def parse_portal_html(html):
"""
Given the Portal search results HTML as a string, return a list of
raw course data dictionaries.
If HTML is bad, raise ScrapeError.
"""
soup = bs4.BeautifulSoup(html, "lxml")
table = soup.find(id="tableCourses")
if not table:
raise ScrapeError("could not find course list table in Portal HTML")
table_body = table.find("tbody")
if not table_body:
raise ScrapeError(
"could not find course list table body in Portal HTML")
table_rows = table_body.find_all("tr", recursive=False)
if not table_rows:
raise ScrapeError(
"could not extract course list table rows from Portal HTML")
raw_courses = []
for row_idx, row in enumerate(table_rows):
if "style" in row.attrs and row.attrs["style"] == "display:none;":
continue
raw_courses.append(parse_table_row(row_idx, row))
return raw_courses
def lingk_csv_data_to_course_descriptions(data):
header, *rows = data
try:
course_code_idx = header.index("courseNumber")
desc_idx = header.index("description")
except ValueError:
raise ScrapeError(f"unexpected header: {header!r}") from None
desc_map = {}
for row in rows:
# We have some rows that are completely empty and some that
# are just whitespace.
if not row or "".join(row).isspace():
continue
if len(row) != len(header):
raise ScrapeError(f"malformed row: {row!r}")
course_code = row[course_code_idx]
try:
course_info = shared.parse_course_code(course_code,
with_section=False)
except ScrapeError:
continue
index_key = tuple(
shared.course_info_as_list(course_info, with_section=False))
description = row[desc_idx]
if not description:
continue
description = " ".join(description.split())
# If two conflicting descriptions for the same course code
# (yep, it happens), pick the one that comes later :/
desc_map[index_key] = description
if len(desc_map) < 100:
raise ScrapeError(
f"Not enough course descriptions: {len(desc_map)}") from None
return desc_map
def parse_table_row(row_idx, row):
"""
Given a Selenium table row and the index, return a dictionary
representing the raw course data for that row.
Raise ScrapeError if the HTML does not have the desired data.
"""
elements = row.find_all("td")
try:
(_add, course_code, name, faculty, seats, status, schedule,
num_credits, begin, end) = elements
except ValueError:
raise ScrapeError("could not extract course list table row elements "
"from Portal HTML (for row {})".format(row_idx))
return {
"course_code": course_code.text,
"course_name": name.text,
"faculty": faculty.text,
"seats": seats.text,
"status": status.text,
"schedule": [stime.text for stime in schedule.find_all("li")],
"credits": num_credits.text,
"begin_date": begin.text,
"end_date": end.text,
}
def get_course_descriptions():
"""
Given a Lingk API key and secret for authentication, return a
dictionary mapping course codes (as can be used on the frontend)
to course descriptions.
Throw ScrapeError if the API is not available or returns bad data.
"""
if util.get_env_boolean("lingk"):
key = os.environ.get("HYPERSCHEDULE_LINGK_KEY")
secret = os.environ.get("HYPERSCHEDULE_LINGK_SECRET")
if not key or not secret:
util.log("Skipping Lingk as key and secret are not set")
return {}
util.log_verbose("Scraping Lingk API")
data = get_lingk_api_data(key, secret)
desc_index = lingk_api_data_to_course_descriptions(data)
else:
util.log_verbose("Scraping Lingk CSV")
data = get_lingk_csv_data()
desc_index = lingk_csv_data_to_course_descriptions(data)
if len(desc_index) < 100:
raise ScrapeError("Not enough course descriptions: {}".format(
len(desc_index)))
return desc_index
def get_lingk_api_data(key, secret):
"""
Return the decoded JSON response from the Lingk API, using the
given key and secret for authentication.
Throw ScrapeError if the API is not available or returns bad data.
"""
# For some bizarre reason the Lingk API sometimes returns 401
# Unauthorized even when you are authenticated correctly. Asking
# again a few times fixes the issue. I don't even want to know
# why.
last_error = None
fails = 0
for i in range(LINGK_RETRY_COUNT):
now = datetime.datetime.utcnow()
date = now.strftime("%a, %d %b %Y %H:%M:%S UTC")
response = requests.get(LINGK_URL,
headers={
"Date":
date,
"Authorization":
get_auth_header(key, secret, date)
})
try:
response.raise_for_status()
return response.json()
except requests.exceptions.HTTPError as e:
fails += 1
util.log_verbose(
"Got auth error from Lingk API ({} of {} allowed)".format(
fails, LINGK_RETRY_COUNT))
time.sleep(1)
last_error = e
continue
except ValueError:
raise ScrapeError("Lingk API returned no data")
except json.decoder.JSONDecodeError:
raise ScrapeError("Lingk API did not return valid JSON")
raise ScrapeError(
"Lingk API returned error response: {}".format(last_error))
def get_portal_html(browser):
"""
Given a Selenium browser object, perform a webscrape of Portal.
Return a tuple (html, term) with the HTML of the course search
results page as a string and the current term (for which courses
were retrieved) also as a string.
Raise ScrapeError if something goes wrong with the browser or
Portal.
"""
util.log_verbose(f"Current scraper IP is {get_ip()}")
util.log_verbose("Scraping Portal")
browser.set_window_size(1920, 1200)
url = ("https://portal.hmc.edu/ICS/Portal_Homepage.jnz?"
"portlet=Course_Schedules&screen=Advanced+Course+Search"
"&screenType=next")
browser.get(url)
term_dropdown = selenium.webdriver.support.ui.Select(
browser.find_element_by_id("pg0_V_ddlTerm"))
term_names = [option.text for option in term_dropdown.options]
terms_info = []
for term_name in term_names:
match = re.match(r"\s*(FA|SP)\s*([0-9]{4})\s*", term_name)
if match:
fall_or_spring, year_str = match.groups()
terms_info.append(
(int(year_str), fall_or_spring == "FA", term_name))
if not terms_info:
raise ScrapeError(
f"couldn't parse any term names (from: {term_names!r})")
term_info = max(terms_info)
term = term_info[2]
term_dropdown.select_by_visible_text(term)
title_input = browser.find_element_by_id("pg0_V_txtTitleRestrictor")
title_input.clear()
title_input.send_keys("?")
search_button = browser.find_element_by_id("pg0_V_btnSearch")
search_button.click()
show_all_checkbox = browser.find_element_by_id("pg0_V_lnkShowAll")
show_all_checkbox.click()
return browser.page_source, " ".join(term.split())
def lingk_api_data_to_course_descriptions(data):
"""
Given the decoded JSON from the Lingk API, return a dictionary
mapping tuples of course information (`with_section` false; see
`shared.course_info_as_list`) to course descriptions.
Throw ScrapeError if the data is malformed.
"""
if not isinstance(data, dict):
raise ScrapeError("Lingk JSON is not map")
if "data" not in data:
raise ScrapeError("Lingk JSON is missing 'data' field")
desc_index = {}
for idx, course in enumerate(data["data"]):
if "description" not in course:
continue
description = course["description"]
if not isinstance(description, str):
raise ScrapeError(
"'description' at index {} is not string".format(idx))
if "courseNumber" not in course:
raise ScrapeError(
"Lingk JSON at index {} is missing 'courseNumber' field".
format(idx))
course_code = course["courseNumber"]
# Special case that doesn't show up on Portal.
if course_code == "ABROAD HM":
continue
course_info = shared.parse_course_code(course_code, with_section=False)
course_key = tuple(
shared.course_info_as_list(course_info, with_section=False))
found_mismatch = (course_key in desc_index
and desc_index[course_key] != description)
if found_mismatch:
raise ScrapeError("Lingk JSON has duplicate course: {}".format(
repr(course_key)))
desc_index[course_key] = description
return desc_index
def try_compute_data(s3, webhook, old_data):
"""
Try to run the scraper and return course data. If something goes
wrong, raise `ScrapeError`. Otherwise, invoke the provided
`Webhook`. `old_data` is the previous course data or `util.Unset`.
"""
scraper_timeout = util.get_env("scraper_timeout")
try:
scraper_timeout = int(scraper_timeout)
if scraper_timeout <= 0:
raise ValueError
except ValueError:
util.warn("Illegal scraper timeout: {}".format(repr(scraper_timeout)))
util.log("Resetting timeout to 60 seconds")
os.environ["HYPERSCHEDULE_SCRAPER_TIMEOUT"] = "60"
scraper_timeout = 60
if old_data is util.Unset:
# For JSON.
old_data = None
try:
util.log("Running scraper")
process = subprocess.Popen(
["python", "-m", "hyperschedule.scrapers.claremont"],
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
)
output, _ = process.communicate(input=json.dumps(old_data).encode(),
timeout=scraper_timeout)
if process.returncode != 0:
raise ScrapeError("scraper failed")
try:
output = output.decode()
except UnicodeDecodeError as e:
raise ScrapeError(
"scraper emitted malformed output: {}".format(e)) from None
if "$delete" in output:
raise ScrapeError("scraper output contains '$delete'")
data = json.loads(output)
if util.get_env_boolean("snitch"):
webhook.get()
if util.get_env_boolean("cache"):
cache_file_write(data)
if util.get_env_boolean("s3_write"):
s3_write(s3, data)
except OSError as e:
raise ScrapeError(
"unexpected error while running scraper: {}".format(e)) from None
except subprocess.TimeoutExpired:
process.kill()
process.communicate()
raise ScrapeError("scraper timed out after {} seconds".format(
scraper_timeout)) from None
except json.decoder.JSONDecodeError:
raise ScrapeError("scraper did not return valid JSON") from None
except requests.exceptions.RequestException as e:
util.warn("failed to reach success webhook: {}".format(e))
return data
def parse_table_row(row_idx, row):
"""
Given a Selenium table row and the index, return a dictionary
representing the raw course data for that row.
Raise ScrapeError if the HTML does not have the desired data.
"""
elements = row.find_all("td")
try:
(
_tb,
_add,
course_code,
name,
_req,
_note,
seats,
status,
faculty_and_schedule,
num_credits,
begin,
end,
) = elements
except ValueError:
raise ScrapeError("could not extract course list table row elements "
f"from Portal HTML (for row {row_idx})")
all_faculty = []
schedule = []
for item in faculty_and_schedule.find_all("li"):
try:
faculty, meeting = item.text.split(" / ")
# This list gets uniquified later.
all_faculty.append(faculty)
schedule.append(meeting)
except ValueError:
# No "/" separator, assumed to mean only schedule (no
# faculty).
schedule.append(item.text)
return {
"course_code": course_code.text,
"course_name": name.text,
"faculty": all_faculty,
"seats": seats.text,
"status": status.text,
"schedule": schedule,
"credits": num_credits.text,
"begin_date": begin.text,
"end_date": end.text,
}
def parse_term_code(term):
"""
Given a term code (e.g. "FA 2018"), return a dictionary with keys:
* year (integer)
* fall (boolean)
* spring (boolean)
"""
match = re.match(r"(FA|SP)\s*(20[0-9]{2})", term)
if not match:
raise ScrapeError("malformed term code: {}".format(repr(term)))
return {
"year": int(match.group(2)),
"fall": match.group(1) == "FA",
"spring": match.group(1) == "SP",
}
def get_courses(desc_index):
"""
Return a tuple containing the list of course objects and the
current term. Takes `desc_index` as returned by
`lingk.get_course_descriptions`.
"""
browser = get_browser()
html, term = get_portal_html(browser)
# Save on memory.
scraper.kill_google_chrome()
# Count how many courses we add descriptions to, so we can fail if
# there aren't enough.
num_descs_added = 0
# Count how many courses we fail to parse, so we can fail if there
# are too many.
num_failed = 0
# Get the first round of raw courses from Portal.
raw_courses_1 = parse_portal_html(html)
# Add course descriptions to them, using the raw course codes.
# Also collect the course codes into a dictionary so that we can
# deduplicate them.
raw_courses_2 = []
course_info_map = collections.defaultdict(list)
for raw_course in raw_courses_1:
try:
course_code = raw_course["course_code"].strip()
course_info = shared.parse_course_code(course_code,
with_section=True)
desc_key = tuple(
shared.course_info_as_list(course_info, with_section=False))
desc = desc_index.get(desc_key)
if desc:
num_descs_added += 1
raw_course["course_description"] = desc
course_info_map[frozendict.frozendict(course_info)].append(
raw_course)
except ScrapeError as err:
util.log_verbose(
f"Failed to parse course: {format_raw_course(raw_course)!r} ({err})"
)
num_failed += 1
continue
raw_courses_2.append(raw_course)
if num_descs_added < 100:
raise ScrapeError(
f"not enough course descriptions added: {num_descs_added}")
# Deduplicate course codes.
raw_courses_3 = []
for course_info, courses in course_info_map.items():
if len(courses) > 1:
if course_info["course_code_suffix"]:
util.log_verbose(
f"Duplicate course with suffix ({len(courses)} copies): "
f"{format_raw_course(courses[0])!r}")
num_failed += len(courses)
continue
if len(courses) > len(string.ascii_uppercase):
util.log_verbose(
f"Duplicate course with too many copies ({len(courses)}): "
f"{format_raw_course(courses[0])!r}")
num_failed += len(courses)
continue
for course, letter in zip(courses, string.ascii_uppercase):
course["course_code_suffix"] = letter
raw_courses_3.extend(courses)
raw_courses = raw_courses_3
courses = []
for raw_course in raw_courses:
try:
courses.append(process_course(raw_course, term))
except ScrapeError as err:
util.log_verbose(
f"Failed to parse course: {format_raw_course(raw_course)!r} ({err})"
)
num_failed += 1
if num_failed >= 10:
raise ScrapeError(f"Too many malformed courses: {num_failed}")
num_succeeded = len(raw_courses) - num_failed
if num_succeeded < 500:
raise ScrapeError(f"Not enough courses: {num_succeeded}")
util.log_verbose(
f"Added descriptions to {num_descs_added} out of {num_succeeded} courses"
)
return courses, term
def process_course(raw_course, term):
"""
Turn a raw course object into something that the frontend can use.
Return a dictionary.
If the raw course object has invalid data, raise ScrapeError.
"""
course_code = raw_course["course_code"].strip()
course_info = shared.parse_course_code(course_code, with_section=True)
course_code = shared.course_info_as_string(course_info)
sort_key = shared.course_info_as_list(course_info, with_section=True)
mutual_exclusion_key = shared.course_info_as_list(course_info,
with_section=False)
course_name = raw_course["course_name"].strip()
if not course_name:
raise ScrapeError("empty string for course name")
faculty = sorted(set(f.strip() for f in raw_course["faculty"]))
if not faculty:
raise ScrapeError("no faculty")
for faculty_name in faculty:
if not faculty_name:
raise ScrapeError("faculty with empty name")
try:
# careful: "∕" (`chr(8725)`) != "/" (`chr(47)`)
filled_seats, total_seats = map(int, raw_course["seats"].split("∕"))
except ValueError as err:
raise ScrapeError(
f"malformed seat count: {raw_course['seats']!r} ({err})")
if filled_seats < 0:
raise ScrapeError(f"negative filled seat count: {filled_seats}")
if total_seats < 0:
raise ScrapeError(f"negative total seat count: {total_seats}")
course_status = raw_course["status"].lower().strip()
if course_status not in ("open", "closed", "reopened"):
raise ScrapeError(f"unknown course status: {course_status!r}")
begin_date = dateutil.parser.parse(raw_course["begin_date"]).date()
end_date = dateutil.parser.parse(raw_course["end_date"]).date()
# First half-semester courses start (spring) January 1 through
# January 31 or (fall) July 15 through September 15. (For some
# reason, MATH 30B in Fall 2017 is listed as starting August 8.)
first_half = datetime.date(
begin_date.year, 1, 1) < begin_date < datetime.date(
begin_date.year, 1, 31) or datetime.date(
begin_date.year, 7, 15) < begin_date < datetime.date(
begin_date.year, 9, 15)
# Second half-semester courses for the spring end May 1 through
# May 31, but there's also frosh chem pt.II which just *has* to be
# different by ending 2/3 of the way through the semester. So we
# also count that by allowing April 1 through April 30. Sigh. Fall
# courses end December 1 through December 31.
second_half = datetime.date(
end_date.year, 4,
1) < end_date < datetime.date(end_date.year, 5, 31) or datetime.date(
end_date.year, 12, 1) < end_date < datetime.date(
end_date.year, 12, 31)
if first_half and second_half:
term_count = 1
terms = [0]
elif first_half and not second_half:
term_count = 2
terms = [0]
elif second_half and not first_half:
term_count = 2
terms = [1]
else:
raise ScrapeError(
f"weird date range "
f"{begin.date.strftime('%F')}-{end_date.strftime('%F')}")
schedule = []
for slot in raw_course["schedule"]:
if re.match(r"To Be Arranged\xa00?0:00 ?- ?0?0:00 ?AM", slot):
continue
match = re.match(SCHEDULE_REGEX, slot)
if not match:
raise ScrapeError(f"malformed schedule slot: {slot!r}")
days, start, end, location = match.groups()
for day in days:
if day not in DAYS_OF_WEEK:
raise ScrapeError(
f"unknown day of week {day!r} in schedule slot {slot!r}")
days = "".join(sorted(set(days), key=DAYS_OF_WEEK.index))
if not days:
raise ScrapeError(f"no days in schedule slot {slot!r}")
if not (start.endswith("AM") or start.endswith("PM")):
start += end[-2:]
try:
start = dateutil.parser.parse(start).time()
except ValueError:
raise ScrapeError(
f"malformed start time {start!r} in schedule slot {slot!r}")
try:
end = dateutil.parser.parse(end).time()
except ValueError:
raise ScrapeError(
f"malformed end time {end!r} in schedule slot {slot!r}")
location = " ".join(location.strip().split())
if not location:
raise ScrapeError("empty string for location")
# Start using camelCase here because we are constructing
# objects that will be returned from the API as JSON -- no
# longer just intermediate objects private to this module.
schedule.append({
"scheduleDays": days,
"scheduleStartTime": start.strftime("%H:%M"),
"scheduleEndTime": end.strftime("%H:%M"),
"scheduleStartDate": begin_date.strftime("%Y-%m-%d"),
"scheduleEndDate": end_date.strftime("%Y-%m-%d"),
"scheduleTermCount": term_count,
"scheduleTerms": terms,
"scheduleLocation": location,
})
if not schedule:
schedule.append({
"scheduleDays": "",
"scheduleStartTime": "00:00",
"scheduleEndTime": "00:00",
"scheduleStartDate": begin_date.strftime("%Y-%m-%d"),
"scheduleEndDate": end_date.strftime("%Y-%m-%d"),
"scheduleTermCount": term_count,
"scheduleTerms": terms,
"scheduleLocation": "N/A",
})
schedule = unique_preserve_order(schedule)
num_credits = raw_course["credits"]
try:
num_credits = float(num_credits)
except ValueError:
raise ScrapeError(f"malformed credit count: {num_credits!r}")
if num_credits < 0.0:
raise ScrapeError(f"negative credit count: {raw_course['credits']}")
if "Colloquium" in course_name and num_credits == 0.0:
num_credits = 0.5
elif re.match("PE ", course_code) and num_credits == 0.0:
num_credits = 1.0
elif num_credits == 0.25:
num_credits = 1.0
elif not re.search(r"HM-", course_code):
num_credits *= 3.0
if num_credits == 9.0:
num_credits = 3.0
num_credits = str(num_credits)
course_description = raw_course["course_description"]
# just urls for now - we could add ratings or would take again percentages later
urls = []
for prof in faculty:
a = RateMyProfAPI(teacher=prof)
# scrape the info from RateMyProfessors site
a.fetch_info()
urls.append(a.get_url())
return {
"courseCode": course_code,
"courseName": course_name,
"courseSortKey": sort_key,
"courseMutualExclusionKey": mutual_exclusion_key,
"courseDescription": course_description,
"courseInstructors": faculty,
"courseInstructorRMPpages": urls,
"courseTerm": term,
"courseSchedule": schedule,
"courseCredits": num_credits,
"courseSeatsTotal": total_seats,
"courseSeatsFilled": filled_seats,
"courseWaitlistLength": None,
"courseEnrollmentStatus": course_status,
}
def parse_course_code(course_code, with_section):
"""
Given a course code in the format used by Portal and Lingk, with
or without a section number ("PHIL 179A HM-01" or just "PHIL179A
HM") as controlled by `with_section`, parse it and return a
dictionary with keys:
- department (string)
- course_number (integer)
- course_code_suffix (string)
- school (string)
- section (integer, or null if with_section is false)
The given course code may also be in the format returned by
`course_info_as_string`.
Throw ScrapeError if parsing fails.
"""
match = re.match(COURSE_REGEX, course_code)
if not match:
raise ScrapeError("malformed course code: {}".format(
repr(course_code)))
department, course_number, num_suffix, school, section = match.groups()
if not department:
raise ScrapeError("empty string for department")
if "/" in department:
raise ScrapeError("department contains slashes: {}".format(
repr(department)))
try:
course_number = int(course_number)
except ValueError:
raise ScrapeError("malformed course number: {}".format(
repr(course_number)))
if course_number <= 0:
raise ScrapeError(
"non-positive course number: {}".format(course_number))
if "/" in num_suffix:
raise ScrapeError("course code suffix contains slashes: {}".format(
repr(num_suffix)))
if not school:
raise ScrapeError("empty string for school")
if "/" in school:
raise ScrapeError("school contains slashes: {}".format(repr(school)))
if bool(section) != bool(with_section):
if with_section:
raise ScrapeError("section missing")
else:
raise ScrapeError("section unexpectedly present: {}".format(
repr(section)))
if section:
try:
section = int(section)
except ValueError:
raise ScrapeError("malformed section number: {}".format(
repr(section)))
if section <= 0:
raise ScrapeError(
"non-positive section number: {}".format(section))
# If section is None, just leave it as is.
return {
"department": department,
"courseNumber": course_number,
"courseCodeSuffix": num_suffix,
"school": school,
"section": section,
}
def process_course(raw_course, term):
"""
Turn a raw course object into something that the frontend can use.
Return a dictionary.
If the raw course object has invalid data, raise ScrapeError.
"""
course_code = raw_course["course_code"].strip()
course_info = shared.parse_course_code(course_code, with_section=True)
course_code = shared.course_info_as_string(course_info)
sort_key = shared.course_info_as_list(course_info, with_section=True)
mutual_exclusion_key = shared.course_info_as_list(course_info,
with_section=False)
course_name = raw_course["course_name"].strip()
if not course_name:
raise ScrapeError("empty string for course name")
faculty = sorted(set(re.split(r"\s*\n\s*", raw_course["faculty"].strip())))
if not faculty:
raise ScrapeError("no faculty")
for faculty_name in faculty:
if not faculty_name:
raise ScrapeError("faculty with empty name")
match = re.match(r"([0-9]+)/([0-9]+)", raw_course["seats"])
if not match:
raise ScrapeError("malformed seat count: {}".format(
repr(raw_course["seats"])))
filled_seats, total_seats = map(int, match.groups())
if filled_seats < 0:
raise ScrapeError(
"negative filled seat count: {}".format(filled_seats))
if total_seats < 0:
raise ScrapeError("negative total seat count: {}".format(total_seats))
course_status = raw_course["status"].lower()
if course_status not in ("open", "closed", "reopened"):
raise ScrapeError("unknown course status: {}".format(
repr(course_status)))
begin_date = dateutil.parser.parse(raw_course["begin_date"]).date()
end_date = dateutil.parser.parse(raw_course["end_date"]).date()
# First half-semester courses start (spring) January 1 through
# January 31 or (fall) July 15 through September 15. (For some
# reason, MATH 30B in Fall 2017 is listed as starting August 8.)
first_half = (datetime.date(begin_date.year, 1, 1) < begin_date <
datetime.date(begin_date.year, 1, 31)
or datetime.date(begin_date.year, 7, 15) < begin_date <
datetime.date(begin_date.year, 9, 15))
# Second half-semester courses for the spring end May 1 through
# May 31, but there's also frosh chem pt.II which just *has* to be
# different by ending 2/3 of the way through the semester. So we
# also count that by allowing April 1 through April 30. Sigh. Fall
# courses end December 1 through December 31.
second_half = (datetime.date(
end_date.year, 4, 1) < end_date < datetime.date(end_date.year, 5, 31)
or datetime.date(end_date.year, 12, 1) < end_date <
datetime.date(end_date.year, 12, 31))
if first_half and second_half:
term_count = 1
terms = [0]
elif first_half and not second_half:
term_count = 2
terms = [0]
elif second_half and not first_half:
term_count = 2
terms = [1]
else:
raise ScrapeError("weird date range {}-{}".format(
begin_date.strftime("%Y-%m-%d"), end_date.strftime("%Y-%m-%d")))
schedule = []
for slot in raw_course["schedule"]:
if slot.startswith("0:00 - 0:00 AM"):
continue
match = re.match(SCHEDULE_REGEX, slot)
if not match:
raise ScrapeError("malformed schedule slot: {}".format(repr(slot)))
days, start, end, location = match.groups()
for day in days:
if day not in DAYS_OF_WEEK:
raise ScrapeError(
"unknown day of week {} in schedule slot {}".format(
repr(day), repr(slot)))
days = "".join(sorted(set(days), key=DAYS_OF_WEEK.index))
if not days:
raise ScrapeError("no days in schedule slot {}".format(repr(slot)))
if not (start.endswith("AM") or start.endswith("PM")):
start += end[-2:]
try:
start = dateutil.parser.parse(start).time()
except ValueError:
raise ScrapeError(
"malformed start time {} in schedule slot {}".format(
repr(start), repr(slot)))
try:
end = dateutil.parser.parse(end).time()
except ValueError:
raise ScrapeError(
"malformed end time {} in schedule slot {}".format(
repr(end), repr(slot)))
location = " ".join(location.strip().split())
if not location:
raise ScrapeError("empty string for location")
# Start using camelCase here because we are constructing
# objects that will be returned from the API as JSON -- no
# longer just intermediate objects private to this module.
schedule.append({
"scheduleDays": days,
"scheduleStartTime": start.strftime("%H:%M"),
"scheduleEndTime": end.strftime("%H:%M"),
"scheduleStartDate": begin_date.strftime("%Y-%m-%d"),
"scheduleEndDate": end_date.strftime("%Y-%m-%d"),
"scheduleTermCount": term_count,
"scheduleTerms": terms,
"scheduleLocation": location,
})
if not schedule:
schedule.append({
"scheduleDays": "",
"scheduleStartTime": "00:00",
"scheduleEndTime": "00:00",
"scheduleStartDate": begin_date.strftime("%Y-%m-%d"),
"scheduleEndDate": end_date.strftime("%Y-%m-%d"),
"scheduleTermCount": term_count,
"scheduleTerms": terms,
"scheduleLocation": "N/A",
})
schedule = unique_preserve_order(schedule)
num_credits = raw_course["credits"]
try:
num_credits = float(num_credits)
except ValueError:
raise ScrapeError("malformed credit count: {}".format(
repr(num_credits)))
if num_credits < 0.0:
raise ScrapeError("negative credit count: {}".format(
raw_course["credits"]))
if "Colloquium" in course_name and num_credits == 0.0:
num_credits = 0.5
elif re.match("PE ", course_code) and num_credits == 0.0:
num_credits = 1.0
elif num_credits == 0.25:
num_credits = 1.0
elif not re.search(r"HM-", course_code):
num_credits *= 3.0
if num_credits == 9.0:
num_credits = 3.0
num_credits = str(num_credits)
course_description = raw_course["course_description"]
return {
"courseCode": course_code,
"courseName": course_name,
"courseSortKey": sort_key,
"courseMutualExclusionKey": mutual_exclusion_key,
"courseDescription": course_description,
"courseInstructors": faculty,
"courseTerm": term,
"courseSchedule": schedule,
"courseCredits": num_credits,
"courseSeatsTotal": total_seats,
"courseSeatsFilled": filled_seats,
"courseWaitlistLength": None,
"courseEnrollmentStatus": course_status,
}
