arsenetar/dupeguru
1
0
Fork 0
mirror of https://github.com/arsenetar/dupeguru.git synced 2025-03-10 05:34:36 +00:00
Code Releases Wiki Activity

Catch MemoryError better in PE's block matching algo

Browse Source 
fixes #264 (for good this time, hopefully)
This commit is contained in:
Virgil Dupras 2014-10-05 22:22:59 -04:00
parent 5b3d5f5d1c
commit 321f8ab406
3 changed files with 43 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
core_pe/matchblock.py
View File

@ -193,13 +193,13 @@ def getmatches(pictures, cache_path, threshold=75, match_scaled=False, j=job.nul
# some wiggle room, log about the incident, and stop matching right here. We then process
# the matches we have. The rest of the process doesn't allocate much and we should be
# alright.
del matches[-1000:] # some wiggle room to ensure we don't run out of memory again.
logging.warning("Ran out of memory when scanning! We had %d matches.", len(matches) + 1000)
del comparisons_to_do, chunks, pictures # some wiggle room for the next statements
logging.warning("Ran out of memory when scanning! We had %d matches.", len(matches))
del matches[-len(matches)//3:] # some wiggle room to ensure we don't run out of memory again.
pool.close()
result = []
myiter = j.iter_with_progress(
iterconsume(matches),
iterconsume(matches, reverse=False),
tr("Verified %d/%d matches"),
every=10,
count=len(matches),

56
hscommon/tests/util_test.py
View File

@ -1,9 +1,9 @@
# Created By: Virgil Dupras
# Created On: 2011-01-11
# Copyright 2014 Hardcoded Software (http://www.hardcoded.net)
#
# This software is licensed under the "BSD" License as described in the "LICENSE" file,
# which should be included with this package. The terms are also available at
#
# This software is licensed under the "BSD" License as described in the "LICENSE" file,
# which should be included with this package. The terms are also available at
# http://www.hardcoded.net/licenses/bsd_license
from io import StringIO
@ -65,6 +65,12 @@ def test_trailiter():
eq_(list(trailiter(['foo', 'bar'], skipfirst=True)), [('foo', 'bar')])
eq_(list(trailiter([], skipfirst=True)), []) # no crash
def test_iterconsume():
# We just want to make sure that we return *all* items and that we're not mistakenly skipping
# one.
eq_(list(range(2500)), list(iterconsume(list(range(2500)))))
eq_(list(reversed(range(2500))), list(iterconsume(list(range(2500)), reverse=False)))
#--- String
def test_escape():
@ -188,63 +194,63 @@ class TestCase_modified_after:
monkeyplus.patch_osstat('first', st_mtime=42)
monkeyplus.patch_osstat('second', st_mtime=41)
assert modified_after('first', 'second')
def test_second_is_modified_after(self, monkeyplus):
monkeyplus.patch_osstat('first', st_mtime=42)
monkeyplus.patch_osstat('second', st_mtime=43)
assert not modified_after('first', 'second')
def test_same_mtime(self, monkeyplus):
monkeyplus.patch_osstat('first', st_mtime=42)
monkeyplus.patch_osstat('second', st_mtime=42)
assert not modified_after('first', 'second')
def test_first_file_does_not_exist(self, monkeyplus):
# when the first file doesn't exist, we return False
monkeyplus.patch_osstat('second', st_mtime=42)
assert not modified_after('does_not_exist', 'second') # no crash
def test_second_file_does_not_exist(self, monkeyplus):
# when the second file doesn't exist, we return True
monkeyplus.patch_osstat('first', st_mtime=42)
assert modified_after('first', 'does_not_exist') # no crash
def test_first_file_is_none(self, monkeyplus):
# when the first file is None, we return False
monkeyplus.patch_osstat('second', st_mtime=42)
assert not modified_after(None, 'second') # no crash
def test_second_file_is_none(self, monkeyplus):
# when the second file is None, we return True
monkeyplus.patch_osstat('first', st_mtime=42)
assert modified_after('first', None) # no crash
class TestCase_delete_if_empty:
def test_is_empty(self, tmpdir):
testpath = Path(str(tmpdir))
assert delete_if_empty(testpath)
assert not testpath.exists()
def test_not_empty(self, tmpdir):
testpath = Path(str(tmpdir))
testpath['foo'].mkdir()
assert not delete_if_empty(testpath)
assert testpath.exists()
def test_with_files_to_delete(self, tmpdir):
testpath = Path(str(tmpdir))
testpath['foo'].open('w')
testpath['bar'].open('w')
assert delete_if_empty(testpath, ['foo', 'bar'])
assert not testpath.exists()
def test_directory_in_files_to_delete(self, tmpdir):
testpath = Path(str(tmpdir))
testpath['foo'].mkdir()
assert not delete_if_empty(testpath, ['foo'])
assert testpath.exists()
def test_delete_files_to_delete_only_if_dir_is_empty(self, tmpdir):
testpath = Path(str(tmpdir))
testpath['foo'].open('w')
@ -252,25 +258,25 @@ class TestCase_delete_if_empty:
assert not delete_if_empty(testpath, ['foo'])
assert testpath.exists()
assert testpath['foo'].exists()
def test_doesnt_exist(self):
# When the 'path' doesn't exist, just do nothing.
delete_if_empty(Path('does_not_exist')) # no crash
def test_is_file(self, tmpdir):
# When 'path' is a file, do nothing.
p = Path(str(tmpdir)) + 'filename'
p.open('w').close()
delete_if_empty(p) # no crash
def test_ioerror(self, tmpdir, monkeypatch):
# if an IO error happens during the operation, ignore it.
def do_raise(*args, **kw):
raise OSError()
monkeypatch.setattr(Path, 'rmdir', do_raise)
delete_if_empty(Path(str(tmpdir))) # no crash
class TestCase_open_if_filename:
def test_file_name(self, tmpdir):
@ -280,7 +286,7 @@ class TestCase_open_if_filename:
assert close
eq_(b'test_data', file.read())
file.close()
def test_opened_file(self):
sio = StringIO()
sio.write('test_data')
@ -288,14 +294,14 @@ class TestCase_open_if_filename:
file, close = open_if_filename(sio)
assert not close
eq_('test_data', file.read())
def test_mode_is_passed_to_open(self, tmpdir):
filepath = str(tmpdir.join('test.txt'))
open(filepath, 'w').close()
file, close = open_if_filename(filepath, 'a')
eq_('a', file.mode)
file.close()
class TestCase_FileOrPath:
def test_path(self, tmpdir):
@ -303,17 +309,17 @@ class TestCase_FileOrPath:
open(filepath, 'wb').write(b'test_data')
with FileOrPath(filepath) as fp:
eq_(b'test_data', fp.read())
def test_opened_file(self):
sio = StringIO()
sio.write('test_data')
sio.seek(0)
with FileOrPath(sio) as fp:
eq_('test_data', fp.read())
def test_mode_is_passed_to_open(self, tmpdir):
filepath = str(tmpdir.join('test.txt'))
open(filepath, 'w').close()
with FileOrPath(filepath, 'a') as fp:
eq_('a', fp.mode)

19
hscommon/util.py
View File

@ -117,23 +117,20 @@ def trailiter(iterable, skipfirst=False):
yield prev, item
prev = item
def iterconsume(seq):
"""Iterate over ``seq`` and discard yielded objects.
def iterconsume(seq, reverse=True):
"""Iterate over ``seq`` and pops yielded objects.
Right after the ``yield``, we replace the element we've just yielded by ``None`` in the
sequence.
Because we use the ``pop()`` method, we reverse ``seq`` before proceeding. If you don't need
to do that, set ``reverse`` to ``False``.
This is useful in tight memory situation where you are looping over a sequence of objects that
are going to be discarded afterwards. If you're creating other objects during that iteration
you might want to use this to avoid ``MemoryError``.
Note that this only works for sequence (index accessible), not all iterables.
"""
# We don't use ``del``, because it would be disastrous performance-wise as the array would have
# to be constantly re-allocated.
for index, elem in enumerate(seq):
seq[index] = None
yield elem
if reverse:
seq.reverse()
while seq:
yield seq.pop()
#--- String related