ilias crawler: Add some unhelpful documentation

This commit is contained in:
I-Al-Istannen 2020-04-22 17:58:19 +02:00
parent eb7df036df
commit a051e3bcca

View File

@ -48,6 +48,9 @@ class IliasCrawler:
@staticmethod @staticmethod
def _url_set_query_param(url: str, param: str, value: str) -> str: def _url_set_query_param(url: str, param: str, value: str) -> str:
"""
Set a query parameter in an url, overwriting existing ones with the same name.
"""
scheme, netloc, path, query, fragment = urlsplit(url) scheme, netloc, path, query, fragment = urlsplit(url)
query_parameters = parse_qs(query) query_parameters = parse_qs(query)
query_parameters[param] = [value] query_parameters[param] = [value]
@ -59,10 +62,13 @@ class IliasCrawler:
""" """
Starts the crawl process, yielding a list of elements to (potentially) download. Starts the crawl process, yielding a list of elements to (potentially) download.
""" """
# Start crawling at the given course
root_url = self._url_set_query_param( root_url = self._url_set_query_param(
self._base_url + "/goto.php", "target", f"crs_{self._course_id}" self._base_url + "/goto.php", "target", f"crs_{self._course_id}"
) )
# And treat it as a folder
return self._crawl_folder(Path(""), root_url) return self._crawl_folder(Path(""), root_url)
def _switch_on_crawled_type( def _switch_on_crawled_type(
@ -77,6 +83,7 @@ class IliasCrawler:
parsed_url = urlparse(url) parsed_url = urlparse(url)
LOGGER.debug("Parsed url: %r", parsed_url) LOGGER.debug("Parsed url: %r", parsed_url)
# file URLs contain "target=file"
if "target=file_" in parsed_url.query: if "target=file_" in parsed_url.query:
LOGGER.debug("Interpreted as file.") LOGGER.debug("Interpreted as file.")
return self._crawl_file(path, link_element, url) return self._crawl_file(path, link_element, url)
@ -86,6 +93,8 @@ class IliasCrawler:
LOGGER.debug("Skipping forum %r", url) LOGGER.debug("Skipping forum %r", url)
return [] return []
# Everything with a ref_id can *probably* be opened to reveal nested things
# video groups, directories, exercises, etc
if "ref_id=" in parsed_url.query: if "ref_id=" in parsed_url.query:
LOGGER.debug("Processing folder-like...") LOGGER.debug("Processing folder-like...")
return self._switch_on_folder_like(path, link_element, url) return self._switch_on_folder_like(path, link_element, url)
@ -99,11 +108,17 @@ class IliasCrawler:
""" """
Crawls a file. Crawls a file.
""" """
# Files have a list of properties (type, modification date, size, etc.)
# In a series of divs.
# Find the parent containing all those divs, so we can filter our what we need
properties_parent: bs4.Tag = link_element.findParent( properties_parent: bs4.Tag = link_element.findParent(
"div", {"class": lambda x: "il_ContainerListItem" in x} "div", {"class": lambda x: "il_ContainerListItem" in x}
).select_one(".il_ItemProperties") ).select_one(".il_ItemProperties")
# The first one is always the filetype
file_type = properties_parent.select_one("span.il_ItemProperty").getText().strip() file_type = properties_parent.select_one("span.il_ItemProperty").getText().strip()
# The rest does not have a stable order. Grab the whole text and reg-ex the date
# out of it
all_properties_text = properties_parent.getText().strip() all_properties_text = properties_parent.getText().strip()
modification_date_match = re.search( modification_date_match = re.search(
r"(((\d+\. \w+ \d+)|(Gestern)|(Heute)), \d+:\d+)", r"(((\d+\. \w+ \d+)|(Gestern)|(Heute)), \d+:\d+)",
@ -116,15 +131,10 @@ class IliasCrawler:
modification_date_str = modification_date_match.group(1) modification_date_str = modification_date_match.group(1)
modification_date = demangle_date(modification_date_str) modification_date = demangle_date(modification_date_str)
# Grab the name from the link text
name = link_element.getText() name = link_element.getText()
full_path = Path(path, name + "." + file_type) full_path = Path(path, name + "." + file_type)
match_result = re.match(r".+target=file_(\d+).+", url)
if match_result is None:
LOGGER.warning("Could not download file %r", url)
return []
return [IliasDownloadInfo(full_path, url, modification_date)] return [IliasDownloadInfo(full_path, url, modification_date)]
def _switch_on_folder_like( def _switch_on_folder_like(
@ -133,8 +143,13 @@ class IliasCrawler:
link_element: bs4.Tag, link_element: bs4.Tag,
url: str url: str
) -> List[IliasDownloadInfo]: ) -> List[IliasDownloadInfo]:
"""
Try crawling something that looks like a folder.
"""
found_parent: Optional[bs4.Tag] = None found_parent: Optional[bs4.Tag] = None
# We look for the outer div of our inner link, to find information around it
# (mostly the icon)
for parent in link_element.parents: for parent in link_element.parents:
if "ilContainerListItemOuter" in parent["class"]: if "ilContainerListItemOuter" in parent["class"]:
found_parent = parent found_parent = parent
@ -144,22 +159,26 @@ class IliasCrawler:
LOGGER.warning("Could not find element icon for %r", url) LOGGER.warning("Could not find element icon for %r", url)
return [] return []
# Find the small descriptive icon to figure out the type
img_tag: Optional[bs4.Tag] = found_parent.select_one("img.ilListItemIcon") img_tag: Optional[bs4.Tag] = found_parent.select_one("img.ilListItemIcon")
if img_tag is None: if img_tag is None:
LOGGER.warning("Could not find image tag for %r", url) LOGGER.warning("Could not find image tag for %r", url)
return [] return []
# A forum
if str(img_tag["src"]).endswith("frm.svg"): if str(img_tag["src"]).endswith("frm.svg"):
LOGGER.debug("Skipping forum at %r", url) LOGGER.debug("Skipping forum at %r", url)
return [] return []
element_path = Path(path, link_element.getText().strip()) element_path = Path(path, link_element.getText().strip())
# An exercise
if str(img_tag["src"]).endswith("icon_exc.svg"): if str(img_tag["src"]).endswith("icon_exc.svg"):
LOGGER.debug("Crawling exercises at %r", url) LOGGER.debug("Crawling exercises at %r", url)
return self._crawl_exercises(element_path, url) return self._crawl_exercises(element_path, url)
# Match the opencast video plugin
if "opencast" in str(img_tag["alt"]).lower(): if "opencast" in str(img_tag["alt"]).lower():
LOGGER.debug("Found video site: %r", url) LOGGER.debug("Found video site: %r", url)
return self._crawl_video_directory(element_path, url) return self._crawl_video_directory(element_path, url)
@ -168,13 +187,22 @@ class IliasCrawler:
return self._crawl_folder(element_path, self._abs_url_from_link(link_element)) return self._crawl_folder(element_path, self._abs_url_from_link(link_element))
def _crawl_video_directory(self, path: Path, url: str) -> List[IliasDownloadInfo]: def _crawl_video_directory(self, path: Path, url: str) -> List[IliasDownloadInfo]:
"""
Crawl the video overview site.
"""
initial_soup = self._get_page(url, {}) initial_soup = self._get_page(url, {})
# The page is actually emtpy but contains a much needed token in the link below.
# That token can be used to fetch the *actual* video listing
content_link: bs4.Tag = initial_soup.select_one("#tab_series a") content_link: bs4.Tag = initial_soup.select_one("#tab_series a")
# Fetch the actual video listing. The given parameters return all videos (max 800)
# in a standalone html page
video_list_soup = self._get_page( video_list_soup = self._get_page(
self._abs_url_from_link(content_link), self._abs_url_from_link(content_link),
{"limit": 800, "cmd": "asyncGetTableGUI", "cmdMode": "asynch"} {"limit": 800, "cmd": "asyncGetTableGUI", "cmdMode": "asynch"}
) )
# Video start links are marked with an "Abspielen" link
video_links: List[bs4.Tag] = video_list_soup.findAll( video_links: List[bs4.Tag] = video_list_soup.findAll(
name="a", text=re.compile(r"\s*Abspielen\s*") name="a", text=re.compile(r"\s*Abspielen\s*")
) )
@ -187,8 +215,13 @@ class IliasCrawler:
return results return results
def _crawl_single_video(self, path: Path, link: bs4.Tag) -> List[IliasDownloadInfo]: def _crawl_single_video(self, path: Path, link: bs4.Tag) -> List[IliasDownloadInfo]:
"""
Crawl a single video based on its "Abspielen" link from the video listing.
"""
video_page_url = self._abs_url_from_link(link) video_page_url = self._abs_url_from_link(link)
# The link is part of a table with multiple columns, describing metadata.
# 6th child (1 indexed) is the modification time string
modification_string = link.parent.parent.parent.select_one( modification_string = link.parent.parent.parent.select_one(
"td.std:nth-child(6)" "td.std:nth-child(6)"
).getText().strip() ).getText().strip()
@ -199,6 +232,11 @@ class IliasCrawler:
).getText().strip() ).getText().strip()
title += ".mp4" title += ".mp4"
# Fetch the actual video page. This is a small wrapper page initializing a javscript
# player. Sadly we can not execute that JS. The actual video stream url is nowhere
# on the page, but defined in a JS object inside a script tag, passed to the player
# library.
# We do the impossible and RegEx the stream JSON object out of the page's HTML source
video_page_soup = self._get_page(video_page_url, {}) video_page_soup = self._get_page(video_page_url, {})
regex: re.Pattern = re.compile( regex: re.Pattern = re.compile(
r"({\"streams\"[\s\S]+?),\s*{\"paella_config_file", re.IGNORECASE r"({\"streams\"[\s\S]+?),\s*{\"paella_config_file", re.IGNORECASE
@ -210,35 +248,49 @@ class IliasCrawler:
return [] return []
json_str = json_match.group(1) json_str = json_match.group(1)
# parse it
json_object = json.loads(json_str) json_object = json.loads(json_str)
# and fetch the video url!
video_url = json_object["streams"][0]["sources"]["mp4"][0]["src"] video_url = json_object["streams"][0]["sources"]["mp4"][0]["src"]
return [IliasDownloadInfo(Path(path, title), video_url, modification_time)] return [IliasDownloadInfo(Path(path, title), video_url, modification_time)]
def _crawl_exercises(self, element_path: Path, url: str) -> List[IliasDownloadInfo]: def _crawl_exercises(self, element_path: Path, url: str) -> List[IliasDownloadInfo]:
"""
Crawl files offered for download in exercises.
"""
soup = self._get_page(url, {}) soup = self._get_page(url, {})
results: List[IliasDownloadInfo] = [] results: List[IliasDownloadInfo] = []
# Each assignment is in an accordion container
assignment_containers: List[bs4.Tag] = soup.select(".il_VAccordionInnerContainer") assignment_containers: List[bs4.Tag] = soup.select(".il_VAccordionInnerContainer")
for container in assignment_containers: for container in assignment_containers:
# Fetch the container name out of the header to use it in the path
container_name = container.select_one(".ilAssignmentHeader").getText().strip() container_name = container.select_one(".ilAssignmentHeader").getText().strip()
# Find all download links in the container (this will contain all the files)
files: List[bs4.Tag] = container.findAll( files: List[bs4.Tag] = container.findAll(
name="a", name="a",
# download links contain the given command class
attrs={"href": lambda x: x and "cmdClass=ilexsubmissiongui" in x}, attrs={"href": lambda x: x and "cmdClass=ilexsubmissiongui" in x},
text="Download" text="Download"
) )
LOGGER.debug("Found exercise container %r", container_name) LOGGER.debug("Found exercise container %r", container_name)
# Parse the end date to use as modification date
# TODO: Return None?
end_date: datetime.datetime = datetime.datetime.now() end_date: datetime.datetime = datetime.datetime.now()
end_date_header: bs4.Tag = container.find(name="div", text="Abgabetermin") end_date_header: bs4.Tag = container.find(name="div", text="Abgabetermin")
if end_date_header is not None: if end_date_header is not None:
end_date_text = end_date_header.findNext("div").getText().strip() end_date_text = end_date_header.findNext("div").getText().strip()
end_date = demangle_date(end_date_text) end_date = demangle_date(end_date_text)
# Grab each file as you now have the link
for file_link in files: for file_link in files:
# Two divs, side by side. Left is the name, right is the link ==> get left
# sibling
file_name = file_link.parent.findPrevious(name="div").getText().strip() file_name = file_link.parent.findPrevious(name="div").getText().strip()
url = self._abs_url_from_link(file_link) url = self._abs_url_from_link(file_link)
@ -253,10 +305,14 @@ class IliasCrawler:
return results return results
def _crawl_folder(self, path: Path, url: str) -> List[IliasDownloadInfo]: def _crawl_folder(self, path: Path, url: str) -> List[IliasDownloadInfo]:
"""
Crawl all files in a folder-like element.
"""
soup = self._get_page(url, {}) soup = self._get_page(url, {})
result: List[IliasDownloadInfo] = [] result: List[IliasDownloadInfo] = []
# Fetch all links and throw them to the general interpreter
links: List[bs4.Tag] = soup.select("a.il_ContainerItemTitle") links: List[bs4.Tag] = soup.select("a.il_ContainerItemTitle")
for link in links: for link in links:
abs_url = self._abs_url_from_link(link) abs_url = self._abs_url_from_link(link)
@ -292,10 +348,13 @@ class IliasCrawler:
@staticmethod @staticmethod
def _is_logged_in(soup: bs4.BeautifulSoup) -> bool: def _is_logged_in(soup: bs4.BeautifulSoup) -> bool:
# Normal ILIAS pages
userlog = soup.find("li", {"id": "userlog"}) userlog = soup.find("li", {"id": "userlog"})
if userlog is not None: if userlog is not None:
LOGGER.debug("Auth: Found #userlog") LOGGER.debug("Auth: Found #userlog")
return True return True
# Video listing embeds do not have complete ILIAS html. Try to match them by
# their video listing table
video_table = soup.find( video_table = soup.find(
recursive=True, recursive=True,
name="table", name="table",
@ -304,6 +363,8 @@ class IliasCrawler:
if video_table is not None: if video_table is not None:
LOGGER.debug("Auth: Found #tbl_xoct.+") LOGGER.debug("Auth: Found #tbl_xoct.+")
return True return True
# The individual video player wrapper page has nothing of the above.
# Match it by its playerContainer.
if soup.select_one("#playerContainer") is not None: if soup.select_one("#playerContainer") is not None:
LOGGER.debug("Auth: Found #playerContainer") LOGGER.debug("Auth: Found #playerContainer")
return True return True