def add_table_ob_to_article(table_html, article_ob, text_mine = True, uploading_user = None):
if uploading_user:
user_uploaded = True
else:
user_uploaded = False
table_soup = BeautifulSoup(table_html, 'lxml')
table_html_cleaned = str(table_soup)
table_html_cleaned = add_id_tags_to_table(table_html_cleaned)
table_text = table_soup.get_text()
table_text = table_text[0:min(9999,len(table_text))]
data_table_ob = m.DataTable.objects.get_or_create(article = article_ob,
table_html = table_html_cleaned,
table_text = table_text,
uploading_user = uploading_user,
user_uploaded = user_uploaded
)[0]
data_table_ob = remove_spurious_table_headers(data_table_ob) # takes care of weird header thing for elsevier xml tables
ds = m.DataSource.objects.get_or_create(data_table=data_table_ob)[0]
# apply initial text mining of ephys concepts to table
if text_mine:
assocDataTableEphysVal(data_table_ob)
# creates data table stat object, relevance is to count and store num of unique ecms that were TMed
data_table_stat = update_data_table_stat(data_table_ob)
return data_table_ob
def ephys_table_identify():
artObs = m.Article.objects.filter(datatable__isnull = False, articlefulltext__isnull = False).distinct()
artObs = artObs.exclude(articlefulltext__articlefulltextstat__data_table_ephys_processed = True)
dataTableObs = m.DataTable.objects.filter(article__in = artObs).distinct()
num_tables = dataTableObs.count()
print 'analyzing %s tables' % num_tables
for i,dt in enumerate(dataTableObs):
prog(i, num_tables)
assocDataTableEphysVal(dt)
art = dt.article
aft_ob = art.get_full_text()
if aft_ob is not None:
aftStatOb = m.ArticleFullTextStat.objects.get_or_create(article_full_text = aft_ob)[0]
aftStatOb.data_table_ephys_processed = True
aftStatOb.save()
def ephys_table_identify_block(pk_inds):
dataTableObs = m.DataTable.objects.filter(pk__in = pk_inds).distinct()
num_tables = dataTableObs.count()
print 'analyzing %s tables in block' % num_tables
for i,dt in enumerate(dataTableObs):
#prog(i, num_tables)
assocDataTableEphysVal(dt)
art = dt.article
print art
aft_ob = art.get_full_text()
if aft_ob is not None:
aftStatOb = m.ArticleFullTextStat.objects.get_or_create(article_full_text = aft_ob)[0]
aftStatOb.data_table_ephys_processed = True
aftStatOb.save()
print i
html_tables = extract_tables_from_xml(aft.get_content(), file_name)
else:
html_tables = extract_tables_from_html(aft.get_content(), file_name)
for table in html_tables:
tableSoup = BeautifulSoup(table)
table_html = str(tableSoup)
table_html = add_id_tags_to_table(table_html)
table_text = tableSoup.get_text()
table_text = table_text[0:min(9999,len(table_text))]
data_table_ob = m.DataTable.objects.get_or_create(article = a, table_html = table_html, table_text = table_text)[0]
data_table_ob = remove_spurious_table_headers(data_table_ob) # takes care of weird header thing for elsevier xml tables
ds = m.DataSource.objects.get_or_create(data_table=data_table_ob)[0]
# apply initial text mining of ephys concepts to table
assocDataTableEphysVal(data_table_ob)
# text mine article level metadata
apply_article_metadata(a)
except Exception, e:
with open('failed_files.txt', 'a') as f:
f.write('%s\\%s' % (file_name, e))
print e
print file_name
finally:
f.close()
# if html_tables is not None:
# do a check to see if tables already exist, if do, just return
# if a.datatable_set.all().count() > 0:
# return a
