api):

    from invenio_webhooks.models import Event

    extra_data = remote_token.remote_account.extra_data

    assert '1' in extra_data['repos']
    assert 'repo-1' in extra_data['repos']['1']['full_name']
    assert '2' in extra_data['repos']
    assert 'repo-2' in extra_data['repos']['2']['full_name']

    # Create the repository that will make the release

    with db.session.begin_nested():
        Repository.enable(tester_id, github_id=1, name='repo-1', hook=1234)
        event = Event(
            receiver_id='github',
            user_id=tester_id,
            payload=fixtures.PAYLOAD('auser', 'repo-1', 1)
        )
        db.session.add(event)

    with patch('invenio_deposit.api.Deposit.indexer'):
        event.process()

        repo_1 = Repository.query.filter_by(name='repo-1', github_id=1).first()
        assert repo_1.releases.count() == 1

        release = repo_1.releases.first()
        assert release.status == ReleaseStatus.PUBLISHED
        assert release.errors is None
        assert release.tag == 'v1.0'
        assert release.record is not None
        assert release.record.get('control_number') == '1'
        record_files = release.record.get('_files')
        assert len(record_files) == 1
        assert record_files[0]['size'] > 0

        bucket = Bucket.get(record_files[0]['bucket'])
        assert bucket is not None
        assert len(bucket.objects) == 1
        assert bucket.objects[0].key == 'auser/repo-1-v1.0.zip'
Beispiel #10
0
def test_handle_payload(app, db, location, tester_id, remote_token,
                        github_api):

    from invenio_webhooks.models import Event

    extra_data = remote_token.remote_account.extra_data

    assert '1' in extra_data['repos']
    assert 'repo-1' in extra_data['repos']['1']['full_name']
    assert '2' in extra_data['repos']
    assert 'repo-2' in extra_data['repos']['2']['full_name']

    # Create the repository that will make the release

    with db.session.begin_nested():
        Repository.enable(tester_id, github_id=1, name='repo-1', hook=1234)
        event = Event(receiver_id='github',
                      user_id=tester_id,
                      payload=fixtures.PAYLOAD('auser', 'repo-1', 1))
        db.session.add(event)

    with patch('invenio_deposit.api.Deposit.indexer'):
        event.process()

        repo_1 = Repository.query.filter_by(name='repo-1', github_id=1).first()
        assert repo_1.releases.count() == 1

        release = repo_1.releases.first()
        assert release.status == ReleaseStatus.PUBLISHED
        assert release.errors is None
        assert release.tag == 'v1.0'
        assert release.record is not None
        assert release.record.get('control_number') == '1'
        record_files = release.record.get('_files')
        assert len(record_files) == 1
        assert record_files[0]['size'] > 0

        bucket = Bucket.get(record_files[0]['bucket'])
        assert bucket is not None
        assert len(bucket.objects) == 1
        assert bucket.objects[0].key == 'auser/repo-1-v1.0.zip'
Beispiel #11
0
def load_github_releases(releases_file):
    """Load GitHub releases information.

    Updates the missing releases and the submission dates.
    """
    from invenio_github.models import Release, Repository, ReleaseStatus
    from sqlalchemy.orm.exc import NoResultFound
    import arrow
    import json
    releases_db = json.load(releases_file)
    with click.progressbar(releases_db) as releases:
        for release in releases:
            repo_name, new_repo_name, gh_repo_id, ra_id, user_id, dep = release
            try:
                repo = Repository.query.filter_by(github_id=gh_repo_id).one()
            except NoResultFound:
                repo = Repository.create(user_id=user_id,
                                         github_id=gh_repo_id,
                                         name=new_repo_name)
            pid = PersistentIdentifier.get(pid_type='recid',
                                           pid_value=str(dep['record_id']))
            rel = Release.query.filter_by(
                repository_id=repo.id,
                record_id=pid.get_assigned_object()).first()
            created = arrow.get(dep['submitted']).datetime.replace(tzinfo=None)
            if rel:
                rel.created = created
            else:
                rel = Release(tag=dep['github_ref'],
                              errors=dep['errors'],
                              record_id=pid.get_assigned_object(),
                              repository_id=repo.id,
                              status=ReleaseStatus.PUBLISHED,
                              created=created)
                db.session.add(rel)
        db.session.commit()
Beispiel #12
0
def test_repository_unbound(app):
    """Test create_badge method."""
    assert Repository(name='org/repo', github_id=1).latest_release() is None
Beispiel #13
0
def test_github_newversion_permissions(app, db, minimal_record, users, g_users,
                                       g_remoteaccounts):
    """Test new version creation permissions for GitHub records."""

    old_owner, new_owner = [User.query.get(u['id']) for u in g_users]

    # Create repository, and set owner to `old_owner`
    repo = Repository.create(
        name='foo/bar', github_id=8000, user_id=old_owner.id, hook=1234)

    # Create concpetrecid for the GitHub records
    conceptrecid = PersistentIdentifier.create(
        'recid', '100', status=PIDStatus.RESERVED)

    def create_deposit_and_record(pid_value, owner):
        """Utility function for creating records and deposits."""
        recid = PersistentIdentifier.create(
            'recid', pid_value, status=PIDStatus.RESERVED)
        pv = PIDVersioning(parent=conceptrecid)
        pv.insert_draft_child(recid)

        depid = PersistentIdentifier.create(
            'depid', pid_value, status=PIDStatus.REGISTERED)
        deposit = ZenodoRecord.create({'_deposit': {'id': depid.pid_value},
                                       'conceptrecid': conceptrecid.pid_value,
                                       'recid': recid.pid_value})
        deposit.commit()
        depid.assign('rec', deposit.id)

        record_metadata = deepcopy(minimal_record)
        record_metadata['_deposit'] = {'id': depid.pid_value}
        record_metadata['conceptrecid'] = conceptrecid.pid_value
        record_metadata['recid'] = int(recid.pid_value)
        record_metadata['owners'] = [owner.id]
        record = ZenodoRecord.create(record_metadata)
        zenodo_record_minter(record.id, record)
        record.commit()

        return (depid, deposit, recid, record)

    # Create first GitHub record (by `old_owner`)
    depid1, d1, recid1, r1 = create_deposit_and_record('101', old_owner)
    rel1 = Release(release_id=111, repository_id=repo.id, record_id=d1.id,
                   status=ReleaseStatus.PUBLISHED)
    db.session.add(rel1)
    db.session.commit()

    assert is_github_versioned(recid1)

    @contextmanager
    def set_identity(user):
        from flask_principal import AnonymousIdentity, Identity
        principal = current_app.extensions['security'].principal
        principal.set_identity(Identity(user))
        yield
        principal.set_identity(AnonymousIdentity())

    with app.test_request_context():
        with set_identity(old_owner):
            assert is_github_owner(old_owner, recid1)
            assert has_update_permission(old_owner, r1)
            assert has_newversion_permission(old_owner, r1)

        with set_identity(new_owner):
            assert not is_github_owner(new_owner, recid1)
            assert not has_update_permission(new_owner, r1)
            assert not has_newversion_permission(new_owner, r1)

    # Change the repository owner
    repo.user_id = new_owner.id
    db.session.add(repo)
    db.session.commit()

    with app.test_request_context():
        with set_identity(old_owner):
            assert not is_github_owner(old_owner, recid1)
            # `old_owner` can edit his record of course
            assert has_update_permission(old_owner, r1)
            assert has_newversion_permission(old_owner, r1)

        with set_identity(new_owner):
            assert is_github_owner(new_owner, recid1)
            # `new_owner` can't edit the `old_owner`'s record
            assert not has_update_permission(new_owner, r1)
            assert not has_newversion_permission(new_owner, r1)

    # Create second GitHub record (by `new_owner`)
    depid2, d2, recid2, r2 = create_deposit_and_record('102', new_owner)
    rel2 = Release(release_id=222, repository_id=repo.id, record_id=d2.id,
                   status=ReleaseStatus.PUBLISHED)
    db.session.add(rel2)
    db.session.commit()

    with app.test_request_context():
        with set_identity(old_owner):
            assert not is_github_owner(old_owner, recid1)
            assert not is_github_owner(old_owner, recid2)
            assert has_update_permission(old_owner, r1)
            # `old_owner` can't edit the `new_owner`'s record
            assert not has_update_permission(old_owner, r2)
            assert not has_newversion_permission(old_owner, r1)
            assert not has_newversion_permission(old_owner, r2)

        with set_identity(new_owner):
            assert is_github_owner(new_owner, recid1)
            assert is_github_owner(new_owner, recid2)
            assert not has_update_permission(new_owner, r1)
            # `new_owner` can edit his newly released record
            assert has_update_permission(new_owner, r2)
            assert has_newversion_permission(new_owner, r1)
            assert has_newversion_permission(new_owner, r2)

    # Create a manual record (by `new_owner`)
    depid3, d3, recid3, r3 = create_deposit_and_record('103', new_owner)
    db.session.commit()

    with app.test_request_context():
        with set_identity(old_owner):
            assert not is_github_owner(old_owner, recid3)
            assert not has_update_permission(old_owner, r3)
            assert not has_newversion_permission(old_owner, r3)

        with set_identity(new_owner):
            assert is_github_owner(new_owner, recid3)
            assert has_update_permission(new_owner, r3)
            assert has_newversion_permission(new_owner, r3)
Beispiel #14
0
def migrate_github_remote_account(gh_db_ra, remote_account_id, logger=None):
    """Migrate the GitHub remote accounts."""
    ra = RemoteAccount.query.filter_by(id=remote_account_id).first()
    for full_repo_name, repo_vals in ra.extra_data['repos'].items():
        if '/' not in full_repo_name:
            if logger is not None:
                logger.warning("Repository migrated: {name} ({id})".format(
                    name=full_repo_name, id=ra.id))
            continue
        if repo_vals['hook']:
            owner, repo_name = full_repo_name.split('/')
            # If repository name is cached, get from database, otherwise fetch
            if full_repo_name in gh_db_ra:
                gh_id, gh_full_name = gh_db_ra[full_repo_name]
            else:
                gh_api = GitHubAPI(ra.user.id)
                gh_id, gh_full_name = fetch_gh_info(full_repo_name, gh_api.api)

            try:
                repo = Repository.get(user_id=ra.user_id, github_id=gh_id,
                                      name=gh_full_name)
            except NoResultFound:
                repo = Repository.create(user_id=ra.user_id, github_id=gh_id,
                                         name=gh_full_name)
            except RepositoryAccessError as e:
                if logger is not None:

                    repo = Repository.query.filter_by(github_id=gh_id).one()
                    logger.warning(
                        "User (uid: {user_id}) repository "
                        "'{repo_name}' from remote account ID:{ra_id} has "
                        "already been claimed by another user ({user2_id})."
                        "Repository ID: {repo_id}.".format(
                            user_id=ra.user.id, repo_name=full_repo_name,
                            ra_id=ra.id, user2_id=repo.user_id,
                            repo_id=repo.id))
                continue
                # TODO: Hook for this user will not be added.
            repo.hook = repo_vals['hook']
            if repo_vals['depositions']:
                for dep in repo_vals['depositions']:
                    try:
                        pid = PersistentIdentifier.get(
                            pid_type='recid', pid_value=str(dep['record_id']))
                        release = Release.query.filter_by(
                            tag=dep['github_ref'], repository_id=repo.id,
                            record_id=pid.get_assigned_object()).first()
                        if not release:
                            release = Release(
                                tag=dep['github_ref'], errors=dep['errors'],
                                record_id=pid.get_assigned_object(),
                                repository_id=repo.id,
                                status=ReleaseStatus.PUBLISHED)
                            # TODO: DO SOMETHING WITH dep['doi']
                            # TODO: Update the date dep['submitted']
                            db.session.add(release)
                    except PIDDoesNotExistError as e:
                        if logger is not None:
                            logger.exception(
                                'Could not create release {tag} for repository'
                                ' {repo_id}, because corresponding PID: {pid} '
                                'does not exist')
                        raise e
    db.session.commit()