def find(cube, query, fields=None, date=None, most_recent=True):
logger.debug('Running Find')
if date is not None:
# we will be doing a timeline query so we need to rename the fields
# WARNING: might not work if some field is a substring of other field
all_fields = get_fields(cube, '__all__')
for f in all_fields:
query = re.sub(f, 'fields.%s' % f, query)
# add the date constraint
query = query + ' and ' + _get_date_pql_string(date)
pql_parser = pql.SchemaFreeParser()
try:
# FIXME: make it a schema aware parser
spec = pql_parser.parse(query)
except Exception as e:
raise ValueError("Invalid Query (%s)" % str(e))
c = get_cube(cube)
_cube = c.get_collection(timeline=(date is not None))
logger.debug('Query: %s' % spec)
fields = get_fields(cube, fields)
if date is not None:
project_d = dict([(f, '$fields.%s' % f) for f in fields])
project_d.update(dict(_id='$id', _start='$start', _end='$end'))
if most_recent:
docs = _cube.aggregate([{'$match': spec},
{'$sort': {'start': -1}},
{'$group': {'_id': '$id',
'fields': {'$first':
'$fields'},
'start': {'$first': '$start'},
'end': {'$first': '$end'},
'id': {'$first': '$id'}}},
{'$project': project_d}])
else:
docs = _cube.aggregate([{'$match': spec},
{'$project': project_d}])
docs = docs['result']
else:
docs = _cube.find(spec, fields)
docs.batch_size(10000000) # hard limit is 16M...
docs = [d for d in docs]
return docs
def index_warehouse(cube, fields, force=False):
'''
NOTE: _id key index is generated automatically by mongo
'''
c = get_cube(cube)
_cube = c.get_collection(admin=True)
fields = get_fields(cube, fields)
result = {}
for field in fields:
name = '%s-tokens' % field
if force or c.get_field_property('index', field):
logger.info(' %s... Indexing Warehouse (%s)%s' %
(YELLOW, field, ENDC))
key = [(field, -1)]
result[field] = _cube.ensure_index(key, name=name)
else:
result[field] = -1
return result
def extract(cube, **kwargs):
logger.info(' Starting Update operation!')
logger.info(' %sCube: %s%s' % (YELLOW, cube, ENDC))
c = get_cube(cube)
logger.debug('%sExtract - Start%s' % (YELLOW, ENDC))
_fields = kwargs.get('fields')
fields = get_fields(cube, _fields)
if fields:
result = {}
for field in fields:
kwargs['field'] = field
logger.debug('%sField: %s%s' % (YELLOW, field, ENDC))
result[field] = c.extract_func(**kwargs)
logger.info('Extract - Complete: (%s.%s): %s' %
(cube, field, result[field]))
else:
result = c.extract_func(**kwargs)
logger.info('Extract - Complete: (%s): %s' % (cube, result))
return result
def fetch(cube, fields, skip=0, limit=0, ids=None):
logger.debug('Running Fetch (skip:%s, limit:%s, ids:%s)' % (
skip, limit, len(ids)))
logger.debug('... Fields: %s' % fields)
c = get_cube(cube)
_cube = c.get_collection()
fields = get_fields(cube, fields)
logger.debug('Return Fields: %s' % fields)
sort = [('_id', 1)]
if ids:
spec = {'_id': {'$in': parse_ids(ids)}}
else:
spec = {}
docs = _cube.find(spec, fields,
skip=skip, limit=limit,
sort=sort)
docs.batch_size(10000000) # hard limit is 16M...
return [d for d in docs]