def _purge_batch(self, batch):
batch_name = batch.name
# just delete batch causes memory to bloat out
# so we do it piece-meal
for issue in batch.issues.all():
for page in issue.pages.all():
page.delete()
# remove coordinates
if os.path.exists(models.coordinates_path(page._url_parts())):
os.remove(models.coordinates_path(page._url_parts()))
issue.delete()
batch.delete()
if self.PROCESS_OCR:
self.solr.delete_query('batch:"%s"' % batch_name)
self.solr.commit()
def _process_coordinates(self, page, coords):
_logger.debug("writing out word coords for %s" %
page.url)
f = open(models.coordinates_path(page._url_parts()), "w")
f.write(gzip_compress(json.dumps(coords)))
f.close()
def _process_coordinates(self, page, coords):
_logger.debug("writing out word coords for %s" %
page.url)
f = open(models.coordinates_path(page._url_parts()), "w")
f.write(gzip_compress(json.dumps(coords)))
f.close()
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
file_path = models.coordinates_path(url_parts)
try:
with gzip.open(file_path, 'rb') as i:
response = HttpResponse(i.read(), content_type='application/json')
return add_cache_tag(response, "lccn=%s" % lccn)
except IOError:
LOGGER.warning('Word coordinates file %s does not exist', file_path)
raise Http404
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
file_path = models.coordinates_path(url_parts)
try:
with gzip.open(file_path, 'rb') as i:
response = HttpResponse(i.read(), content_type='application/json')
return add_cache_tag(response, "lccn=%s" % lccn)
except IOError:
LOGGER.warning('Word coordinates file %s does not exist', file_path)
raise Http404
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
try:
f = open(models.coordinates_path(url_parts))
except IOError:
return HttpResponse()
data = f.read()
r = HttpResponse(mimetype='application/json')
r['Content-Encoding'] = 'gzip'
r['Content-Length'] = len(data)
r.write(data)
f.close()
return r
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
try:
f = open(models.coordinates_path(url_parts))
except IOError:
return HttpResponse()
data = f.read()
r = HttpResponse(mimetype='application/json')
r['Content-Encoding'] = 'gzip'
r['Content-Length'] = len(data)
r.write(data)
f.close()
return r
def _process_coordinates(self, page, coords):
LOGGER.debug("writing out word coords for %s", page.url)
fd, path = tempfile.mkstemp(text="w", suffix=".coordinates", dir=settings.TEMP_STORAGE) # get a temp file in case the coordinates dir is a NFS or S3 mount which have poor multiple write performance
f = open(path, "w")
f.write(gzip_compress(json.dumps(coords)))
f.close()
os.close(fd)
final_path = models.coordinates_path(page._url_parts())
try:
shutil.move(path, final_path)
except Exception:
LOGGER.warn("Could not move coordinates to [%s]. Waiting 5 seconds and trying again in case of network mount", final_path)
time.sleep(5)
shutil.move(path, final_path)
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = {
"lccn": lccn,
"date": date,
"edition": edition,
"sequence": sequence
}
file_path = models.coordinates_path(url_parts)
try:
with gzip.open(file_path, "rb") as i:
return HttpResponse(i.read(), content_type="application/json")
except IOError:
LOGGER.warning("Word coordinates file %s does not exist", file_path)
raise Http404
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
try:
file_data = gzip.open(models.coordinates_path(url_parts), 'rb')
except IOError:
return HttpResponse()
data = json.load(file_data)
non_lexemes = re.compile('''^[^a-zA-Z0-9]+|[^a-zA-Z0-9]+$|'s$''')
return_coords = data.copy()
# reset coords to {} and build afresh, getting rid of unwanted punctuations
return_coords['coords'] = {}
for key in data.get('coords'):
return_coords['coords'][re.sub(non_lexemes, '', key)] = data['coords'][key]
r = HttpResponse(mimetype='application/json')
r.write(json.dumps(return_coords))
return r
def coordinates(request, lccn, date, edition, sequence, words=None):
url_parts = dict(lccn=lccn, date=date, edition=edition, sequence=sequence)
try:
file_data = gzip.open(models.coordinates_path(url_parts), 'rb')
except IOError:
return HttpResponse()
data = json.load(file_data)
non_lexemes = re.compile('''^[^a-zA-Z0-9]+|[^a-zA-Z0-9]+$|'s$''')
return_coords = data.copy()
# reset coords to {} and build afresh, getting rid of unwanted punctuations
return_coords['coords'] = {}
for key in data.get('coords'):
return_coords['coords'][re.sub(non_lexemes, '', key)] = data['coords'][key]
r = HttpResponse(content_type='application/json')
r.write(json.dumps(return_coords))
return r
def _process_coordinates(self, page, coords):
LOGGER.debug("writing out word coords for %s", page.url)
# We'll use a temporary file in case the coordinates dir is configured
# to a network filesystem which has poor update performance
# characteristics
fd, path = tempfile.mkstemp(text="w",
suffix=".coordinates",
dir=settings.TEMP_STORAGE)
f = open(path, "w")
f.write(gzip_compress(json.dumps(coords)))
f.close()
os.close(fd)
final_path = models.coordinates_path(page._url_parts())
try:
shutil.move(path, final_path)
except Exception:
LOGGER.warning(
'Could not move coordinates to "%s". Waiting 5 seconds before trying again…',
final_path)
time.sleep(5)
shutil.move(path, final_path)
