phantop/quatalog-scraper
1
0
Fork 0
mirror of https://github.com/quatalog/quatalog.git synced 2025-11-29 15:55:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

Improve credit count parsing

Browse source
This commit is contained in:
powe97 2024-03-06 16:13:13 -06:00
parent 9c374bf130
commit 69d8946f37
No known key found for this signature in database
GPG key ID: 7D1663B10978D1BA
1 changed files with 4 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
transfer_scraper/main.py
View file

@ -232,9 +232,10 @@ def parse_course_td(td, note=None):
cr_delim = ( cr_delim = (
len(course_info) len(course_info)
- 1 - 1
- list(bool(re.search(r"\(", s)) for s in course_info[::-1]).index(True) - list(bool(re.search(r"^\([0-9]", s.strip())) for s in course_info[::-1]).index(True)
) )
except ValueError: assert bool(re.search(r"[0-9]\)"), course_info[-1])
except (ValueError, AssertionError):
cr_delim = len(course_info) cr_delim = len(course_info)
# note serves as a credit count override, since the RPI-side credit counts # note serves as a credit count override, since the RPI-side credit counts
@ -245,7 +246,7 @@ def parse_course_td(td, note=None):
"catalog": td.find_element(By.TAG_NAME, "span").text, "catalog": td.find_element(By.TAG_NAME, "span").text,
} }
if note is None: if note is None:
out.update({"credits": str(" ".join(course_info[cr_delim:])[1:-1])}), out.update({"credits": str(" ".join(course_info[cr_delim:])[1:-1]).strip()}),
return out return out
else: else:
out.update({"note": note}) out.update({"note": note})