def test_1_http_upload_fail(self):
requests.head = ResponseMockFactory.head_fail
requests.get = ResponseMockFactory.get_fail
url= "http://fail"
file_upload = models.FileUpload()
file_upload.set_id()
file_upload.upload("testuser", url, status="exists")
upload_dir = app.config.get("UPLOAD_DIR")
path = os.path.join(upload_dir, file_upload.local_filename)
self.cleanup_paths.append(path)
job = models.BackgroundJob()
result = ingestarticles.http_upload(job, path, file_upload)
assert result is False
assert file_upload.status == "failed"
assert file_upload.error is not None and file_upload.error != ""
assert file_upload.error_details is None
assert list(file_upload.failure_reasons.keys()) == []
# now try it with an actual exception
url= "http://except"
file_upload.upload("testuser", url, status="exists")
result = ingestarticles.http_upload(job, path, file_upload)
assert result is False
assert file_upload.status == "failed"
assert file_upload.error is not None and file_upload.error != ""
assert file_upload.error_details is None
assert list(file_upload.failure_reasons.keys()) == []
def test_3_submit_retry(self):
app.config["HUEY_TASKS"]["ingest_articles"]["retries"] = 1
fu = models.FileUpload()
fu.validated("doaj")
fu.save()
job = models.BackgroundJob()
params = {}
params["ingest_articles__file_upload_id"] = fu.id
params["ingest_articles__attempts"] = 0
job.params = params
job.save(blocking=True)
# this assumes that huey is in always eager mode, and thus this immediately calls the async task,
# which in turn calls execute, which ultimately calls run
with self.assertRaises(RetryException):
ingestarticles.IngestArticlesBackgroundTask.submit(job)
job = models.BackgroundJob.pull(job.id)
assert job.params.get("ingest_articles__attempts") == 1
assert job.status == "processing"
# now do it again, to see the retry cause the job to fail on the second attempt as per the config
with self.assertRaises(RetryException):
ingestarticles.IngestArticlesBackgroundTask.submit(job)
job = models.BackgroundJob.pull(job.id)
assert job.params.get("ingest_articles__attempts") == 2
assert job.status == "error"
lastmods = list(set(lastmods))
lastmods.sort()
for lm in lastmods:
for obj in lookup[lm]:
attempted += 1
publisher = obj["publisher"]
filename = obj["filename"]
id = obj["id"]
f = id + ".xml"
xml_file = os.path.join(xml_dir, f)
uploaded = datetime.fromtimestamp(lm).strftime("%Y-%m-%dT%H:%M:%SZ")
upload = models.FileUpload()
upload.set_schema(xwalk.format_name)
upload.upload(publisher, filename)
upload.set_created(uploaded)
upload.set_id()
# now try and parse the file
doc = None
try:
doc = etree.parse(open(xml_file))
except:
failed += 1
print f, "Malformed XML"
malformed_writer.writerow(
[f, publisher, filename, uploaded, acc.email])
upload.failed("Unable to parse file")
def _url_upload(cls, username, url, schema, previous):
# first define a few functions
def __http_upload(record, previous, url):
# first thing to try is a head request, supporting redirects
head = requests.head(url, allow_redirects=True)
if head.status_code == requests.codes.ok:
return __ok(record, previous)
# if we get to here, the head request failed. This might be because the file
# isn't there, but it might also be that the server doesn't support HEAD (a lot
# of webapps [including this one] don't implement it)
#
# so we do an interruptable get request instead, so we don't download too much
# unnecessary content
get = requests.get(url, stream=True)
get.close()
if get.status_code == requests.codes.ok:
return __ok(record, previous)
return __fail(
record,
previous,
error='error while checking submitted file reference: {0}'.
format(get.status_code))
def __ftp_upload(record, previous, parsed_url):
# 1. find out whether the file exists
# 2. that's it, return OK
# We might as well check if the file exists using the SIZE command.
# If the FTP server does not support SIZE, our article ingestion
# script is going to refuse to process the file anyway, so might as
# well get a failure now.
# Also it's more of a faff to check file existence using LIST commands.
try:
f = ftplib.FTP(parsed_url.hostname, parsed_url.username,
parsed_url.password)
r = f.sendcmd(
'TYPE I'
) # SIZE is not usually allowed in ASCII mode, so set to binary mode
if not r.startswith('2'):
return __fail(
record,
previous,
error='could not set binary '
'mode in target FTP server while checking file exists')
if f.size(parsed_url.path) < 0:
# this will either raise an error which will get caught below
# or, very rarely, will return an invalid size
return __fail(
record,
previous,
error='file does not seem to exist on FTP server')
except Exception as e:
return __fail(record,
previous,
error='error during FTP file existence check: ' +
str(e.args))
return __ok(record, previous)
def __ok(record, previous):
record.exists()
record.save()
previous.insert(0, record)
return record.id
def __fail(record, previous, error):
message = 'The URL could not be accessed; ' + error
record.failed(message)
record.save()
previous.insert(0, record)
raise BackgroundException(message)
# prep a record to go into the index, to record this upload. The filename is the url
record = models.FileUpload()
record.upload(username, url)
record.set_id()
record.set_schema(
schema) # although it could be wrong, this will get checked later
# now we attempt to verify that the file is retrievable
try:
# first, determine if ftp or http
parsed_url = urlparse(url)
if parsed_url.scheme in ['http', "https"]:
return __http_upload(record, previous, url)
elif parsed_url.scheme == 'ftp':
return __ftp_upload(record, previous, parsed_url)
else:
return __fail(
record,
previous,
error=
'unsupported URL scheme "{0}". Only HTTP(s) and FTP are supported.'
.format(parsed_url.scheme))
except BackgroundException as e:
raise
except Exception as e:
return __fail(record,
previous,
error="please check it before submitting again; " +
e.message)
def _file_upload(cls, username, f, schema, previous):
# prep a record to go into the index, to record this upload
record = models.FileUpload()
record.upload(username, f.filename)
record.set_id()
# the file path that we are going to write to
xml = os.path.join(app.config.get("UPLOAD_DIR", "."),
record.local_filename)
# it's critical here that no errors cause files to get left behind unrecorded
try:
# write the incoming file out to the XML file
f.save(xml)
# save the index entry
record.save()
except:
# if we can't record either of these things, we need to back right off
try:
file_failed(xml)
except:
pass
try:
record.delete()
except:
pass
raise BackgroundException(
"Failed to upload file - please contact an administrator")
xwalk_name = app.config.get("ARTICLE_CROSSWALKS", {}).get(schema)
xwalk = plugin.load_class(xwalk_name)()
# now we have the record in the index and on disk, we can attempt to
# validate it
try:
with open(xml) as handle:
xwalk.validate_file(handle)
record.validated(schema)
record.save()
previous.insert(0, record)
return record.id
except IngestException as e:
record.failed(e.message, e.inner_message)
try:
file_failed(xml)
except:
pass
record.save()
previous.insert(0, record)
raise BackgroundException("Failed to upload file: " + e.message +
"; " + str(e.inner_message))
except Exception as e:
record.failed("File system error when reading file")
try:
file_failed(xml)
except:
pass
record.save()
previous.insert(0, record)
raise BackgroundException(
"Failed to upload file - please contact an administrator")
def _file_upload(f, schema, previous):
# prep a record to go into the index, to record this upload
record = models.FileUpload()
record.upload(current_user.id, f.filename)
record.set_id()
# the file path that we are going to write to
xml = os.path.join(app.config.get("UPLOAD_DIR", "."), record.local_filename)
# it's critical here that no errors cause files to get left behind unrecorded
try:
# write the incoming file out to the XML file
f.save(xml)
# save the index entry
record.save()
except:
# if we can't record either of these things, we need to back right off
try:
os.remove(xml)
except:
pass
try:
record.delete()
except:
pass
flash("Failed to upload file - please contact an administrator", "error")
return render_template('publisher/uploadfile.html', previous=previous)
# now we have the record in the index and on disk, we can attempt to
# validate it
try:
actual_schema = None
with open(xml) as handle:
actual_schema = article.check_schema(handle, schema)
except:
# file is a dud, so remove it
try:
os.remove(xml)
except:
pass
# if we're unable to validate the file, we should record this as
# a file error.
record.failed("Unable to parse file")
record.save()
previous = [record] + previous
flash("Failed to parse file - it is invalid XML; please fix it before attempting to upload again.", "error")
return render_template('publisher/uploadfile.html', previous=previous)
if actual_schema:
record.validated(actual_schema)
record.save()
previous = [record] + previous # add the new record to the previous records
flash("File successfully uploaded - it will be processed shortly", "success")
return render_template('publisher/uploadfile.html', previous=previous)
else:
record.failed("File could not be validated against a known schema")
record.save()
os.remove(xml)
previous = [record] + previous
flash("File could not be validated against a known schema; please fix this before attempting to upload again", "error")
return render_template('publisher/uploadfile.html', previous=previous)
