def get(self, uuid):
args = versioned_parser.parse_args()
version = args.get('version')
bag = database.get_bag(uuid, version=version)
if not bag:
abort(404)
bag_etag = 'bag-{}-version-{}'.format(uuid, bag['version'])
if request.if_none_match.contains_weak(bag_etag):
return None, 304
headers = {}
headers['Last-Modified'] = format_date_time(bag['created_at'])
data = json.loads(bag['data'])
defs, defs_ctx = database.get_periods_and_context(data['items'])
data['@id'] = identifier.prefix('bags/%s' % uuid)
data['creator'] = bag['created_by']
data['items'] = defs
response = api.make_response(data, 200, headers)
response.set_etag(bag_etag, weak=True)
if version is None:
return cache.no_time(response)
else:
return cache.long_time(response)
def create_or_update_bag(uuid, creator_id, data):
db = get_db()
c = get_db().cursor()
c.execute('''
SELECT MAX(version) AS max_version
FROM bag
WHERE uuid = ?''', (uuid.hex,))
row = c.fetchone()
version = 0 if row['max_version'] is None else row['max_version'] + 1
if version > 0:
data['wasRevisionOf'] = identifier.prefix('bags/{}?version={}'.format(
uuid, row['max_version']))
c.execute('''
INSERT INTO bag (
uuid,
version,
created_by,
data,
owners)
VALUES (?, ?, ?, ?, ?)''',
(uuid.hex,
version,
creator_id,
json.dumps(data),
json.dumps([creator_id])))
c.close()
db.commit()
return version
def get_item(extract_item, id, version=None):
dataset = get_dataset(version=version)
o = json.loads(dataset['data'])
item = extract_item(identifier.prefix(id), o, raiseErrors=True)
item['@context'] = o['@context']
if version is not None:
item['@context']['__version'] = version
return item
def redirect_to_last_update(entity_id, version):
if version is None:
return None
v = database.find_version_of_last_update(
identifier.prefix(entity_id), version)
if v is None:
abort(404)
if v == int(version):
return None
return redirect(request.path + '?version={}'.format(v), code=301)
def attach_to_dataset(o):
if len(o) > 0:
if app.config['CANONICAL']:
o['primaryTopicOf'] = {
'id': identifier.prefix(request.full_path[1:]),
'inDataset': {
'id': identifier.prefix('d'),
'changes': identifier.prefix('h#changes')
}
}
else:
o['primaryTopicOf'] = {
'id': request.url,
'inDataset': {
'id': url_for('abstract_dataset', _external=True),
'changes': url_for('history', _external=True) + '#changes'
}
}
return o
def get(self, definition_id):
version = request.args.get('version')
new_location = redirect_to_last_update(definition_id, version)
if new_location is not None:
return new_location
dataset = database.get_dataset(version=version)
o = json.loads(dataset['data'])
if 'periodCollections' not in o:
abort(404)
definition_key = identifier.prefix(definition_id)
collection_key = identifier.prefix(definition_id[:5])
if collection_key not in o['periodCollections']:
abort_gone_or_not_found(collection_key)
collection = o['periodCollections'][collection_key]
if definition_key not in collection['definitions']:
abort_gone_or_not_found(definition_key)
definition = collection['definitions'][definition_key]
definition['collection'] = collection_key
definition['@context'] = o['@context']
return attach_to_dataset(definition)
def create_or_update_graph(id, data):
db = get_db()
c = get_db().cursor()
c.execute('''
SELECT MAX(version) AS max_version
FROM graph
WHERE id = ?''', (id,))
row = c.fetchone()
version = 0 if row['max_version'] is None else row['max_version'] + 1
if version > 0:
data['wasRevisionOf'] = identifier.prefix(
'graphs/{}?version={}'.format(id, row['max_version']))
c.execute('''
INSERT INTO graph (
id,
version,
data)
VALUES (?, ?, ?)''',
(id,
version,
json.dumps(data)))
c.close()
db.commit()
return version
def make_nanopub(period_id, version):
cursor = database.get_db().cursor()
cursor.execute(
'''
SELECT
patch.id as patch_id,
patch.merged_at,
patch.merged_by,
patch.created_by,
dataset.data
FROM patch_request AS patch
LEFT JOIN dataset ON patch.resulted_in = dataset.id
WHERE
patch.created_entities LIKE ?
OR
patch.updated_entities LIKE ?
ORDER BY patch.id ASC
LIMIT ?, 1;
''',
('%"' + identifier.prefix(period_id) + '"%',
'%"' + identifier.prefix(period_id) + '"%',
version - 1)
)
result = cursor.fetchone()
if not result:
raise PeriodNotFoundError(
'Could not find version {} of period {}'.format(
version, period_id))
data = json.loads(result['data'])
authority_id = identifier.prefix(
period_id[:identifier.AUTHORITY_SEQUENCE_LENGTH + 1])
authority = data['authorities'][authority_id]
source = authority['source']
period = authority['periods'][identifier.prefix(period_id)]
period['authority'] = authority_id
nanopub_uri = '{}/nanopub{}'.format(
identifier.prefix(period_id), version)
patch_uri = identifier.prefix('h#change-{}'.format(result['patch_id']))
context = data['@context'].copy()
context['np'] = 'http://nanopub.org/nschema#'
context['pub'] = data['@context']['@base'] + nanopub_uri + '#'
context['prov'] = 'http://www.w3.org/ns/prov#'
# TODO: Pop "source" from period and include it in the provenance
# graph?
return {
"@context": context,
"@graph": [
{
"@id": "pub:head",
"@graph": {
"@id": nanopub_uri,
"@type": "np:Nanopublication",
"np:hasAssertion": as_uri("pub:assertion"),
"np:hasProvenance": as_uri("pub:provenance"),
"np:hasPublicationInfo": as_uri("pub:pubinfo"),
}
},
{
"@id": "pub:assertion",
"@graph": [period]
},
{
"@id": "pub:provenance",
"@graph": [
{
"@id": 'pub:assertion',
"dc:source": source
}
]
},
{
"@id": "pub:pubinfo",
"@graph": [
{
"@id": nanopub_uri,
"prov:wasGeneratedBy": as_uri(patch_uri),
"prov:asGeneratedAtTime": result['merged_at'],
"prov:wasAttributedTo": [
as_uri(result['merged_by']),
as_uri(result['created_by'])
]
}
]
}
]
}
def attach_to_dataset(o):
o['primaryTopicOf'] = {'id': identifier.prefix(request.path[1:]),
'inDataset': identifier.prefix('d')}
return o
def history():
g = Graph()
changelog = Collection(g, URIRef("#changelog"))
cursor = database.get_db().cursor()
for row in cursor.execute(
"""
SELECT
id,
created_at,
created_by,
updated_by,
merged_at,
merged_by,
applied_to,
resulted_in,
created_entities,
updated_entities,
removed_entities
FROM patch_request
WHERE merged = 1
ORDER BY id ASC
"""
).fetchall():
change = URIRef("#change-{}".format(row["id"]))
patch = URIRef("#patch-{}".format(row["id"]))
g.add((patch, FOAF.page, PERIODO[identifier.prefix(url_for("patch", id=row["id"]))]))
g.add((change, PROV.startedAtTime, Literal(utils.isoformat(row["created_at"]), datatype=XSD.dateTime)))
g.add((change, PROV.endedAtTime, Literal(utils.isoformat(row["merged_at"]), datatype=XSD.dateTime)))
dataset = PERIODO[identifier.prefix(url_for("abstract_dataset"))]
version_in = PERIODO[identifier.prefix(url_for("abstract_dataset", version=row["applied_to"]))]
g.add((version_in, PROV.specializationOf, dataset))
version_out = PERIODO[identifier.prefix(url_for("abstract_dataset", version=row["resulted_in"]))]
g.add((version_out, PROV.specializationOf, dataset))
g.add((change, PROV.used, version_in))
g.add((change, PROV.used, patch))
g.add((change, PROV.generated, version_out))
def add_entity_version(entity_id):
entity = PERIODO[entity_id]
entity_version = PERIODO[entity_id + "?version={}".format(row["resulted_in"])]
g.add((entity_version, PROV.specializationOf, entity))
g.add((change, PROV.generated, entity_version))
return entity_version
for entity_id in json.loads(row["created_entities"]):
add_entity_version(entity_id)
for entity_id in json.loads(row["updated_entities"]):
entity_version = add_entity_version(entity_id)
prev_entity_version = PERIODO[entity_id + "?version={}".format(row["applied_to"])]
g.add((entity_version, PROV.wasRevisionOf, prev_entity_version))
for entity_id in json.loads(row["removed_entities"]):
g.add((change, PROV.invalidated, PERIODO[entity_id]))
for field, term in (("created_by", "submitted"), ("updated_by", "updated"), ("merged_by", "merged")):
if row[field] == "initial-data-loader":
continue
agent = URIRef(row[field])
association = URIRef("#patch-{}-{}".format(row["id"], term))
g.add((change, PROV.wasAssociatedWith, agent))
g.add((change, PROV.qualifiedAssociation, association))
g.add((association, PROV.agent, agent))
g.add((association, PROV.hadRole, PERIODO[identifier.prefix(url_for("vocab") + "#" + term)]))
changelog.append(change)
def ordering(o):
if o["@id"] == "#changelog":
# sort first
return " "
return o["@id"]
jsonld = json.loads(g.serialize(format="json-ld", context=CONTEXT).decode("utf-8"))
jsonld["history"] = sorted(jsonld["history"], key=ordering)
return json.dumps(jsonld, sort_keys=True)
