sporada/nopaque
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/nopaque.git synced 2025-11-03 20:02:47 +00:00
Code Issues Projects Releases Wiki Activity

First metadata addition

Browse Source
This commit is contained in:
Stephan Porada
2020-04-29 16:15:11 +02:00
parent eb5e5c3253
commit d54a3d3f33
8 changed files with 65 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
app/corpora/events.py
View File

@@ -7,6 +7,7 @@ from ..events import connected_sessions
from ..models import Corpus, User
import cqi
import math
from app import logger
'''
@@ -36,14 +37,27 @@ def corpus_analysis_get_meta_data(corpus_id):
db_corpus = Corpus.query.get(corpus_id)
# TODO: Check if current user is actually the creator of the corpus?
metadata = {}
metadata["corpus_name"] = db_corpus.title
metadata["corpus_description"] = db_corpus.description
metadata["corpus_creation_date"] = db_corpus.creation_date.isoformat()
metadata['corpus_name'] = db_corpus.title
metadata['corpus_description'] = db_corpus.description
metadata['corpus_creation_date'] = db_corpus.creation_date.isoformat()
metadata['corpus_last_edited_date'] = db_corpus.last_edited_date.isoformat()
# get meta data from corpus in cqp server
client = corpus_analysis_clients.get(request.sid)
client_corpus = client.corpora.get('CORPUS')
metadata["corpus_properties"] = client_corpus.attrs['properties']
# metadata["corpus_properties"] = client.api.corpus_properties('CORPUS')
metadata['corpus_properties'] = client_corpus.attrs['properties']
metadata['corpus_size_tokens'] = client_corpus.attrs['size']
texts = {}
text_attr = client_corpus.structural_attributes.get('text')
for text_id in range(0, (text_attr.attrs['size'])):
texts[text_id] = {}
logger.warning('Text ID: {}'.format(text_id))
struct_attrs = client_corpus.structural_attributes.list(filters={'part_of': text_attr})
for struct_attr in struct_attrs:
logger.warning(struct_attr)
for struct_id in list(range(0, struct_attr.attrs['size'])):
logger.warning(struct_attr.values_by_ids([struct_id]))
payload = metadata
response = {'code': 200, 'desc': 'Corpus meta data', 'msg': 'OK', 'payload': payload}
socketio.emit('corpus_analysis_send_meta_data', response, room=request.sid)

2
app/corpora/tasks.py
View File

@@ -1,3 +1,4 @@
from datetime import datetime
from .. import db
from ..decorators import background
from ..models import Corpus, CorpusFile
@@ -45,6 +46,7 @@ def build_corpus(corpus_id, *args, **kwargs):
master_element_tree.write(output_file, xml_declaration=True,
encoding='utf-8')
corpus.status = 'submitted'
corpus.last_edited_date = datetime.utcnow()
db.session.commit()