def _sort(self, data_dict, fields_types):
sort = data_dict.get('sort')
if not sort:
q = data_dict.get('q')
if q:
if isinstance(q, basestring):
return [self._ts_rank_alias()]
elif isinstance(q, dict):
return [
self._ts_rank_alias(field) for field in q
if field not in fields_types
]
else:
return []
clauses = datastore_helpers.get_list(sort, False)
clause_parsed = []
for clause in clauses:
field, sort = self._parse_sort_clause(clause, fields_types)
clause_parsed.append(u'{0} {1}'.format(
datastore_helpers.identifier(field), sort))
return clause_parsed
def datastore_search(self, context, data_dict, fields_types, query_dict):
fields = data_dict.get('fields')
if fields:
field_ids = datastore_helpers.get_list(fields)
else:
field_ids = fields_types.keys()
ts_query, rank_column = self._textsearch_query(data_dict)
limit = data_dict.get('limit', 100)
offset = data_dict.get('offset', 0)
sort = self._sort(data_dict, fields_types)
where = self._where(data_dict, fields_types)
select_cols = [u'"{0}"'.format(field_id) for field_id in field_ids] +\
[u'count(*) over() as "_full_count" %s' % rank_column]
query_dict['distinct'] = data_dict.get('distinct', False)
query_dict['select'] += select_cols
query_dict['ts_query'] = ts_query
query_dict['sort'] += sort
query_dict['where'] += where
query_dict['limit'] = limit
query_dict['offset'] = offset
return query_dict
def datastore_search(self, context, data_dict, fields_types, query_dict):
fields = data_dict.get('fields')
if fields:
field_ids = datastore_helpers.get_list(fields)
else:
field_ids = fields_types.keys()
ts_query, rank_column = self._textsearch_query(data_dict)
limit = data_dict.get('limit', 100)
offset = data_dict.get('offset', 0)
sort = self._sort(data_dict, fields_types)
where = self._where(data_dict, fields_types)
select_cols = [u'"{0}"'.format(field_id) for field_id in field_ids] +\
[u'count(*) over() as "_full_count" %s' % rank_column]
query_dict['distinct'] = data_dict.get('distinct', False)
query_dict['select'] += select_cols
query_dict['ts_query'] = ts_query
query_dict['sort'] += sort
query_dict['where'] += where
query_dict['limit'] = limit
query_dict['offset'] = offset
return query_dict
def create_alias(context, data_dict):
aliases = datastore_helpers.get_list(data_dict.get('aliases'))
if aliases is not None:
# delete previous aliases
previous_aliases = _get_aliases(context, data_dict)
for alias in previous_aliases:
sql_alias_drop_string = u'DROP VIEW "{0}"'.format(alias)
context['connection'].execute(sql_alias_drop_string)
try:
for alias in aliases:
sql_alias_string = u'''CREATE VIEW "{alias}"
AS SELECT * FROM "{main}"'''.format(
alias=alias,
main=data_dict['resource_id']
)
res_ids = _get_resources(context, alias)
if res_ids:
raise ValidationError({
'alias': [(u'The alias "{0}" already exists.').format(
alias)]
})
context['connection'].execute(sql_alias_string)
except DBAPIError, e:
if e.orig.pgcode in [_PG_ERR_CODE['duplicate_table'],
_PG_ERR_CODE['duplicate_alias']]:
raise ValidationError({
'alias': [u'"{0}" already exists'.format(alias)]
})
def create_alias(context, data_dict):
aliases = datastore_helpers.get_list(data_dict.get('aliases'))
if aliases is not None:
# delete previous aliases
previous_aliases = _get_aliases(context, data_dict)
for alias in previous_aliases:
sql_alias_drop_string = u'DROP VIEW "{0}"'.format(alias)
context['connection'].execute(sql_alias_drop_string)
try:
for alias in aliases:
sql_alias_string = u'''CREATE VIEW "{alias}"
AS SELECT * FROM "{main}"'''.format(
alias=alias, main=data_dict['resource_id'])
res_ids = _get_resources(context, alias)
if res_ids:
raise ValidationError({
'alias':
[(u'The alias "{0}" already exists.').format(alias)]
})
context['connection'].execute(sql_alias_string)
except DBAPIError, e:
if e.orig.pgcode in [
_PG_ERR_CODE['duplicate_table'],
_PG_ERR_CODE['duplicate_alias']
]:
raise ValidationError(
{'alias': [u'"{0}" already exists'.format(alias)]})
def validate(context, data_dict):
fields_types = _get_fields_types(context, data_dict)
data_dict_copy = copy.deepcopy(data_dict)
# TODO: Convert all attributes that can be a comma-separated string to
# lists
if 'fields' in data_dict_copy:
fields = datastore_helpers.get_list(data_dict_copy['fields'])
data_dict_copy['fields'] = fields
if 'sort' in data_dict_copy:
fields = datastore_helpers.get_list(data_dict_copy['sort'], False)
data_dict_copy['sort'] = fields
for plugin in p.PluginImplementations(interfaces.ITimeseries):
data_dict_copy = plugin.datastore_validate(context,
data_dict_copy,
fields_types)
# Remove default elements in data_dict
del data_dict_copy['connection_url']
del data_dict_copy['resource_id']
data_dict_copy.pop('id', None)
for key, values in data_dict_copy.iteritems():
if not values:
continue
if isinstance(values, basestring):
value = values
elif isinstance(values, (list, tuple)):
value = values[0]
elif isinstance(values, dict):
value = values.keys()[0]
else:
value = values
raise ValidationError({
key: [u'invalid value "{0}"'.format(value)]
})
return True
def validate(context, data_dict):
fields_types = _get_fields_types(context, data_dict)
data_dict_copy = copy.deepcopy(data_dict)
# TODO: Convert all attributes that can be a comma-separated string to
# lists
if 'fields' in data_dict_copy:
fields = datastore_helpers.get_list(data_dict_copy['fields'])
data_dict_copy['fields'] = fields
if 'sort' in data_dict_copy:
fields = datastore_helpers.get_list(data_dict_copy['sort'], False)
data_dict_copy['sort'] = fields
for plugin in p.PluginImplementations(interfaces.ITimeseries):
data_dict_copy = plugin.datastore_validate(context, data_dict_copy,
fields_types)
# Remove default elements in data_dict
del data_dict_copy['connection_url']
del data_dict_copy['resource_id']
data_dict_copy.pop('id', None)
for key, values in data_dict_copy.iteritems():
if not values:
continue
if isinstance(values, basestring):
value = values
elif isinstance(values, (list, tuple)):
value = values[0]
elif isinstance(values, dict):
value = values.keys()[0]
else:
value = values
raise ValidationError({key: [u'invalid value "{0}"'.format(value)]})
return True
def _sort(self, data_dict, fields_types):
sort = data_dict.get('sort')
if not sort:
q = data_dict.get('q')
if q:
if isinstance(q, basestring):
return [self._ts_rank_alias()]
elif isinstance(q, dict):
return [self._ts_rank_alias(field) for field in q
if field not in fields_types]
else:
return []
clauses = datastore_helpers.get_list(sort, False)
clause_parsed = []
for clause in clauses:
field, sort = self._parse_sort_clause(clause, fields_types)
clause_parsed.append(
u'{0} {1}'.format(datastore_helpers.identifier(field), sort))
return clause_parsed
def datastore_create(context, data_dict):
'''Adds a new table to the DataStore.
The datastore_create action allows you to post JSON data to be
stored against a resource. This endpoint also supports altering tables,
aliases and indexes and bulk insertion. This endpoint can be called multiple
times to initially insert more data, add fields, change the aliases or indexes
as well as the primary keys.
To create an empty datastore resource and a CKAN resource at the same time,
provide ``resource`` with a valid ``package_id`` and omit the ``resource_id``.
If you want to create a datastore resource from the content of a file,
provide ``resource`` with a valid ``url``.
See :ref:`fields` and :ref:`records` for details on how to lay out records.
:param resource_id: resource id that the data is going to be stored against.
:type resource_id: string
:param force: set to True to edit a read-only resource
:type force: bool (optional, default: False)
:param resource: resource dictionary that is passed to
:meth:`~ckan.logic.action.create.resource_create`.
Use instead of ``resource_id`` (optional)
:type resource: dictionary
:param aliases: names for read only aliases of the resource. (optional)
:type aliases: list or comma separated string
:param fields: fields/columns and their extra metadata. (optional)
:type fields: list of dictionaries
:param records: the data, eg: [{"dob": "2005", "some_stuff": ["a", "b"]}] (optional)
:type records: list of dictionaries
:param primary_key: fields that represent a unique key (optional)
:type primary_key: list or comma separated string
:param indexes: indexes on table (optional)
:type indexes: list or comma separated string
Please note that setting the ``aliases``, ``indexes`` or ``primary_key`` replaces the exising
aliases or constraints. Setting ``records`` appends the provided records to the resource.
**Results:**
:returns: The newly created data object.
:rtype: dictionary
See :ref:`fields` and :ref:`records` for details on how to lay out records.
'''
schema = context.get('schema', dsschema.datastore_create_schema())
records = data_dict.pop('records', None)
resource = data_dict.pop('resource', None)
data_dict, errors = _validate(data_dict, schema, context)
resource_dict = None
if records:
data_dict['records'] = records
if resource:
data_dict['resource'] = resource
if errors:
raise p.toolkit.ValidationError(errors)
p.toolkit.check_access('datastore_ts_create', context, data_dict)
if 'resource' in data_dict and 'resource_id' in data_dict:
raise p.toolkit.ValidationError({
'resource': ['resource cannot be used with resource_id']
})
if not 'resource' in data_dict and not 'resource_id' in data_dict:
raise p.toolkit.ValidationError({
'resource_id': ['resource_id or resource required']
})
if 'resource' in data_dict:
has_url = 'url' in data_dict['resource']
if 'retention' in data_dict['resource']:
try:
retention = int(data_dict['resource']['retention'])
if retention < 1 or retention > 100:
raise Exception()
except:
raise p.toolkit.ValidationError({'resource': [
'Retention must be an integer from 1-100']})
# A datastore only resource does not have a url in the db
data_dict['resource'].setdefault('url', '_datastore_only_resource')
resource_dict = p.toolkit.get_action('resource_create')(
context, data_dict['resource'])
data_dict['resource_id'] = resource_dict['id']
# create resource from file
if has_url:
if not p.plugin_loaded('datapusher'):
raise p.toolkit.ValidationError({'resource': [
'The datapusher has to be enabled.']})
p.toolkit.get_action('datapusher_submit')(context, {
'resource_id': resource_dict['id'],
'set_url_type': True
})
# since we'll overwrite the datastore resource anyway, we
# don't need to create it here
return
# create empty resource
else:
# no need to set the full url because it will be set in before_show
resource_dict['url_type'] = 'datastore'
p.toolkit.get_action('resource_update')(context, resource_dict)
else:
if not data_dict.pop('force', False):
resource_id = data_dict['resource_id']
_check_read_only(context, resource_id)
data_dict['connection_url'] = config['ckan.datastore.write_url']
# validate aliases
aliases = datastore_helpers.get_list(data_dict.get('aliases', []))
for alias in aliases:
if not db._is_valid_table_name(alias):
raise p.toolkit.ValidationError({
'alias': [u'"{0}" is not a valid alias name'.format(alias)]
})
# create a private datastore resource, if necessary
model = _get_or_bust(context, 'model')
resource = model.Resource.get(data_dict['resource_id'])
legacy_mode = 'ckan.datastore.read_url' not in config
if not legacy_mode and resource.package.private:
data_dict['private'] = True
try:
result = db.create(context, data_dict)
except db.InvalidDataError as err:
raise p.toolkit.ValidationError(unicode(err))
# Set the datastore_active flag on the resource if necessary
if resource.extras.get('datastore_active') is not True:
log.debug(
'Setting datastore_active=True on resource {0}'.format(resource.id)
)
# issue #3245: race condition
update_dict = {'datastore_active': True}
# get extras(for entity update) and package_id(for search index update)
res_query = model.Session.query(
model.resource_table.c.extras,
model.resource_table.c.package_id
).filter(
model.Resource.id == data_dict['resource_id']
)
extras, package_id = res_query.one()
# update extras in database for record and its revision
extras.update(update_dict)
res_query.update({'extras': extras}, synchronize_session=False)
model.Session.query(model.resource_revision_table).filter(
model.ResourceRevision.id == data_dict['resource_id'],
model.ResourceRevision.current is True
).update({'extras': extras}, synchronize_session=False)
model.Session.commit()
# get package with updated resource from solr
# find changed resource, patch it and reindex package
psi = search.PackageSearchIndex()
solr_query = search.PackageSearchQuery()
q = {
'q': 'id:"{0}"'.format(package_id),
'fl': 'data_dict',
'wt': 'json',
'fq': 'site_id:"%s"' % config.get('ckan.site_id'),
'rows': 1
}
for record in solr_query.run(q)['results']:
solr_data_dict = json.loads(record['data_dict'])
for resource in solr_data_dict['resources']:
if resource['id'] == data_dict['resource_id']:
resource.update(update_dict)
psi.index_package(solr_data_dict)
break
result.pop('id', None)
result.pop('private', None)
result.pop('connection_url')
datastore_helpers.remove_autogen(result)
return result
def create_indexes(context, data_dict):
connection = context['connection']
indexes = datastore_helpers.get_list(data_dict.get('indexes'))
# primary key is not a real primary key
# it's just a unique key
primary_key = datastore_helpers.get_list(data_dict.get('primary_key'))
sql_index_tmpl = u'CREATE {unique} INDEX "{name}" ON "{res_id}"'
sql_index_string_method = sql_index_tmpl + u' USING {method}({fields})'
sql_index_string = sql_index_tmpl + u' ({fields})'
sql_index_strings = []
fields = _get_fields(context, data_dict)
field_ids = _pluck('id', fields)
json_fields = [x['id'] for x in fields if x['type'] == 'nested']
fts_indexes = _build_fts_indexes(connection,
data_dict,
sql_index_string_method,
fields)
sql_index_strings = sql_index_strings + fts_indexes
if indexes is not None:
_drop_indexes(context, data_dict, False)
else:
indexes = []
if primary_key is not None:
_drop_indexes(context, data_dict, True)
indexes.append(primary_key)
for index in indexes:
if not index:
continue
index_fields = datastore_helpers.get_list(index)
for field in index_fields:
if field not in field_ids:
raise ValidationError({
'index': [
u'The field "{0}" is not a valid column name.'.format(
index)]
})
fields_string = u', '.join(
['(("{0}").json::text)'.format(field)
if field in json_fields else
'"%s"' % field
for field in index_fields])
sql_index_strings.append(sql_index_string.format(
res_id=data_dict['resource_id'],
unique='unique' if index == primary_key else '',
name=_generate_index_name(data_dict['resource_id'], fields_string),
fields=fields_string))
sql_index_strings = map(lambda x: x.replace('%', '%%'), sql_index_strings)
current_indexes = _get_index_names(context['connection'],
data_dict['resource_id'])
for sql_index_string in sql_index_strings:
has_index = [c for c in current_indexes
if sql_index_string.find(c) != -1]
if not has_index:
connection.execute(sql_index_string)
create_timestamp_index(context, data_dict)
def create_indexes(context, data_dict):
connection = context['connection']
indexes = datastore_helpers.get_list(data_dict.get('indexes'))
# primary key is not a real primary key
# it's just a unique key
primary_key = datastore_helpers.get_list(data_dict.get('primary_key'))
sql_index_tmpl = u'CREATE {unique} INDEX "{name}" ON "{res_id}"'
sql_index_string_method = sql_index_tmpl + u' USING {method}({fields})'
sql_index_string = sql_index_tmpl + u' ({fields})'
sql_index_strings = []
fields = _get_fields(context, data_dict)
field_ids = _pluck('id', fields)
json_fields = [x['id'] for x in fields if x['type'] == 'nested']
fts_indexes = _build_fts_indexes(connection, data_dict,
sql_index_string_method, fields)
sql_index_strings = sql_index_strings + fts_indexes
if indexes is not None:
_drop_indexes(context, data_dict, False)
else:
indexes = []
if primary_key is not None:
_drop_indexes(context, data_dict, True)
indexes.append(primary_key)
for index in indexes:
if not index:
continue
index_fields = datastore_helpers.get_list(index)
for field in index_fields:
if field not in field_ids:
raise ValidationError({
'index': [
u'The field "{0}" is not a valid column name.'.format(
index)
]
})
fields_string = u', '.join([
'(("{0}").json::text)'.format(field)
if field in json_fields else '"%s"' % field
for field in index_fields
])
sql_index_strings.append(
sql_index_string.format(
res_id=data_dict['resource_id'],
unique='unique' if index == primary_key else '',
name=_generate_index_name(data_dict['resource_id'],
fields_string),
fields=fields_string))
sql_index_strings = map(lambda x: x.replace('%', '%%'), sql_index_strings)
current_indexes = _get_index_names(context['connection'],
data_dict['resource_id'])
for sql_index_string in sql_index_strings:
has_index = [
c for c in current_indexes if sql_index_string.find(c) != -1
]
if not has_index:
connection.execute(sql_index_string)
create_timestamp_index(context, data_dict)