def upload_annotated_file(request):
handle_uploaded_file(request.FILES['file_to_analyze'])
filename = str(request.FILES['file_to_analyze'])
normalized = check_if_normalized(
open(upload_location + filename).readlines())
try:
# File that are already annotated are assumed to be eligible
t = import_textfile(upload_location + filename, True, normalized)
if type(t) == str:
return JsonResponse({'success': 0, 'error_meta': t})
md5 = get_md5(t)
try:
existing = UploadedFile.objects.get(pk=md5)
t.normalized = existing.normalized
except:
pass
finally:
os.remove(upload_location + filename)
if not t:
return JsonResponse({'success': 0})
request = set_session(request, t)
return JsonResponse({'success': 1})
def save_plugin_data(self, request=None):
"""
Saving the plugin data and moving the file.
"""
image = self.cleaned_data.get('image', None)
if image:
saved_image = handle_uploaded_file(image)
self.cleaned_data['image'] = saved_image
def handle(self, username):
if self.is_valid():
info = self.cleaned_data
# existed asset
if 'asset' in info:
if info['asset'].uid.user.username == username:
# Store temporary file
handle_uploaded_file(info['file'], info['asset'], is_final=False)
return {'non_existed': False,
'aid': info['asset'].aid,
'owner': info['asset'].uid.user.username,
'new_mime_type': info['new_mime_type'],
'new_nice_type': info['new_nice_type'],
'is_existed_owner': info['asset'].uid.user.username == username,
'new_keywords': info['tags']}
# success adding new asset
asset = Asset()
asset.populate(username, info)
handle_uploaded_file(info['file'], asset, is_final=True)
return {'non_existed': True, 'asset': asset, 'aid': asset.aid}
def annotate_uploaded_file(request):
tmp_dir = tempfile.mkdtemp() + "/" # Used for the pipeline
normalized = False
use_paste = False
pasted_text = None
for key in request.POST:
if key == 'use_paste':
if checkbox_to_bool(request.POST[key]):
use_paste = True
elif key == 'pasted_text':
pasted_text = request.POST[key]
elif key == 'checkNormalization':
normalized = checkbox_to_bool(request.POST[key])
if use_paste:
tf = tempfile.NamedTemporaryFile(delete=False)
with codecs.open(tf.name, 'wb', encoding='utf-8') as f:
f.write(pasted_text)
file_path = tf.name
filename = 'paste ' + datetime.now().strftime("%H:%M:%S") + '.txt'
shutil.move(tf.name, upload_location + filename)
options = get_optparse(request,
upload_location + filename,
tmp_dir,
custom_filename=filename)
else:
handle_uploaded_file(request.FILES['file_to_annotate'])
filename = str(request.FILES['file_to_annotate'])
original_filename = filename
if settings.PRODUCTION and os.path.splitext(filename)[1] != ".txt":
try:
subprocess.call(
['unoconv', '--format=txt', upload_location + filename])
stdout_file = open(
upload_location + os.path.splitext(filename)[0] + ".txt2",
"w")
subprocess.call([
'iconv', '-f', 'iso-8859-1', '-t', 'utf-8',
upload_location + os.path.splitext(filename)[0] + ".txt"
],
stdout=stdout_file)
shutil.move(
upload_location + os.path.splitext(filename)[0] + ".txt2",
upload_location + os.path.splitext(filename)[0] + ".txt")
filename = os.path.splitext(filename)[0] + ".txt"
finally:
os.remove(upload_location + original_filename)
options = get_optparse(request, upload_location + filename, tmp_dir)
# If the user has removed some column, the text can't be used for analysis
text_eligible = False
if len(options.columns.split(",")) == 13:
text_eligible = True
try:
if request.session['language'] == 'en':
annotated_file_path = pipeline_en.run(options)
else:
annotated_file_path = pipeline.run(options)
finally:
shutil.rmtree(tmp_dir)
@timing
def import_t(request, annotated_file_path, text_eligible, normalized):
return import_textfile(request, annotated_file_path, text_eligible,
normalized)
# The second arg here is whether to use metadata or not
try:
t = import_t(request, annotated_file_path, text_eligible, normalized)
finally:
try:
os.remove(annotated_file_path)
except:
pass
if not t:
return JsonResponse({'success': 0})
if type(t) == str:
return JsonResponse({'success': 0, 'error_meta': t})
request = set_session(request, t)
return JsonResponse({'success': 1})
