jprdonnelly/congress
1
0
Fork 0
mirror of https://github.com/unitedstates/congress.git synced 2025-12-25 01:00:04 -05:00
Code Issues Packages Projects Releases Wiki Activity

Handle multiple committees that a nomination is referred to (referred_to is now an array)

Browse Source
This commit is contained in:
Eric Mill
2013-09-06 18:20:43 -04:00
parent 20432c7f1b
commit 8533ec016b
1 changed files with 6 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
tasks/nomination_info.py
View File

@@ -60,7 +60,8 @@ def parse_nomination(nomination_id, body, options):
#the markup on these pages is a disaster, so we're going to use a heuristic based on boldface, inline tags followed by text
for pair in doc.xpath('//span[@class="elabel"]|//strong'):
if pair.tail:
label, data = pair.text.replace(':', '').strip(), pair.tail.strip()
text = pair.text or pair.text_content()
label, data = text.replace(':', '').strip(), pair.tail.strip()
# handle actions separately
if label.split(" ")[-1] == "Action":
@@ -90,8 +91,10 @@ def parse_nomination(nomination_id, body, options):
elif label == "Control Number":
info["control_number"] = data
elif label == "Referred to":
info["referred_to"] = data
elif label.lower() == "referred to":
if info.get("referred_to", None) == None:
info["referred_to"] = []
info["referred_to"].append(data)
elif label == "Reported by":
info["reported_by"] = data