CIRCL
/
lookyloo
mirror of https://github.com/CIRCL/lookyloo
2
0
Fork 0
Code Issues Releases Wiki Activity

chg: update the cache at the right place.

pull/184/head
Raphaël Vinot 2021-03-20 21:54:46 +01:00
parent b1fbf9ff74
commit fa6b4701c0
2 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
bin/background_indexer.py
View File

@ -36,6 +36,7 @@ class BackgroundIndexer(AbstractManager):
try: try:
self.logger.info(f'Build pickle for {uuid}: {uuid_path.parent.name}') self.logger.info(f'Build pickle for {uuid}: {uuid_path.parent.name}')
self.lookyloo.get_crawled_tree(uuid) self.lookyloo.get_crawled_tree(uuid)
self.logger.info(f'Pickle for {uuid} build.')
except NoValidHarFile: except NoValidHarFile:
self.logger.warning(f'Unable to build pickle for {uuid}: {uuid_path.parent.name}') self.logger.warning(f'Unable to build pickle for {uuid}: {uuid_path.parent.name}')
# The capture is not working, moving it away. # The capture is not working, moving it away.
@ -44,10 +45,6 @@ class BackgroundIndexer(AbstractManager):
def _check_indexes(self): def _check_indexes(self):
for cache in self.lookyloo.sorted_capture_cache(): for cache in self.lookyloo.sorted_capture_cache():
if cache.incomplete_redirects:
# FIXME: this is dirty and needs to be moved.
self.lookyloo._set_capture_cache(cache.capture_dir, force=True)
cache = self.lookyloo.capture_cache(cache.uuid) # type: ignore
if self.lookyloo.is_public_instance and cache.no_index: if self.lookyloo.is_public_instance and cache.no_index:
# Capture unindexed # Capture unindexed
continue continue

3
lookyloo/lookyloo.py
View File

@ -485,6 +485,9 @@ class Lookyloo():
if not c: if not c:
continue continue
c = CaptureCache(c) c = CaptureCache(c)
if c.incomplete_redirects:
self._set_capture_cache(c.capture_dir, force=True)
c = self.capture_cache(c.uuid)
if hasattr(c, 'timestamp'): if hasattr(c, 'timestamp'):
all_cache.append(c) all_cache.append(c)
all_cache.sort(key=operator.attrgetter('timestamp'), reverse=True) all_cache.sort(key=operator.attrgetter('timestamp'), reverse=True)