def create_article(metadata):
tmpdir = tempfile.mkdtemp()
pmc_file = '%s/article.tar.gz' %(tmpdir)
print('Downloading: %s' %(metadata['uid']))
urllib.request.urlretrieve(metadata['download_url'], pmc_file)
tar = tarfile.open(pmc_file, "r:gz")
tar.extractall(tmpdir)
files = glob('%s/%s/*' %(tmpdir,metadata['pmcid']))
images = [x for x in files if imghdr.what(x) is not None]
pdf_files = [x for x in files if x.lower().endswith('pdf')]
xml_file = [x for x in files if x.lower().endswith('xml')]
images = images + pdf_files
general_client.upload_dataset(images=images,
texts=xml_file,
collection=collection,
uid=metadata['uid'],
metadata=metadata)
shutil.rmtree(tmpdir)
######################################################################
# Signals
######################################################################
评论列表
文章目录