mirror of
https://github.com/internetarchive/warcprox.git
synced 2025-01-18 13:22:09 +01:00
Revert warc to previous behavior
If record_id is available, write it to REFERS_TO header.
This commit is contained in:
parent
ad8ba43c3d
commit
424f236126
@ -50,6 +50,7 @@ class WarcRecordBuilder:
|
|||||||
url=recorded_url.url, warc_date=warc_date,
|
url=recorded_url.url, warc_date=warc_date,
|
||||||
data=response_header_block,
|
data=response_header_block,
|
||||||
warc_type=warctools.WarcRecord.REVISIT,
|
warc_type=warctools.WarcRecord.REVISIT,
|
||||||
|
refers_to=recorded_url.dedup_info.get('id'),
|
||||||
refers_to_target_uri=recorded_url.dedup_info['url'],
|
refers_to_target_uri=recorded_url.dedup_info['url'],
|
||||||
refers_to_date=recorded_url.dedup_info['date'],
|
refers_to_date=recorded_url.dedup_info['date'],
|
||||||
payload_digest=warcprox.digest_str(recorded_url.response_recorder.payload_digest, self.base32),
|
payload_digest=warcprox.digest_str(recorded_url.response_recorder.payload_digest, self.base32),
|
||||||
@ -86,8 +87,8 @@ class WarcRecordBuilder:
|
|||||||
|
|
||||||
def build_warc_record(self, url, warc_date=None, recorder=None, data=None,
|
def build_warc_record(self, url, warc_date=None, recorder=None, data=None,
|
||||||
concurrent_to=None, warc_type=None, content_type=None, remote_ip=None,
|
concurrent_to=None, warc_type=None, content_type=None, remote_ip=None,
|
||||||
profile=None, refers_to_target_uri=None, refers_to_date=None,
|
profile=None, refers_to=None, refers_to_target_uri=None,
|
||||||
payload_digest=None):
|
refers_to_date=None, payload_digest=None):
|
||||||
|
|
||||||
if warc_date is None:
|
if warc_date is None:
|
||||||
warc_date = warctools.warc.warc_datetime_str(datetime.datetime.utcnow())
|
warc_date = warctools.warc.warc_datetime_str(datetime.datetime.utcnow())
|
||||||
@ -104,6 +105,8 @@ class WarcRecordBuilder:
|
|||||||
headers.append((warctools.WarcRecord.IP_ADDRESS, remote_ip))
|
headers.append((warctools.WarcRecord.IP_ADDRESS, remote_ip))
|
||||||
if profile is not None:
|
if profile is not None:
|
||||||
headers.append((warctools.WarcRecord.PROFILE, profile))
|
headers.append((warctools.WarcRecord.PROFILE, profile))
|
||||||
|
if refers_to is not None:
|
||||||
|
headers.append((warctools.WarcRecord.REFERS_TO, refers_to))
|
||||||
if refers_to_target_uri is not None:
|
if refers_to_target_uri is not None:
|
||||||
headers.append((warctools.WarcRecord.REFERS_TO_TARGET_URI, refers_to_target_uri))
|
headers.append((warctools.WarcRecord.REFERS_TO_TARGET_URI, refers_to_target_uri))
|
||||||
if refers_to_date is not None:
|
if refers_to_date is not None:
|
||||||
|
Loading…
x
Reference in New Issue
Block a user