def _get_zenodo_doi(label, local_filename):
try:
# Python 2
from HTMLParser import HTMLParser
bytes2text = lambda x: x
except ImportError:
# Python 3
from html.parser import HTMLParser
def bytes2text(b):
return b.decode(encoding="utf8")
class ZenodoParser(HTMLParser):
def handle_starttag(self, tag, attrs):
if tag == "link":
attrs = dict(attrs)
if attrs.get("rel") == "alternate" \
and attrs.get("type") != "application/rss+xml":
self.link_href = attrs.get("href")
self.link_type = attrs.get("type")
zenodo_url = "http://dx.doi.org/" + label
parser = ZenodoParser()
source = url.urlopen(zenodo_url)
try:
parser.feed(bytes2text(source.read()))
finally:
source.close()
assert parser.link_type == "application/octet-stream"
download_url = parser.link_href
url.urlretrieve(download_url, local_filename)
return local_filename
def _get_figshare_doi(label, local_filename):
figshare_url = "http://api.figshare.com/v1/articles/%s" % label
try:
response = url.urlopen(figshare_url)
json_data = response.read().decode("utf-8")
except url.HTTPError:
raise ValueError("Not a figshare DOI: %s" % label)
article_details = json.loads(json_data)
download_url = article_details['items'][0]['files'][0]['download_url']
url.urlretrieve(download_url, local_filename)
return local_filename
def find_in_library(paper_ref):
ref_type, label = split_paper_ref(paper_ref)
assert ref_type in ["doi", "local"]
if ref_type == "local":
filename = label + '.ap'
for dir in library:
full_name = os.path.join(dir, "local", filename)
if os.path.exists(full_name):
return full_name
raise IOError(2, "No such ActivePaper: '%s' (filename: %s)"
% (paper_ref, full_name))
elif ref_type == "doi":
local_filename = os.path.join(library[0], label + ".ap")
if os.path.exists(local_filename):
return local_filename
# Only figshare is supported for downloading at the moment.
# There doesn't seem to be a way to download an
# arbitrary digital object through its DOI.
figshare_url = "http://api.figshare.com/v1/articles/%s" % label
try:
response = url.urlopen(figshare_url)
json_data = response.read().decode("utf-8")
except urllib2.HTTPError:
raise ValueError("Not a figshare DOI: %s" % label)
article_details = json.loads(json_data)
download_url = article_details['items'][0]['files'][0]['download_url']
dir_name = os.path.join(library[0], label.split("/")[0])
if not os.path.exists(dir_name):
os.mkdir(dir_name)
url.urlretrieve(download_url, local_filename)
return local_filename