discard_names = set(discard_names)
else:
discard_names = set()
if not os.path.exists(kt_filecache):
os.makedirs(kt_filecache)
kt_appid = Babe.get_config_with_env("kontagent", "KT_APPID", kwargs)
for hour in enumerate_period_per_hour(start_time, end_time, referent_timezone):
url = get_url(hour, kt_user, kt_pass, kt_appid)
log.info("Kontagent: retrieving list: %s" % url)
s = urllib.urlopen(url).read()
if s == "No files available":
continue
file_urls = json.loads(s)
if sample_mode and len(file_urls) > 0:
# Sample mode: just process the first file.
file_urls = file_urls[:1]
p = Pool(8)
downloaded_files = p.map(lambda url: read_url_with_cache(url, kt_user, kt_pass, kt_filecache), file_urls)
p.close()
header = kt_msg.replace(partition=[("date", datetime.date(hour.year, hour.month, hour.day)), ("hour", hour.hour)])
yield header
gzips = [Popen(['gzip', '-d', '-c', f], stdin=PIPE, stdout=PIPE) for f in downloaded_files]
for gzip in gzips:
for row in process_file(hour, gzip.stdout, discard_names):
yield row
gzip.stdin.close()
gzip.wait()
yield StreamFooter()
Babe.register("pull_kontagent", pull_kontagent)
from pybabe import Babe, StreamHeader, StreamMeta
def unpivot(stream, common_fields,
unpivot_name_field, unpivot_value_field):
"""Unpivot a table. Keep fields lines, use other as values"""
for row in stream:
if isinstance(row, StreamHeader):
header = row.replace(
fields=common_fields + [unpivot_name_field, unpivot_value_field])
other_fields = [field for field in row.fields
if not field in common_fields]
yield header
elif isinstance(row, StreamMeta):
yield row
else:
commons = [getattr(row, StreamHeader.keynormalize(f)) for f in common_fields]
for field in other_fields:
yield header.t._make(commons + [field, getattr(row, StreamHeader.keynormalize(field))])
Babe.register('unpivot', unpivot)
from pybabe import Babe, StreamHeader, StreamMeta
def unpivot(stream, common_fields, unpivot_name_field, unpivot_value_field):
"""Unpivot a table. Keep fields lines, use other as values"""
for row in stream:
if isinstance(row, StreamHeader):
header = row.replace(fields=common_fields +
[unpivot_name_field, unpivot_value_field])
other_fields = [
field for field in row.fields if not field in common_fields
]
yield header
elif isinstance(row, StreamMeta):
yield row
else:
commons = [
getattr(row, StreamHeader.keynormalize(f))
for f in common_fields
]
for field in other_fields:
yield header.t._make(
commons +
[field,
getattr(row, StreamHeader.keynormalize(field))])
Babe.register('unpivot', unpivot)
s = urllib.urlopen(url).read()
if s == "No files available":
continue
file_urls = json.loads(s)
if sample_mode and len(
file_urls) > 0: # Sample mode: just process the first file.
file_urls = file_urls[:1]
p = Pool(8)
downloaded_files = p.map(
lambda url: read_url_with_cache(url, kt_user, kt_pass, kt_filecache
), file_urls)
p.close()
header = kt_msg.replace(partition=[(
"date",
datetime.date(hour.year, hour.month, hour.day)), ("hour",
hour.hour)])
yield header
gzips = [
Popen(['gzip', '-d', '-c', f], stdin=PIPE, stdout=PIPE)
for f in downloaded_files
]
for gzip in gzips:
for row in process_file(hour, gzip.stdout, discard_names):
yield row
gzip.stdin.close()
gzip.wait()
yield StreamFooter()
Babe.register("pull_kontagent", pull_kontagent)
