Add start of metadata extract
Former-commit-id: 51210e2d4385a4942d9d7a380a4c16c16811f98b
| |
#http://packages.python.org/CouchDB/client.html |
| |
import couchdb |
| |
from BeautifulSoup import BeautifulSoup |
| |
|
| |
couch = couchdb.Server('http://127.0.0.1:5984/') |
| |
|
| |
# select database |
| |
docsdb = couch['disclosr-documents'] |
| |
|
| |
for row in docsdb.view('app/getMetadataExtractRequired'): |
| |
print row.id |
| |
html = docsdb.get_attachment(row.id,row.value.iterkeys().next()).read() |
| |
metadata = [] |
| |
# http://www.crummy.com/software/BeautifulSoup/documentation.html |
| |
soup = BeautifulSoup(html) |
| |
metatags = soup.meta |
| |
for metatag in metatags: |
| |
print metatag['name'] |
| |
doc = docsdb.get(row.id) |
| |
//doc['metadata'] = metadata |
| |
//docsdb.save(doc) |
| |
|