def purge_batch(self, batch_name):
event = LoadBatchEvent(batch_name=batch_name, message="starting purge")
event.save()
try:
batch = self._get_batch(batch_name)
self._purge_batch(batch)
event = LoadBatchEvent(batch_name=batch_name, message="purged")
event.save()
# clean up symlinks if exists
link_name = os.path.join(settings.BATCH_STORAGE, batch_name)
if os.path.islink(link_name):
LOGGER.info("Removing symlink %s", link_name)
os.remove(link_name)
except Exception as e:
msg = "purge failed: %s" % e
LOGGER.exception(msg)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
raise BatchLoaderException(msg)
batch_source += "/"
batch_name = _normalize_batch_name(batch_name)
if not strict:
try:
batch = Batch.objects.get(name=batch_name)
_logger.info("Batch already loaded: %s" % batch_name)
return batch
except Batch.DoesNotExist, e:
pass
_logger.info("loading batch: %s" % batch_name)
t0 = time()
times = []
event = LoadBatchEvent(batch_name=batch_name, message="starting load")
event.save()
batch = None
try:
# build a Batch object for the batch location
batch = self._get_batch(batch_name, batch_source, create=True)
self._sanity_check_batch(batch)
# stash it away for processing later on
self.current_batch = batch
# parse the batch.xml and load up each issue mets file
doc = etree.parse(batch.validated_batch_url)
for e in doc.xpath('ndnp:reel', namespaces=ns):
class BatchLoader(object):
"""This class allows you to load a batch into the database. A loader
object serves as a context for a particular batch loading job.
"""
def __init__(self,
process_ocr=True,
process_coordinates=True,
in_copyright=False):
"""Create a BatchLoader.
The process_ocr parameter is used (mainly in testing) when we don't
want to spend time actually extracting ocr text and indexing.
"""
self.PROCESS_OCR = process_ocr
if self.PROCESS_OCR:
self.solr = SolrConnection(settings.SOLR)
self.PROCESS_COORDINATES = process_coordinates
self.IN_COPYRIGHT = in_copyright
def _find_batch_file(self, batch):
# look for batch_1.xml, BATCH_1.xml, etc
for alias in [
"batch_1.xml", "BATCH_1.xml", "batchfile_1.xml", "batch_2.xml",
"BATCH_2.xml", "batch.xml"
]:
if os.path.isfile(os.path.join(batch.storage_url, alias)):
return alias
else:
continue
else:
raise BatchLoaderException(
"could not find batch_1.xml (or any of its aliases) in '%s' -- has the batch been validated?"
% batch.path)
def _sanity_check_batch(self, batch):
if not os.path.exists(batch.path):
raise BatchLoaderException("batch does not exist at %s" %
batch.path)
# b = urllib2.urlopen(batch.url)
batch.validated_batch_file = self._find_batch_file(batch)
def load_batch(self, batch_path, strict=True):
"""Load a batch, and return a Batch instance for the batch
that was loaded.
loader.load_batch('/path/to/batch_curiv_ahwahnee_ver01')
"""
self.pages_processed = 0
logging.info("loading batch at %s", batch_path)
dirname, batch_name = os.path.split(batch_path.rstrip("/"))
if dirname:
batch_source = None
link_name = os.path.join(settings.BATCH_STORAGE, batch_name)
if batch_path != link_name and not os.path.islink(link_name):
_logger.info("creating symlink %s -> %s", batch_path,
link_name)
os.symlink(batch_path, link_name)
else:
batch_source = os.path.join(settings.BATCH_STORAGE, batch_name)
if not batch_source.endswith("/"):
batch_source += "/"
batch_name = _normalize_batch_name(batch_name)
if not strict:
try:
batch = Batch.objects.get(name=batch_name)
_logger.info("Batch already loaded: %s" % batch_name)
return batch
except Batch.DoesNotExist, e:
pass
_logger.info("loading batch: %s" % batch_name)
t0 = time()
times = []
event = LoadBatchEvent(batch_name=batch_name, message="starting load")
event.save()
batch = None
try:
# build a Batch object for the batch location
batch = self._get_batch(batch_name, batch_source, create=True)
self._sanity_check_batch(batch)
# stash it away for processing later on
self.current_batch = batch
# parse the batch.xml and load up each issue mets file
doc = etree.parse(batch.validated_batch_url)
for e in doc.xpath('ndnp:reel', namespaces=ns):
reel_number = e.attrib['reelNumber'].strip()
try:
reel = models.Reel.objects.get(number=reel_number,
batch=batch)
except models.Reel.DoesNotExist, e:
reel = models.Reel(number=reel_number, batch=batch)
reel.save()
for e in doc.xpath('ndnp:issue', namespaces=ns):
mets_url = urlparse.urljoin(batch.storage_url, e.text)
try:
issue = self._load_issue(mets_url)
except ValueError, e:
_logger.exception(e)
continue
reset_queries()
times.append((time() - t0, self.pages_processed))
def purge_batch(self, batch_name):
event = LoadBatchEvent(batch_name=batch_name, message="starting purge")
event.save()
try:
batch = self._get_batch(batch_name)
self._purge_batch(batch)
event = LoadBatchEvent(batch_name=batch_name, message="purged")
event.save()
# clean up symlinks if exists
link_name = os.path.join(settings.BATCH_STORAGE, batch_name)
if os.path.islink(link_name):
_logger.info("Removing symlink %s", link_name)
os.remove(link_name)
except Exception, e:
msg = "purge failed: %s" % e
_logger.error(msg)
_logger.exception(e)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
raise BatchLoaderException(msg)
mets_url = urlparse.urljoin(batch.storage_url, e.text)
try:
issue = self._load_issue(mets_url)
except ValueError, e:
_logger.exception(e)
continue
reset_queries()
times.append((time() - t0, self.pages_processed))
# commit new changes to the solr index, if we are indexing
if self.PROCESS_OCR:
self.solr.commit()
batch.save()
msg = "processed %s pages" % batch.page_count
event = LoadBatchEvent(batch_name=batch_name, message=msg)
_logger.info(msg)
event.save()
# _chart(times)
except Exception, e:
msg = "unable to load batch: %s" % e
_logger.error(msg)
_logger.exception(e)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
try:
self.purge_batch(batch_name)
except Exception, pbe:
_logger.error("purge batch failed for failed load batch: %s" %
pbe)
batch_source += "/"
batch_name = _normalize_batch_name(batch_name)
if not strict:
try:
batch = Batch.objects.get(name=batch_name)
_logger.info("Batch already loaded: %s" % batch_name)
return batch
except Batch.DoesNotExist, e:
pass
_logger.info("loading batch: %s" % batch_name)
t0 = time()
times = []
event = LoadBatchEvent(batch_name=batch_name, message="starting load")
event.save()
batch = None
try:
# build a Batch object for the batch location
batch = self._get_batch(batch_name, batch_source, create=True)
self._sanity_check_batch(batch)
# stash it away for processing later on
self.current_batch = batch
# parse the batch.xml and load up each issue mets file
doc = etree.parse(batch.validated_batch_url)
for e in doc.xpath('ndnp:reel', namespaces=ns):
def load_batch(self, batch_path, strict=True):
"""Load a batch, and return a Batch instance for the batch
that was loaded.
loader.load_batch('/path/to/batch_curiv_ahwahnee_ver01')
"""
LOGGER.info("loading batch at %s", batch_path)
dirname, batch_name = os.path.split(batch_path.rstrip("/"))
if dirname:
batch_source = None
link_name = os.path.join(settings.BATCH_STORAGE, batch_name)
if batch_path != link_name and not os.path.islink(link_name):
LOGGER.info("creating symlink %s -> %s", batch_path, link_name)
os.symlink(batch_path, link_name)
else:
batch_source = urlparse.urljoin(settings.BATCH_STORAGE, batch_name)
if not batch_source.endswith("/"):
batch_source += "/"
batch_name = _normalize_batch_name(batch_name)
if not strict:
try:
batch = Batch.objects.get(name=batch_name)
LOGGER.info("Batch already loaded: %s", batch_name)
return batch
except Batch.DoesNotExist as e:
pass
LOGGER.info("loading batch: %s", batch_name)
event = LoadBatchEvent(batch_name=batch_name, message="starting load")
event.save()
batch = None
try:
# build a Batch object for the batch location
batch = self._get_batch(batch_name, batch_source, create=True)
self._sanity_check_batch(batch)
# stash it away for processing later on
self.current_batch = batch
# parse the batch.xml and load up each issue mets file
doc = etree.parse(batch.validated_batch_url)
for e in doc.xpath('ndnp:reel', namespaces=ns):
reel_number = e.attrib['reelNumber'].strip()
try:
reel = models.Reel.objects.get(number=reel_number,
batch=batch)
except models.Reel.DoesNotExist as e:
reel = models.Reel(number=reel_number, batch=batch)
reel.save()
for e in doc.xpath('ndnp:issue', namespaces=ns):
mets_url = urlparse.urljoin(batch.storage_url, e.text)
try:
issue, pages = self._load_issue(mets_url)
except ValueError as e:
LOGGER.exception(e)
continue
# commit new changes to the solr index, if we are indexing
if self.PROCESS_OCR:
LOGGER.info("Adding pages to solr index from issue %s", issue.title)
for page in pages:
LOGGER.debug("indexing ocr for: %s", page.url)
self.solr.add(**page.solr_doc)
page.indexed = True
page.save()
if self.PROCESS_OCR:
LOGGER.info("Committing solr index")
self.solr.commit()
batch.save()
msg = "processed %s pages" % batch.page_count
LOGGER.info(msg)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
except Exception as e:
msg = "unable to load batch: %s" % e
LOGGER.exception(msg)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
try:
self.purge_batch(batch_name)
except Exception as pbe:
LOGGER.error("purge batch failed for failed load batch: %s", pbe)
LOGGER.exception(pbe)
raise BatchLoaderException(msg)
if settings.IS_PRODUCTION:
batch.released = datetime.now()
batch.save()
return batch
def purge_batch(self, batch_name):
batch_name = _normalize_batch_name(batch_name)
try:
batch = self._get_batch(batch_name)
except Batch.DoesNotExist:
LOGGER.info("Batch %s does not exist", batch_name)
return
event = LoadBatchEvent(batch_name=batch_name, message="starting purge")
event.save()
try:
self._purge_batch(batch)
event = LoadBatchEvent(batch_name=batch_name, message="purged")
event.save()
# clean up symlinks if exists
link_name = os.path.join(settings.BATCH_STORAGE, batch_name)
if os.path.islink(link_name):
LOGGER.info("Removing symlink %s", link_name)
os.remove(link_name)
except Exception as e:
msg = "purge failed: %s" % e
LOGGER.exception(msg)
event = LoadBatchEvent(batch_name=batch_name, message=msg)
event.save()
raise BatchLoaderException(msg)
