backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-15 00:03:28 +01:00
Code Issues Projects Releases Activity

Add configuration options for dedup (#597)

Browse Source 
* Add configuration options for dedup

Signed-off-by: Lukas Straub <lukasstraub2@web.de>

* Add documentation for new dedup_index configuration options

Signed-off-by: Lukas Straub <lukasstraub2@web.de>
This commit is contained in:
Lukey3332 2021-01-27 02:06:18 +01:00 committed by GitHub
parent 04d0586244
commit ddf3207e40
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 43 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
docs/manual/configuring.rst
View File

@ -267,7 +267,6 @@ The full set of configurable options (with their default settings) is as follows
filename_template: my-warc-{timestamp}-{hostname}-{random}.warc.gz
source_filter: live
The required ``source_coll`` setting specifies the source collection from which to load content that will be recorded.
Most likely this will be the :ref:`live-web` collection, which should also be defined.
However, it could be any other collection, allowing for "extraction" from other collections or remote web archives.
@ -295,6 +294,20 @@ If running with auto indexing, the WARC will also get automatically indexed and
As a shortcut, ``recorder: live`` can also be used to specify only the ``source_coll`` option.
Optionally, a ``dedup_index`` key can be placed under the ``recorder`` key to enable deduplication of responses via an index::
recorder:
...
dedup_index:
type: redis
dupe_policy: revisit
redis_url: 'redis://localhost/2/{coll}:cdxj'
For ``type`` currently only ``redis`` is supported.
The ``dupe_policy`` key specifies what will hapen when a duplicate response is found. Can be ``duplicate``, to write duplicate responses, ``revisit``, to write a revisit record or ``skip`` to ignore duplicates and don't write anything to the WARC.
The ``redis_url`` key specifies which redis database to use and the template for the sorted-set key to use.
.. _auto-fetch:

31
pywb/apps/frontendapp.py
View File

@ -10,6 +10,8 @@ from wsgiprox.wsgiprox import WSGIProxMiddleware
from pywb.recorder.multifilewarcwriter import MultiFileWARCWriter
from pywb.recorder.recorderapp import RecorderApp
from pywb.recorder.filters import SkipDupePolicy, WriteDupePolicy, WriteRevisitDupePolicy
from pywb.recorder.redisindexer import WritableRedisIndexer
from pywb.utils.loaders import load_yaml_config
from pywb.utils.geventserver import GeventServer
@ -207,8 +209,33 @@ class FrontEndApp(object):
else:
recorder_coll = recorder_config['source_coll']
# TODO: support dedup
dedup_index = None
if 'dedup_index' in recorder_config:
dedup_config = recorder_config['dedup_index']
else:
dedup_config = None
if dedup_config:
type = dedup_config.get('type')
if type != 'redis':
msg = 'Invalid option for dedup_index: type: {0}'
raise Exception(msg.format(type))
dupe_policy = dedup_config.get('dupe_policy')
if dupe_policy == 'duplicate':
dupe_policy = WriteDupePolicy()
elif dupe_policy == 'revisit':
dupe_policy = WriteRevisitDupePolicy()
elif dupe_policy == 'skip':
dupe_policy = SkipDupePolicy()
else:
msg = 'Invalid option for dedup_index: dupe_policy: {0}'
raise Exception(msg.format(dupe_policy))
dedup_index = WritableRedisIndexer(redis_url=dedup_config.get('redis_url'),
dupe_policy=dupe_policy)
else:
dedup_index = None
warc_writer = MultiFileWARCWriter(self.warcserver.archive_paths,
max_size=int(recorder_config.get('rollover_size', 1000000000)),
max_idle_secs=int(recorder_config.get('rollover_idle_secs', 600)),