backup/warcprox
1
0
Fork 0
mirror of https://github.com/internetarchive/warcprox.git synced 2025-01-18 13:22:09 +01:00
Code Issues Projects Releases Activity

use rethinkdb native time type for captures table timestamp

Browse Source
This commit is contained in:
Noah Levitt 2015-11-18 02:00:48 +00:00
parent df31068c80
commit 927419645b
3 changed files with 14 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
warcprox/bigtable.py
View File

@ -10,6 +10,7 @@ import os
import hashlib
import threading
import datetime
import rethinkstuff
class RethinkCaptures:
"""Inserts in batches every 0.5 seconds"""
@ -97,8 +98,7 @@ class RethinkCaptures:
"id": "{} {}".format(canon_surt[:20], records[0].id.decode("utf-8")[10:-1]),
"abbr_canon_surt": canon_surt[:150],
"canon_surt": canon_surt,
# "timestamp": re.sub(r"[^0-9]", "", records[0].date.decode("utf-8")),
"timestamp": records[0].date.decode("utf-8"),
"timestamp": recorded_url.timestamp.replace(tzinfo=rethinkstuff.UTC),
"url": recorded_url.url.decode("utf-8"),
"offset": records[0].offset,
"filename": os.path.basename(records[0].warc_filename),
@ -140,7 +140,11 @@ class RethinkCapturesDedup:
raw_digest = base64.b16decode(value_str, casefold=True)
entry = self.captures_db.find_response_by_digest(algo, raw_digest, bucket)
if entry:
dedup_info = {"url":entry["url"].encode("utf-8"), "date":entry["timestamp"].encode("utf-8"), "id":entry["warc_id"].encode("utf-8")}
dedup_info = {
"url": entry["url"].encode("utf-8"),
"date": entry["timestamp"].strftime("%Y-%m-%dT%H:%M:%SZ").encode("utf-8"),
"id": entry["warc_id"].encode("utf-8")
}
return dedup_info
else:
return None

2
warcprox/dedup.py
View File

@ -65,7 +65,7 @@ class DedupDb(object):
if (records[0].get_header(warctools.WarcRecord.TYPE) == warctools.WarcRecord.RESPONSE
and recorded_url.response_recorder.payload_size() > 0):
digest_key = warcprox.digest_str(recorded_url.response_recorder.payload_digest,
self.options.base32)
self.options.base32)
if recorded_url.warcprox_meta and "captures-bucket" in recorded_url.warcprox_meta:
self.save(digest_key, records[0], bucket=recorded_url.warcprox_meta["captures-bucket"])
else: