mirrors/pferd
1
0
Fork 0
mirror of https://github.com/Garmelon/PFERD.git synced 2023-12-21 10:23:01 +01:00
Code Issues Packages Projects Releases Wiki Activity

Canonize meeting names to a properly formatted date

Browse Source
This commit is contained in:
I-Al-Istannen 2020-12-30 14:32:59 +01:00
parent 1c2b6bf994
commit 9b048a9cfc
1 changed files with 21 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
PFERD/ilias/crawler.py
View File

@ -40,6 +40,7 @@ class IliasElementType(Enum):
REGULAR_FILE = "REGULAR_FILE" REGULAR_FILE = "REGULAR_FILE"
VIDEO_FILE = "VIDEO_FILE" VIDEO_FILE = "VIDEO_FILE"
FORUM = "FORUM" FORUM = "FORUM"
MEETING = "MEETING"
EXTERNAL_LINK = "EXTERNAL_LINK" EXTERNAL_LINK = "EXTERNAL_LINK"
def is_folder(self) -> bool: def is_folder(self) -> bool:
@ -241,6 +242,8 @@ class IliasCrawler:
entries_to_process += self._crawl_video_directory(entry.path, url) entries_to_process += self._crawl_video_directory(entry.path, url)
continue continue
PRETTY.warning(f"Unknown type: {entry.entry_type}!")
return result return result
def _crawl_folder(self, folder_path: Path, url: str) -> List[IliasCrawlerEntry]: def _crawl_folder(self, folder_path: Path, url: str) -> List[IliasCrawlerEntry]:
@ -269,6 +272,19 @@ class IliasCrawler:
if element_type == IliasElementType.REGULAR_FILE: if element_type == IliasElementType.REGULAR_FILE:
result += self._crawl_file(folder_path, link, abs_url) result += self._crawl_file(folder_path, link, abs_url)
elif element_type == IliasElementType.MEETING:
meeting_name = str(element_path.name)
date_portion_str = meeting_name.split(" - ")[0]
date_portion = demangle_date(date_portion_str)
if not date_portion:
result += [IliasCrawlerEntry(element_path, abs_url, element_type, None)]
continue
rest_of_name = meeting_name.removeprefix(date_portion_str)
new_name = datetime.datetime.strftime(date_portion, "%Y-%m-%d, %H:%M") + rest_of_name
new_path = Path(folder_path, _sanitize_path_name(new_name))
result += [IliasCrawlerEntry(new_path, abs_url, IliasElementType.REGULAR_FOLDER, None)]
elif element_type is not None: elif element_type is not None:
result += [IliasCrawlerEntry(element_path, abs_url, element_type, None)] result += [IliasCrawlerEntry(element_path, abs_url, element_type, None)]
else: else:
@ -320,6 +336,8 @@ class IliasCrawler:
""" """
# pylint: disable=too-many-return-statements # pylint: disable=too-many-return-statements
found_parent: Optional[bs4.Tag] = None
# We look for the outer div of our inner link, to find information around it # We look for the outer div of our inner link, to find information around it
# (mostly the icon) # (mostly the icon)
for parent in link_element.parents: for parent in link_element.parents:
@ -350,6 +368,9 @@ class IliasCrawler:
if str(img_tag["src"]).endswith("frm.svg"): if str(img_tag["src"]).endswith("frm.svg"):
return IliasElementType.FORUM return IliasElementType.FORUM
if str(img_tag["src"]).endswith("sess.svg"):
return IliasElementType.MEETING
return IliasElementType.REGULAR_FOLDER return IliasElementType.REGULAR_FOLDER
@staticmethod @staticmethod