Bump version to 3.4.2

This downloads all forum posts when needed and saves each thread in its
own html file, named after the thread title.

.github/workflows/build-and-release.yml

@@ -14,7 +14,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, windows-latest, macos-latest]
-        python: ["3.8"]
+        python: ["3.9"]
     steps:
 
       - uses: actions/checkout@v2

CHANGELOG.md

@@ -22,6 +22,53 @@ ambiguous situations.
 
 ## Unreleased
 
+## 3.4.2 - 2022-10-26
+
+### Added
+- Recognize and crawl content pages in cards
+- Recognize and ignore surveys
+
+### Fixed
+- Forum crawling crashing when a thread has no messages at all
+- Forum crawling crashing when a forum has no threads at all
+- Ilias login failing in some cases
+- Crawling of paginated future meetings
+- IPD crawler handling of URLs without trailing slash
+
+## 3.4.1 - 2022-08-17
+
+### Added
+- Download of page descriptions
+- Forum download support
+- `pass` authenticator
+
+### Changed
+- Add `cpp` extension to default `link_regex` of IPD crawler
+- Mention hrefs in IPD crawler's `--explain` output for users of `link_regex` option
+- Simplify default IPD crawler `link_regex`
+
+### Fixed
+- IPD crawler crashes on some sites
+- Meeting name normalization for yesterday, today and tomorrow
+- Crawling of meeting file previews
+- Login with new login button html layout
+- Descriptions for courses are now placed in the correct subfolder when
+  downloading the whole desktop
+
+## 3.4.0 - 2022-05-01
+
+### Added
+- Message when Shibboleth entitlements need to be manually reviewed
+- Links to unofficial packages and repology in the readme
+
+### Changed
+- Increase minimum supported Python version to 3.9
+- Support video listings with more columns
+- Use UTF-8 when reading/writing the config file
+
+### Fixed
+- Crash during authentication when the Shibboleth session is still valid
+
 ## 3.3.1 - 2022-01-15
 
 ### Fixed

CONFIG.md

@@ -146,7 +146,7 @@ requests is likely a good idea.
 - `target`: URL to a KIT-IPD page
 - `link_regex`: A regex that is matched against the `href` part of links. If it
   matches, the given link is downloaded as a file. This is used to extract
-  files from KIT-IPD pages. (Default: `^.*/[^/]*\.(?:pdf|zip|c|java)$`)
+  files from KIT-IPD pages. (Default: `^.*?[^/]+\.(pdf|zip|c|cpp|java)$`)
 
 ### The `kit-ilias-web` crawler
 
@@ -223,6 +223,23 @@ is stored in the keyring.
 - `keyring_name`: The service name PFERD uses for storing credentials. (Default:
   `PFERD`)
 
+### The `pass` authenticator
+
+This authenticator queries the [`pass` password manager][3] for a username and
+password. It tries to be mostly compatible with [browserpass][4] and
+[passff][5], so see those links for an overview of the format. If PFERD fails
+to load your password, you can use the `--explain` flag to see why.
+
+- `passname`: The name of the password to use (Required)
+- `username_prefixes`: A comma-separated list of username line prefixes
+  (Default: `login,username,user`)
+- `password_prefixes`: A comma-separated list of password line prefixes
+  (Default: `password,pass,secret`)
+
+[3]: <https://www.passwordstore.org/> "Pass: The Standard Unix Password Manager"
+[4]: <https://github.com/browserpass/browserpass-extension#organizing-password-store> "Organizing password store"
+[5]: <https://github.com/passff/passff#multi-line-format> "Multi-line format"
+
 ### The `tfa` authenticator
 
 This authenticator prompts the user on the console for a two-factor
@@ -316,7 +333,7 @@ is a regular expression and `TARGET` an f-string based template. If a path
 matches `SOURCE`, the output path is created using `TARGET` as template.
 `SOURCE` is automatically anchored.
 
-`TARGET` uses Python's [format string syntax][3]. The *n*-th capturing group can
+`TARGET` uses Python's [format string syntax][6]. The *n*-th capturing group can
 be referred to as `{g<n>}` (e.g. `{g3}`). `{g0}` refers to the original path.
 If capturing group *n*'s contents are a valid integer, the integer value is
 available as `{i<n>}` (e.g. `{i3}`). If capturing group *n*'s contents are a
@@ -337,7 +354,7 @@ Example: `f(oo+)/be?ar -re-> B{g1.upper()}H/fear`
 - Converts `fooooo/bear` into `BOOOOOH/fear`
 - Converts `foo/bar/baz` into `BOOH/fear/baz`
 
-[3]: <https://docs.python.org/3/library/string.html#format-string-syntax> "Format String Syntax"
+[6]: <https://docs.python.org/3/library/string.html#format-string-syntax> "Format String Syntax"
 
 ### The `-name-re->` arrow
 

PFERD/__main__.py

@@ -159,3 +159,7 @@ def main() -> None:
         sys.exit(1)
     else:
         pferd.print_report()
+
+
+if __name__ == "__main__":
+    main()

PFERD/auth/__init__.py

@@ -5,6 +5,7 @@ from ..config import Config
 from .authenticator import Authenticator, AuthError, AuthLoadError, AuthSection  # noqa: F401
 from .credential_file import CredentialFileAuthenticator, CredentialFileAuthSection
 from .keyring import KeyringAuthenticator, KeyringAuthSection
+from .pass_ import PassAuthenticator, PassAuthSection
 from .simple import SimpleAuthenticator, SimpleAuthSection
 from .tfa import TfaAuthenticator
 
@@ -19,6 +20,8 @@ AUTHENTICATORS: Dict[str, AuthConstructor] = {
         CredentialFileAuthenticator(n, CredentialFileAuthSection(s), c),
     "keyring": lambda n, s, c:
         KeyringAuthenticator(n, KeyringAuthSection(s)),
+    "pass": lambda n, s, c:
+        PassAuthenticator(n, PassAuthSection(s)),
     "simple": lambda n, s, c:
         SimpleAuthenticator(n, SimpleAuthSection(s)),
     "tfa": lambda n, s, c:

PFERD/auth/credential_file.py

@@ -20,8 +20,10 @@ class CredentialFileAuthenticator(Authenticator):
 
         path = config.default_section.working_dir() / section.path()
         try:
-            with open(path) as f:
+            with open(path, encoding="utf-8") as f:
                 lines = list(f)
+        except UnicodeDecodeError:
+            raise AuthLoadError(f"Credential file at {fmt_real_path(path)} is not encoded using UTF-8")
         except OSError as e:
             raise AuthLoadError(f"No credential file at {fmt_real_path(path)}") from e
 

PFERD/auth/pass_.py

@@ -0,0 +1,98 @@
+import re
+import subprocess
+from typing import List, Tuple
+
+from ..logging import log
+from .authenticator import Authenticator, AuthError, AuthSection
+
+
+class PassAuthSection(AuthSection):
+    def passname(self) -> str:
+        if (value := self.s.get("passname")) is None:
+            self.missing_value("passname")
+        return value
+
+    def username_prefixes(self) -> List[str]:
+        value = self.s.get("username_prefixes", "login,username,user")
+        return [prefix.lower() for prefix in value.split(",")]
+
+    def password_prefixes(self) -> List[str]:
+        value = self.s.get("password_prefixes", "password,pass,secret")
+        return [prefix.lower() for prefix in value.split(",")]
+
+
+class PassAuthenticator(Authenticator):
+    PREFIXED_LINE_RE = r"([a-zA-Z]+):\s?(.*)"  # to be used with fullmatch
+
+    def __init__(self, name: str, section: PassAuthSection) -> None:
+        super().__init__(name)
+
+        self._passname = section.passname()
+        self._username_prefixes = section.username_prefixes()
+        self._password_prefixes = section.password_prefixes()
+
+    async def credentials(self) -> Tuple[str, str]:
+        log.explain_topic("Obtaining credentials from pass")
+
+        try:
+            log.explain(f"Calling 'pass show {self._passname}'")
+            result = subprocess.check_output(["pass", "show", self._passname], text=True)
+        except subprocess.CalledProcessError as e:
+            raise AuthError(f"Failed to get password info from {self._passname}: {e}")
+
+        prefixed = {}
+        unprefixed = []
+        for line in result.strip().splitlines():
+            if match := re.fullmatch(self.PREFIXED_LINE_RE, line):
+                prefix = match.group(1).lower()
+                value = match.group(2)
+                log.explain(f"Found prefixed line {line!r} with prefix {prefix!r}, value {value!r}")
+                if prefix in prefixed:
+                    raise AuthError(f"Prefix {prefix} specified multiple times")
+                prefixed[prefix] = value
+            else:
+                log.explain(f"Found unprefixed line {line!r}")
+                unprefixed.append(line)
+
+        username = None
+        for prefix in self._username_prefixes:
+            log.explain(f"Looking for username at prefix {prefix!r}")
+            if prefix in prefixed:
+                username = prefixed[prefix]
+                log.explain(f"Found username {username!r}")
+                break
+
+        password = None
+        for prefix in self._password_prefixes:
+            log.explain(f"Looking for password at prefix {prefix!r}")
+            if prefix in prefixed:
+                password = prefixed[prefix]
+                log.explain(f"Found password {password!r}")
+                break
+
+        if password is None and username is None:
+            log.explain("No username and password found so far")
+            log.explain("Using first unprefixed line as password")
+            log.explain("Using second unprefixed line as username")
+        elif password is None:
+            log.explain("No password found so far")
+            log.explain("Using first unprefixed line as password")
+        elif username is None:
+            log.explain("No username found so far")
+            log.explain("Using first unprefixed line as username")
+
+        if password is None:
+            if not unprefixed:
+                log.explain("Not enough unprefixed lines left")
+                raise AuthError("Password could not be determined")
+            password = unprefixed.pop(0)
+            log.explain(f"Found password {password!r}")
+
+        if username is None:
+            if not unprefixed:
+                log.explain("Not enough unprefixed lines left")
+                raise AuthError("Username could not be determined")
+            username = unprefixed.pop(0)
+            log.explain(f"Found username {username!r}")
+
+        return username, password

PFERD/cli/command_kit_ilias_web.py

@@ -62,6 +62,11 @@ GROUP.add_argument(
     action=BooleanOptionalAction,
     help="crawl and download videos"
 )
+GROUP.add_argument(
+    "--forums",
+    action=BooleanOptionalAction,
+    help="crawl and download forum posts"
+)
 GROUP.add_argument(
     "--http-timeout", "-t",
     type=float,
@@ -90,6 +95,8 @@ def load(
         section["link_redirect_delay"] = str(args.link_redirect_delay)
     if args.videos is not None:
         section["videos"] = "yes" if args.videos else "no"
+    if args.forums is not None:
+        section["forums"] = "yes" if args.forums else "no"
     if args.http_timeout is not None:
         section["http_timeout"] = str(args.http_timeout)
 

PFERD/config.py

@@ -120,7 +120,7 @@ class Config:
         # Using config.read_file instead of config.read because config.read
         # would just ignore a missing file and carry on.
         try:
-            with open(path) as f:
+            with open(path, encoding="utf-8") as f:
                 parser.read_file(f, source=str(path))
         except FileNotFoundError:
             raise ConfigLoadError(path, "File does not exist")
@@ -128,6 +128,8 @@ class Config:
             raise ConfigLoadError(path, "That's a directory, not a file")
         except PermissionError:
             raise ConfigLoadError(path, "Insufficient permissions")
+        except UnicodeDecodeError:
+            raise ConfigLoadError(path, "File is not encoded using UTF-8")
 
     def dump(self, path: Optional[Path] = None) -> None:
         """
@@ -154,12 +156,12 @@ class Config:
             try:
                 # x = open for exclusive creation, failing if the file already
                 # exists
-                with open(path, "x") as f:
+                with open(path, "x", encoding="utf-8") as f:
                     self._parser.write(f)
             except FileExistsError:
                 print("That file already exists.")
                 if asyncio.run(prompt_yes_no("Overwrite it?", default=False)):
-                    with open(path, "w") as f:
+                    with open(path, "w", encoding="utf-8") as f:
                         self._parser.write(f)
                 else:
                     raise ConfigDumpError(path, "File already exists")

PFERD/crawl/crawler.py

@@ -1,9 +1,10 @@
 import asyncio
 import os
 from abc import ABC, abstractmethod
+from collections.abc import Awaitable, Coroutine
 from datetime import datetime
 from pathlib import Path, PurePath
-from typing import Any, Awaitable, Callable, Dict, List, Optional, Sequence, Set, Tuple, TypeVar
+from typing import Any, Callable, Dict, List, Optional, Sequence, Set, Tuple, TypeVar
 
 from ..auth import Authenticator
 from ..config import Config, Section
@@ -58,7 +59,7 @@ def noncritical(f: Wrapped) -> Wrapped:
     return wrapper  # type: ignore
 
 
-AWrapped = TypeVar("AWrapped", bound=Callable[..., Awaitable[Optional[Any]]])
+AWrapped = TypeVar("AWrapped", bound=Callable[..., Coroutine[Any, Any, Optional[Any]]])
 
 
 def anoncritical(f: AWrapped) -> AWrapped:

PFERD/crawl/http_crawler.py

@@ -108,7 +108,7 @@ class HttpCrawler(Crawler):
 
     def _load_cookies_from_file(self, path: Path) -> None:
         jar: Any = http.cookies.SimpleCookie()
-        with open(path) as f:
+        with open(path, encoding="utf-8") as f:
             for i, line in enumerate(f):
                 # Names of headers are case insensitive
                 if line[:11].lower() == "set-cookie:":
@@ -121,7 +121,7 @@ class HttpCrawler(Crawler):
         jar: Any = http.cookies.SimpleCookie()
         for morsel in self._cookie_jar:
             jar[morsel.key] = morsel
-        with open(path, "w") as f:
+        with open(path, "w", encoding="utf-8") as f:
             f.write(jar.output(sep="\n"))
             f.write("\n")  # A trailing newline is just common courtesy
 

PFERD/crawl/ilias/ilias_html_cleaner.py

@@ -0,0 +1,91 @@
+from bs4 import BeautifulSoup, Comment, Tag
+
+_STYLE_TAG_CONTENT = """
+    .ilc_text_block_Information {
+      background-color: #f5f7fa;
+    }
+    div.ilc_text_block_Standard {
+      margin-bottom: 10px;
+      margin-top: 10px;
+    }
+    span.ilc_text_inline_Strong {
+      font-weight: bold;
+    }
+
+    .accordion-head {
+      background-color: #f5f7fa;
+      padding: 0.5rem 0;
+    }
+
+    h3 {
+      margin-top: 0.5rem;
+      margin-bottom: 1rem;
+    }
+
+    br.visible-break {
+      margin-bottom: 1rem;
+    }
+
+    article {
+      margin: 0.5rem 0;
+    }
+
+    body {
+      padding: 1em;
+      grid-template-columns: 1fr min(60rem, 90%) 1fr;
+      line-height: 1.2;
+    }
+"""
+
+_ARTICLE_WORTHY_CLASSES = [
+    "ilc_text_block_Information",
+    "ilc_section_Attention",
+    "ilc_section_Link",
+]
+
+
+def insert_base_markup(soup: BeautifulSoup) -> BeautifulSoup:
+    head = soup.new_tag("head")
+    soup.insert(0, head)
+
+    simplecss_link: Tag = soup.new_tag("link")
+    # <link rel="stylesheet" href="https://cdn.simplecss.org/simple.css">
+    simplecss_link["rel"] = "stylesheet"
+    simplecss_link["href"] = "https://cdn.simplecss.org/simple.css"
+    head.append(simplecss_link)
+
+    # Basic style tags for compat
+    style: Tag = soup.new_tag("style")
+    style.append(_STYLE_TAG_CONTENT)
+    head.append(style)
+
+    return soup
+
+
+def clean(soup: BeautifulSoup) -> BeautifulSoup:
+    for block in soup.find_all(class_=lambda x: x in _ARTICLE_WORTHY_CLASSES):
+        block.name = "article"
+
+    for block in soup.find_all("h3"):
+        block.name = "div"
+
+    for block in soup.find_all("h1"):
+        block.name = "h3"
+
+    for block in soup.find_all(class_="ilc_va_ihcap_VAccordIHeadCap"):
+        block.name = "h3"
+        block["class"] += ["accordion-head"]
+
+    for dummy in soup.select(".ilc_text_block_Standard.ilc_Paragraph"):
+        children = list(dummy.children)
+        if not children:
+            dummy.decompose()
+        if len(children) > 1:
+            continue
+        if type(children[0]) == Comment:
+            dummy.decompose()
+
+    for hrule_imposter in soup.find_all(class_="ilc_section_Separator"):
+        hrule_imposter.insert(0, soup.new_tag("hr"))
+
+    return soup

PFERD/crawl/ilias/kit_ilias_html.py

@@ -3,7 +3,7 @@ import re
 from dataclasses import dataclass
 from datetime import date, datetime, timedelta
 from enum import Enum
-from typing import List, Optional, Union
+from typing import Dict, List, Optional, Union
 from urllib.parse import urljoin, urlparse
 
 from bs4 import BeautifulSoup, Tag
@@ -24,6 +24,7 @@ class IliasElementType(Enum):
     LINK = "link"
     BOOKING = "booking"
     MEETING = "meeting"
+    SURVEY = "survey"
     VIDEO = "video"
     VIDEO_PLAYER = "video_player"
     VIDEO_FOLDER = "video_folder"
@@ -55,6 +56,21 @@ class IliasPageElement:
         return self.url
 
 
+@dataclass
+class IliasDownloadForumData:
+    url: str
+    form_data: Dict[str, Union[str, List[str]]]
+    empty: bool
+
+
+@dataclass
+class IliasForumThread:
+    title: str
+    title_tag: Tag
+    content_tag: Tag
+    mtime: Optional[datetime]
+
+
 class IliasPage:
 
     def __init__(self, soup: BeautifulSoup, _page_url: str, source_element: Optional[IliasPageElement]):
@@ -85,13 +101,72 @@ class IliasPage:
         log.explain("Page is a normal folder, searching for elements")
         return self._find_normal_entries()
 
+    def get_description(self) -> Optional[BeautifulSoup]:
+        def is_interesting_class(name: str) -> bool:
+            return name in ["ilCOPageSection", "ilc_Paragraph", "ilc_va_ihcap_VAccordIHeadCap"]
+
+        paragraphs: List[Tag] = self._soup.findAll(class_=is_interesting_class)
+        if not paragraphs:
+            return None
+
+        # Extract bits and pieces into a string and parse it again.
+        # This ensures we don't miss anything and weird structures are resolved
+        # somewhat gracefully.
+        raw_html = ""
+        for p in paragraphs:
+            if p.find_parent(class_=is_interesting_class):
+                continue
+
+            # Ignore special listings (like folder groupings)
+            if "ilc_section_Special" in p["class"]:
+                continue
+
+            raw_html += str(p) + "\n"
+        raw_html = f"<body>\n{raw_html}\n</body>"
+
+        return BeautifulSoup(raw_html, "html.parser")
+
+    def get_download_forum_data(self) -> Optional[IliasDownloadForumData]:
+        form = self._soup.find("form", attrs={"action": lambda x: x and "fallbackCmd=showThreads" in x})
+        if not form:
+            return None
+        post_url = self._abs_url_from_relative(form["action"])
+
+        thread_ids = [f["value"] for f in form.find_all(attrs={"name": "thread_ids[]"})]
+
+        form_data: Dict[str, Union[str, List[ſtr]]] = {
+            "thread_ids[]": thread_ids,
+            "selected_cmd2": "html",
+            "select_cmd2": "Ausführen",
+            "selected_cmd": "",
+        }
+
+        return IliasDownloadForumData(url=post_url, form_data=form_data, empty=len(thread_ids) == 0)
+
     def get_next_stage_element(self) -> Optional[IliasPageElement]:
+        if self._is_forum_page():
+            if "trows=800" in self._page_url:
+                return None
+            log.explain("Requesting *all* forum threads")
+            return self._get_show_max_forum_entries_per_page_url()
         if self._is_ilias_opencast_embedding():
+            log.explain("Unwrapping opencast embedding")
             return self.get_child_elements()[0]
         if self._page_type == IliasElementType.VIDEO_FOLDER_MAYBE_PAGINATED:
+            log.explain("Unwrapping video pagination")
             return self._find_video_entries_paginated()[0]
+        if self._contains_collapsed_future_meetings():
+            log.explain("Requesting *all* future meetings")
+            return self._uncollapse_future_meetings_url()
         return None
 
+    def _is_forum_page(self) -> bool:
+        read_more_btn = self._soup.find(
+            "button",
+            attrs={"onclick": lambda x: x and "cmdClass=ilobjforumgui&cmd=markAllRead" in x}
+        )
+        return read_more_btn is not None
+
     def _is_video_player(self) -> bool:
         return "paella_config_file" in str(self._soup)
 
@@ -135,6 +210,16 @@ class IliasPage:
             return False
         return "target=copa_" in link.get("value")
 
+    def _contains_collapsed_future_meetings(self) -> bool:
+        return self._uncollapse_future_meetings_url() is not None
+
+    def _uncollapse_future_meetings_url(self) -> Optional[IliasPageElement]:
+        element = self._soup.find("a", attrs={"href": lambda x: x and "crs_next_sess=1" in x})
+        if not element:
+            return None
+        link = self._abs_url_from_link(element)
+        return IliasPageElement(IliasElementType.FOLDER, link, "show all meetings")
+
     def _player_to_video(self) -> List[IliasPageElement]:
         # 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
@@ -169,6 +254,19 @@ class IliasPage:
 
         return items
 
+    def _get_show_max_forum_entries_per_page_url(self) -> Optional[IliasPageElement]:
+        correct_link = self._soup.find(
+            "a",
+            attrs={"href": lambda x: x and "trows=800" in x and "cmd=showThreads" in x}
+        )
+
+        if not correct_link:
+            return None
+
+        link = self._abs_url_from_link(correct_link)
+
+        return IliasPageElement(IliasElementType.FORUM, link, "show all forum threads")
+
     def _find_personal_desktop_entries(self) -> List[IliasPageElement]:
         items: List[IliasPageElement] = []
 
@@ -280,11 +378,22 @@ class IliasPage:
 
     def _listed_video_to_element(self, link: Tag) -> IliasPageElement:
         # The link is part of a table with multiple columns, describing metadata.
-        # 6th child (1 indexed) is the modification time string
+        # 6th or 7th child (1 indexed) is the modification time string. Try to find it
+        # by parsing backwards from the end and finding something that looks like a date
+        modification_time = None
+        row: Tag = link.parent.parent.parent
+        column_count = len(row.select("td.std"))
+        for index in range(column_count, 0, -1):
             modification_string = link.parent.parent.parent.select_one(
-            "td.std:nth-child(6)"
+                f"td.std:nth-child({index})"
             ).getText().strip()
+            if re.search(r"\d+\.\d+.\d+ - \d+:\d+", modification_string):
                 modification_time = datetime.strptime(modification_string, "%d.%m.%Y - %H:%M")
+                break
+
+        if modification_time is None:
+            log.warn(f"Could not determine upload time for {link}")
+            modification_time = datetime.now()
 
         title = link.parent.parent.parent.select_one("td.std:nth-child(3)").getText().strip()
         title += ".mp4"
@@ -417,6 +526,12 @@ class IliasPage:
             element_type = self._find_type_from_link(element_name, link, abs_url)
             description = self._find_link_description(link)
 
+            # The last meeting on every page is expanded by default.
+            # Its content is then shown inline *and* in the meeting page itself.
+            # We should skip the inline content.
+            if element_type != IliasElementType.MEETING and self._is_in_expanded_meeting(link):
+                continue
+
             if not element_type:
                 continue
             if element_type == IliasElementType.MEETING:
@@ -434,6 +549,26 @@ class IliasPage:
 
         return result
 
+    def _is_in_expanded_meeting(self, tag: Tag) -> bool:
+        """
+        Returns whether a file is part of an expanded meeting.
+        Has false positives for meetings themselves as their title is also "in the expanded meeting content".
+        It is in the same general div and this whole thing is guesswork.
+        Therefore, you should check for meetings before passing them in this function.
+        """
+        parents: List[Tag] = list(tag.parents)
+        for parent in parents:
+            if not parent.get("class"):
+                continue
+
+            # We should not crawl files under meetings
+            if "ilContainerListItemContentCB" in parent.get("class"):
+                link: Tag = parent.parent.find("a")
+                type = IliasPage._find_type_from_folder_like(link, self._page_url)
+                return type == IliasElementType.MEETING
+
+        return False
+
     def _find_upwards_folder_hierarchy(self, tag: Tag) -> List[str]:
         """
         Interprets accordions and expandable blocks as virtual folders and returns them
@@ -596,6 +731,10 @@ class IliasPage:
             return IliasElementType.TEST
         if "fold" in icon["class"]:
             return IliasElementType.FOLDER
+        if "copa" in icon["class"]:
+            return IliasElementType.FOLDER
+        if "svy" in icon["class"]:
+            return IliasElementType.SURVEY
 
         _unexpected_html_warning()
         log.warn_contd(f"Could not extract type from {icon} for card title {card_title}")
@@ -675,6 +814,10 @@ class IliasPage:
         if img_tag is None:
             img_tag = found_parent.select_one("img.icon")
 
+        if img_tag is None and found_parent.find("a", attrs={"href": lambda x: x and "crs_next_sess=" in x}):
+            log.explain("Found session expansion button, skipping it as it has no content")
+            return None
+
         if img_tag is None:
             _unexpected_html_warning()
             log.warn_contd(f"Tried to figure out element type, but did not find an image for {url}")
@@ -709,17 +852,26 @@ class IliasPage:
         Normalizes meeting names, which have a relative time as their first part,
         to their date in ISO format.
         """
-        date_portion_str = meeting_name.split(" - ")[0]
+
+        # This checks whether we can reach a `:` without passing a `-`
+        if re.search(r"^[^-]+: ", meeting_name):
+            # Meeting name only contains date: "05. Jan 2000:"
+            split_delimiter = ":"
+        else:
+            # Meeting name contains date and start/end times: "05. Jan 2000, 16:00 - 17:30:"
+            split_delimiter = ", "
+
+        # We have a meeting day without time
+        date_portion_str = meeting_name.split(split_delimiter)[0]
         date_portion = demangle_date(date_portion_str)
 
+        # We failed to parse the date, bail out
         if not date_portion:
             return meeting_name
 
-        rest_of_name = meeting_name
+        # Replace the first section with the absolute date
-        if rest_of_name.startswith(date_portion_str):
+        rest_of_name = split_delimiter.join(meeting_name.split(split_delimiter)[1:])
-            rest_of_name = rest_of_name[len(date_portion_str):]
+        return datetime.strftime(date_portion, "%Y-%m-%d") + split_delimiter + rest_of_name
-
-        return datetime.strftime(date_portion, "%Y-%m-%d, %H:%M") + rest_of_name
 
     def _abs_url_from_link(self, link_tag: Tag) -> str:
         """
@@ -744,35 +896,47 @@ english_months = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep',
 
 def demangle_date(date_str: str, fail_silently: bool = False) -> Optional[datetime]:
     """
-    Demangle a given date in one of the following formats:
+    Demangle a given date in one of the following formats (hour/minute part is optional):
     "Gestern, HH:MM"
     "Heute, HH:MM"
     "Morgen, HH:MM"
     "dd. mon yyyy, HH:MM
     """
     try:
+        # Normalize whitespace because users
         date_str = re.sub(r"\s+", " ", date_str)
+
         date_str = re.sub("Gestern|Yesterday", _format_date_english(_yesterday()), date_str, re.I)
         date_str = re.sub("Heute|Today", _format_date_english(date.today()), date_str, re.I)
         date_str = re.sub("Morgen|Tomorrow",  _format_date_english(_tomorrow()), date_str, re.I)
+        date_str = date_str.strip()
         for german, english in zip(german_months, english_months):
             date_str = date_str.replace(german, english)
             # Remove trailing dots for abbreviations, e.g. "20. Apr. 2020" -> "20. Apr 2020"
             date_str = date_str.replace(english + ".", english)
 
-        # We now have a nice english String in the format: "dd. mmm yyyy, hh:mm"
+        # We now have a nice english String in the format: "dd. mmm yyyy, hh:mm" or "dd. mmm yyyy"
+
+        # Check if we have a time as well
+        if ", " in date_str:
             day_part, time_part = date_str.split(",")
+        else:
+            day_part = date_str.split(",")[0]
+            time_part = None
+
         day_str, month_str, year_str = day_part.split(" ")
 
         day = int(day_str.strip().replace(".", ""))
         month = english_months.index(month_str.strip()) + 1
         year = int(year_str.strip())
 
+        if time_part:
             hour_str, minute_str = time_part.split(":")
             hour = int(hour_str)
             minute = int(minute_str)
-
             return datetime(year, month, day, hour, minute)
+
+        return datetime(year, month, day)
     except Exception:
         if not fail_silently:
             log.warn(f"Date parsing failed for {date_str!r}")
@@ -794,3 +958,45 @@ def _tomorrow() -> date:
 
 def _sanitize_path_name(name: str) -> str:
     return name.replace("/", "-").replace("\\", "-").strip()
+
+
+def parse_ilias_forum_export(forum_export: BeautifulSoup) -> List[IliasForumThread]:
+    elements = []
+    for p in forum_export.select("body > p"):
+        title_tag = p
+        content_tag = p.find_next_sibling("ul")
+
+        if not content_tag:
+            # ILIAS allows users to delete the initial post while keeping the thread open
+            # This produces empty threads without *any* content.
+            # I am not sure why you would want this, but ILIAS makes it easy to do.
+            continue
+
+        title = p.find("b").text
+        if ":" in title:
+            title = title[title.find(":") + 1:]
+        title = title.strip()
+        mtime = _guess_timestamp_from_forum_post_content(content_tag)
+        elements.append(IliasForumThread(title, title_tag, content_tag, mtime))
+
+    return elements
+
+
+def _guess_timestamp_from_forum_post_content(content: Tag) -> Optional[datetime]:
+    posts: Optional[Tag] = content.select(".ilFrmPostHeader > span.small")
+    if not posts:
+        return None
+
+    newest_date: Optional[datetime] = None
+
+    for post in posts:
+        text = post.text.strip()
+        text = text[text.rfind("|") + 1:]
+        date = demangle_date(text, fail_silently=True)
+        if not date:
+            continue
+
+        if not newest_date or newest_date < date:
+            newest_date = date
+
+    return newest_date

PFERD/crawl/ilias/kit_ilias_web_crawler.py

@@ -1,7 +1,8 @@
 import asyncio
 import re
+from collections.abc import Awaitable, Coroutine
 from pathlib import PurePath
-from typing import Any, Awaitable, Callable, Dict, List, Optional, Set, TypeVar, Union, cast
+from typing import Any, Callable, Dict, List, Optional, Set, Union, cast
 
 import aiohttp
 import yarl
@@ -13,13 +14,21 @@ from ...config import Config
 from ...logging import ProgressBar, log
 from ...output_dir import FileSink, Redownload
 from ...utils import fmt_path, soupify, url_set_query_param
-from ..crawler import CrawlError, CrawlToken, CrawlWarning, DownloadToken, anoncritical
+from ..crawler import AWrapped, CrawlError, CrawlToken, CrawlWarning, DownloadToken, anoncritical
 from ..http_crawler import HttpCrawler, HttpCrawlerSection
 from .file_templates import Links
-from .kit_ilias_html import IliasElementType, IliasPage, IliasPageElement
+from .ilias_html_cleaner import clean, insert_base_markup
+from .kit_ilias_html import (IliasElementType, IliasForumThread, IliasPage, IliasPageElement,
+                             _sanitize_path_name, parse_ilias_forum_export)
 
 TargetType = Union[str, int]
 
+_ILIAS_URL = "https://ilias.studium.kit.edu"
+
+
+class KitShibbolethBackgroundLoginSuccessful():
+    pass
+
 
 class KitIliasWebCrawlerSection(HttpCrawlerSection):
     def target(self) -> TargetType:
@@ -33,7 +42,7 @@ class KitIliasWebCrawlerSection(HttpCrawlerSection):
         if target == "desktop":
             # Full personal desktop
             return target
-        if target.startswith("https://ilias.studium.kit.edu"):
+        if target.startswith(_ILIAS_URL):
             # ILIAS URL
             return target
 
@@ -65,6 +74,9 @@ class KitIliasWebCrawlerSection(HttpCrawlerSection):
     def videos(self) -> bool:
         return self.s.getboolean("videos", fallback=False)
 
+    def forums(self) -> bool:
+        return self.s.getboolean("forums", fallback=False)
+
 
 _DIRECTORY_PAGES: Set[IliasElementType] = set([
     IliasElementType.EXERCISE,
@@ -82,8 +94,6 @@ _VIDEO_ELEMENTS: Set[IliasElementType] = set([
     IliasElementType.VIDEO_FOLDER_MAYBE_PAGINATED,
 ])
 
-AWrapped = TypeVar("AWrapped", bound=Callable[..., Awaitable[Optional[Any]]])
-
 
 def _iorepeat(attempts: int, name: str, failure_is_error: bool = False) -> Callable[[AWrapped], AWrapped]:
     def decorator(f: AWrapped) -> AWrapped:
@@ -177,12 +187,13 @@ instance's greatest bottleneck.
             section.tfa_auth(authenticators),
         )
 
-        self._base_url = "https://ilias.studium.kit.edu"
+        self._base_url = _ILIAS_URL
 
         self._target = section.target()
         self._link_file_redirect_delay = section.link_redirect_delay()
         self._links = section.links()
         self._videos = section.videos()
+        self._forums = section.forums()
         self._visited_urls: Set[str] = set()
 
     async def _run(self) -> None:
@@ -216,27 +227,44 @@ instance's greatest bottleneck.
         cl = maybe_cl  # Not mypy's fault, but explained here: https://github.com/python/mypy/issues/2608
 
         elements: List[IliasPageElement] = []
+        # A list as variable redefinitions are not propagated to outer scopes
+        description: List[BeautifulSoup] = []
 
         @_iorepeat(3, "crawling url")
         async def gather_elements() -> None:
             elements.clear()
             async with cl:
-                soup = await self._get_page(url)
+                next_stage_url: Optional[str] = url
+                current_parent = None
 
-                if expected_id is not None:
+                # Duplicated code, but the root page is special - we want to avoid fetching it twice!
+                while next_stage_url:
+                    soup = await self._get_page(next_stage_url)
+
+                    if current_parent is None and expected_id is not None:
                         perma_link_element: Tag = soup.find(id="current_perma_link")
                         if not perma_link_element or "crs_" not in perma_link_element.get("value"):
                             raise CrawlError("Invalid course id? Didn't find anything looking like a course")
 
-                # Duplicated code, but the root page is special - we want to avoid fetching it twice!
+                    log.explain_topic(f"Parsing HTML page for {fmt_path(cl.path)}")
-                log.explain_topic("Parsing root HTML page")
+                    log.explain(f"URL: {next_stage_url}")
-                log.explain(f"URL: {url}")
+                    page = IliasPage(soup, next_stage_url, current_parent)
-                page = IliasPage(soup, url, None)
+                    if next_element := page.get_next_stage_element():
+                        current_parent = next_element
+                        next_stage_url = next_element.url
+                    else:
+                        next_stage_url = None
+
                 elements.extend(page.get_child_elements())
+                if description_string := page.get_description():
+                    description.append(description_string)
 
         # Fill up our task list with the found elements
         await gather_elements()
 
+        if description:
+            await self._download_description(PurePath("."), description[0])
+
         elements.sort(key=lambda e: e.id())
 
         tasks: List[Awaitable[None]] = []
@@ -252,7 +280,7 @@ instance's greatest bottleneck.
         url: str,
         parent: IliasPageElement,
         path: PurePath,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         maybe_cl = await self.crawl(path)
         if not maybe_cl:
             return None
@@ -266,6 +294,8 @@ instance's greatest bottleneck.
         cl: CrawlToken,
     ) -> None:
         elements: List[IliasPageElement] = []
+        # A list as variable redefinitions are not propagated to outer scopes
+        description: List[BeautifulSoup] = []
 
         @_iorepeat(3, "crawling folder")
         async def gather_elements() -> None:
@@ -286,10 +316,15 @@ instance's greatest bottleneck.
                         next_stage_url = None
 
                 elements.extend(page.get_child_elements())
+                if description_string := page.get_description():
+                    description.append(description_string)
 
         # Fill up our task list with the found elements
         await gather_elements()
 
+        if description:
+            await self._download_description(cl.path, description[0])
+
         elements.sort(key=lambda e: e.id())
 
         tasks: List[Awaitable[None]] = []
@@ -310,7 +345,7 @@ instance's greatest bottleneck.
         self,
         parent_path: PurePath,
         element: IliasPageElement,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         if element.url in self._visited_urls:
             raise CrawlWarning(
                 f"Found second path to element {element.name!r} at {element.url!r}. Aborting subpath"
@@ -320,26 +355,42 @@ instance's greatest bottleneck.
         element_path = PurePath(parent_path, element.name)
 
         if element.type in _VIDEO_ELEMENTS:
-            log.explain_topic(f"Decision: Crawl video element {fmt_path(element_path)}")
             if not self._videos:
-                log.explain("Video crawling is disabled")
+                log.status(
-                log.explain("Answer: no")
+                    "[bold bright_black]",
+                    "Ignored",
+                    fmt_path(element_path),
+                    "[bright_black](enable with option 'videos')"
+                )
                 return None
-            else:
-                log.explain("Video crawling is enabled")
-                log.explain("Answer: yes")
 
         if element.type == IliasElementType.FILE:
             return await self._handle_file(element, element_path)
         elif element.type == IliasElementType.FORUM:
-            log.explain_topic(f"Decision: Crawl {fmt_path(element_path)}")
+            if not self._forums:
-            log.explain("Forums are not supported")
+                log.status(
-            log.explain("Answer: No")
+                    "[bold bright_black]",
+                    "Ignored",
+                    fmt_path(element_path),
+                    "[bright_black](enable with option 'forums')"
+                )
                 return None
+            return await self._handle_forum(element, element_path)
         elif element.type == IliasElementType.TEST:
-            log.explain_topic(f"Decision: Crawl {fmt_path(element_path)}")
+            log.status(
-            log.explain("Tests contain no relevant files")
+                "[bold bright_black]",
-            log.explain("Answer: No")
+                "Ignored",
+                fmt_path(element_path),
+                "[bright_black](tests contain no relevant data)"
+            )
+            return None
+        elif element.type == IliasElementType.SURVEY:
+            log.status(
+                "[bold bright_black]",
+                "Ignored",
+                fmt_path(element_path),
+                "[bright_black](surveys contain no relevant data)"
+            )
             return None
         elif element.type == IliasElementType.LINK:
             return await self._handle_link(element, element_path)
@@ -360,7 +411,7 @@ instance's greatest bottleneck.
         self,
         element: IliasPageElement,
         element_path: PurePath,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         log.explain_topic(f"Decision: Crawl Link {fmt_path(element_path)}")
         log.explain(f"Links type is {self._links}")
 
@@ -407,7 +458,7 @@ instance's greatest bottleneck.
         self,
         element: IliasPageElement,
         element_path: PurePath,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         log.explain_topic(f"Decision: Crawl Booking Link {fmt_path(element_path)}")
         log.explain(f"Links type is {self._links}")
 
@@ -426,6 +477,19 @@ instance's greatest bottleneck.
 
         return self._download_booking(element, link_template_maybe, maybe_dl)
 
+    @anoncritical
+    @_iorepeat(1, "downloading description")
+    async def _download_description(self, parent_path: PurePath, description: BeautifulSoup) -> None:
+        path = parent_path / "Description.html"
+        dl = await self.download(path, redownload=Redownload.ALWAYS)
+        if not dl:
+            return
+
+        async with dl as (bar, sink):
+            description = clean(insert_base_markup(description))
+            sink.file.write(description.prettify().encode("utf-8"))
+            sink.done()
+
     @anoncritical
     @_iorepeat(3, "resolving booking")
     async def _download_booking(
@@ -443,7 +507,7 @@ instance's greatest bottleneck.
             if hdrs.LOCATION not in resp.headers:
                 return soupify(await resp.read()).select_one("a").get("href").strip()
 
-        self._authenticate()
+        await self._authenticate()
 
         async with self.session.get(export_url, allow_redirects=False) as resp:
             # No redirect means we were authenticated
@@ -456,7 +520,7 @@ instance's greatest bottleneck.
         self,
         element: IliasPageElement,
         element_path: PurePath,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         # Copy old mapping as it is likely still relevant
         if self.prev_report:
             self.report.add_custom_value(
@@ -564,7 +628,7 @@ instance's greatest bottleneck.
         self,
         element: IliasPageElement,
         element_path: PurePath,
-    ) -> Optional[Awaitable[None]]:
+    ) -> Optional[Coroutine[Any, Any, None]]:
         maybe_dl = await self.download(element_path, mtime=element.mtime)
         if not maybe_dl:
             return None
@@ -607,6 +671,72 @@ instance's greatest bottleneck.
         if not await try_stream():
             raise CrawlError("File streaming failed after authenticate()")
 
+    async def _handle_forum(
+        self,
+        element: IliasPageElement,
+        element_path: PurePath,
+    ) -> Optional[Coroutine[Any, Any, None]]:
+        maybe_cl = await self.crawl(element_path)
+        if not maybe_cl:
+            return None
+        return self._crawl_forum(element, maybe_cl)
+
+    @_iorepeat(3, "crawling forum")
+    @anoncritical
+    async def _crawl_forum(self, element: IliasPageElement, cl: CrawlToken) -> None:
+        elements: List[IliasForumThread] = []
+
+        async with cl:
+            next_stage_url = element.url
+            while next_stage_url:
+                log.explain_topic(f"Parsing HTML page for {fmt_path(cl.path)}")
+                log.explain(f"URL: {next_stage_url}")
+
+                soup = await self._get_page(next_stage_url)
+                page = IliasPage(soup, next_stage_url, None)
+
+                if next := page.get_next_stage_element():
+                    next_stage_url = next.url
+                else:
+                    break
+
+            download_data = page.get_download_forum_data()
+            if not download_data:
+                raise CrawlWarning("Failed to extract forum data")
+            if download_data.empty:
+                log.explain("Forum had no threads")
+                elements = []
+                return
+            html = await self._post_authenticated(download_data.url, download_data.form_data)
+            elements = parse_ilias_forum_export(soupify(html))
+
+        elements.sort(key=lambda elem: elem.title)
+
+        tasks: List[Awaitable[None]] = []
+        for elem in elements:
+            tasks.append(asyncio.create_task(self._download_forum_thread(cl.path, elem)))
+
+        # And execute them
+        await self.gather(tasks)
+
+    @anoncritical
+    @_iorepeat(3, "saving forum thread")
+    async def _download_forum_thread(
+        self,
+        parent_path: PurePath,
+        element: IliasForumThread,
+    ) -> None:
+        path = parent_path / (_sanitize_path_name(element.title) + ".html")
+        maybe_dl = await self.download(path, mtime=element.mtime)
+        if not maybe_dl:
+            return
+
+        async with maybe_dl as (bar, sink):
+            content = element.title_tag.prettify()
+            content += element.content_tag.prettify()
+            sink.file.write(content.encode("utf-8"))
+            sink.done()
+
     async def _get_page(self, url: str) -> BeautifulSoup:
         auth_id = await self._current_auth_id()
         async with self.session.get(url) as request:
@@ -624,6 +754,30 @@ instance's greatest bottleneck.
                 return soup
         raise CrawlError("get_page failed even after authenticating")
 
+    async def _post_authenticated(
+        self,
+        url: str,
+        data: dict[str, Union[str, List[str]]]
+    ) -> BeautifulSoup:
+        auth_id = await self._current_auth_id()
+
+        form_data = aiohttp.FormData()
+        for key, val in data.items():
+            form_data.add_field(key, val)
+
+        async with self.session.post(url, data=form_data(), allow_redirects=False) as request:
+            if request.status == 200:
+                return await request.read()
+
+        # We weren't authenticated, so try to do that
+        await self.authenticate(auth_id)
+
+        # Retry once after authenticating. If this fails, we will die.
+        async with self.session.post(url, data=data, allow_redirects=False) as request:
+            if request.status == 200:
+                return await request.read()
+        raise CrawlError("post_authenticated failed even after authenticating")
+
     # We repeat this as the login method in shibboleth doesn't handle I/O errors.
     # Shibboleth is quite reliable as well, the repeat is likely not critical here.
     @ _iorepeat(3, "Login", failure_is_error=True)
@@ -635,7 +789,7 @@ instance's greatest bottleneck.
         # Normal ILIAS pages
         mainbar: Optional[Tag] = soup.find(class_="il-maincontrols-metabar")
         if mainbar is not None:
-            login_button = mainbar.find("button", attrs={"data-action": lambda x: x and "login.php" in x})
+            login_button = mainbar.find(attrs={"href": lambda x: x and "login.php" in x})
             shib_login = soup.find(id="button_shib_login")
             return not login_button and not shib_login
 
@@ -680,14 +834,17 @@ class KitShibbolethLogin:
 
         # Equivalent: Click on "Mit KIT-Account anmelden" button in
         # https://ilias.studium.kit.edu/login.php
-        url = "https://ilias.studium.kit.edu/shib_login.php"
+        url = f"{_ILIAS_URL}/shib_login.php"
         data = {
             "sendLogin": "1",
             "idp_selection": "https://idp.scc.kit.edu/idp/shibboleth",
             "il_target": "",
             "home_organization_selection": "Weiter",
         }
-        soup: BeautifulSoup = await _shib_post(sess, url, data)
+        soup: Union[BeautifulSoup, KitShibbolethBackgroundLoginSuccessful] = await _shib_post(sess, url, data)
+
+        if isinstance(soup, KitShibbolethBackgroundLoginSuccessful):
+            return
 
         # Attempt to login using credentials, if necessary
         while not self._login_successful(soup):
@@ -710,6 +867,12 @@ class KitShibbolethLogin:
             }
             soup = await _post(sess, url, data)
 
+            if soup.find(id="attributeRelease"):
+                raise CrawlError(
+                    "ILIAS Shibboleth entitlements changed! "
+                    "Please log in once in your browser and review them"
+                )
+
             if self._tfa_required(soup):
                 soup = await self._authenticate_tfa(sess, soup)
 
@@ -720,7 +883,7 @@ class KitShibbolethLogin:
         # (or clicking "Continue" if you have JS disabled)
         relay_state = soup.find("input", {"name": "RelayState"})
         saml_response = soup.find("input", {"name": "SAMLResponse"})
-        url = "https://ilias.studium.kit.edu/Shibboleth.sso/SAML2/POST"
+        url = f"{_ILIAS_URL}/Shibboleth.sso/SAML2/POST"
         data = {  # using the info obtained in the while loop above
             "RelayState": relay_state["value"],
             "SAMLResponse": saml_response["value"],
@@ -769,24 +932,41 @@ async def _post(session: aiohttp.ClientSession, url: str, data: Any) -> Beautifu
         return soupify(await response.read())
 
 
-async def _shib_post(session: aiohttp.ClientSession, url: str, data: Any) -> BeautifulSoup:
+async def _shib_post(
+    session: aiohttp.ClientSession,
+    url: str,
+    data: Any
+) -> Union[BeautifulSoup, KitShibbolethBackgroundLoginSuccessful]:
     """
     aiohttp unescapes '/' and ':' in URL query parameters which is not RFC compliant and rejected
     by Shibboleth. Thanks a lot. So now we unroll the requests manually, parse location headers and
     build encoded URL objects ourselves... Who thought mangling location header was a good idea??
     """
+    log.explain_topic("Shib login POST")
     async with session.post(url, data=data, allow_redirects=False) as response:
         location = response.headers.get("location")
+        log.explain(f"Got location {location!r}")
         if not location:
-            raise CrawlWarning(f"Login failed, no location header present at {url}")
+            raise CrawlWarning(f"Login failed (1), no location header present at {url}")
         correct_url = yarl.URL(location, encoded=True)
+        log.explain(f"Corrected location to {correct_url!r}")
+
+        if str(correct_url).startswith(_ILIAS_URL):
+            log.explain("ILIAS recognized our shib token and logged us in in the background, returning")
+            return KitShibbolethBackgroundLoginSuccessful()
 
         async with session.get(correct_url, allow_redirects=False) as response:
-            as_yarl = yarl.URL(response.url)
             location = response.headers.get("location")
+            log.explain(f"Redirected to {location!r} with status {response.status}")
+            # If shib still still has a valid session, it will directly respond to the request
+            if location is None:
+                log.explain("Shib recognized us, returning its response directly")
+                return soupify(await response.read())
 
+            as_yarl = yarl.URL(response.url)
+            # Probably not needed anymore, but might catch a few weird situations with a nicer message
             if not location or not as_yarl.host:
-                raise CrawlWarning(f"Login failed, no location header present at {correct_url}")
+                raise CrawlWarning(f"Login failed (2), no location header present at {correct_url}")
 
             correct_url = yarl.URL.build(
                 scheme=as_yarl.scheme,
@@ -794,6 +974,7 @@ async def _shib_post(session: aiohttp.ClientSession, url: str, data: Any) -> Bea
                 path=location,
                 encoded=True
             )
+            log.explain(f"Corrected location to {correct_url!r}")
 
             async with session.get(correct_url, allow_redirects=False) as response:
                 return soupify(await response.read())

PFERD/crawl/kit_ipd_crawler.py

@@ -24,10 +24,13 @@ class KitIpdCrawlerSection(HttpCrawlerSection):
         if not target.startswith("https://"):
             self.invalid_value("target", target, "Should be a URL")
 
+        if not target.endswith("/"):
+            target = target + "/"
+
         return target
 
     def link_regex(self) -> Pattern[str]:
-        regex = self.s.get("link_regex", r"^.*/[^/]*\.(?:pdf|zip|c|java)$")
+        regex = self.s.get("link_regex", r"^.*?[^/]+\.(pdf|zip|c|cpp|java)$")
         return re.compile(regex)
 
 
@@ -45,7 +48,7 @@ class KitIpdFolder:
     def explain(self) -> None:
         log.explain_topic(f"Folder {self.name!r}")
         for file in self.files:
-            log.explain(f"File {file.name!r}")
+            log.explain(f"File {file.name!r} (href={file.url!r})")
 
     def __hash__(self) -> int:
         return self.name.__hash__()
@@ -113,7 +116,7 @@ class KitIpdCrawler(HttpCrawler):
             else:
                 file = self._extract_file(element)
                 items.add(file)
-                log.explain_topic(f"Orphan file {file.name!r}")
+                log.explain_topic(f"Orphan file {file.name!r} (href={file.url!r})")
                 log.explain("Attributing it to root folder")
 
         return items
@@ -161,4 +164,10 @@ class KitIpdCrawler(HttpCrawler):
 
     async def get_page(self) -> BeautifulSoup:
         async with self.session.get(self._url) as request:
-            return soupify(await request.read())
+            # The web page for Algorithmen für Routenplanung contains some
+            # weird comments that beautifulsoup doesn't parse correctly. This
+            # hack enables those pages to be crawled, and should hopefully not
+            # cause issues on other pages.
+            content = (await request.read()).decode("utf-8")
+            content = re.sub(r"<!--.*?-->", "", content)
+            return soupify(content.encode("utf-8"))

PFERD/logging.py

@@ -68,7 +68,7 @@ class Log:
         if self._download_progress.task_ids:
             elements.append(self._download_progress)
 
-        group = Group(*elements)  # type: ignore
+        group = Group(*elements)
         self._live.update(group)
 
     @contextmanager
@@ -197,7 +197,7 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
         if self.output_explain:
             self.print(f"  {escape(text)}")
 
-    def status(self, style: str, action: str, text: str) -> None:
+    def status(self, style: str, action: str, text: str, suffix: str = "") -> None:
         """
         Print a status update while crawling. Allows markup in the "style"
         argument which will be applied to the "action" string.
@@ -205,7 +205,7 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
 
         if self.output_status:
             action = escape(f"{action:<{self.STATUS_WIDTH}}")
-            self.print(f"{style}{action}[/] {escape(text)}")
+            self.print(f"{style}{action}[/] {escape(text)} {suffix}")
 
     def report(self, text: str) -> None:
         """

PFERD/output_dir.py

@@ -503,7 +503,7 @@ class OutputDirectory:
         try:
             self._prev_report = Report.load(self._report_path)
             log.explain("Loaded report successfully")
-        except (OSError, json.JSONDecodeError, ReportLoadError) as e:
+        except (OSError, UnicodeDecodeError, json.JSONDecodeError, ReportLoadError) as e:
             log.explain("Failed to load report")
             log.explain(str(e))
 

PFERD/report.py

@@ -100,10 +100,10 @@ class Report:
     @classmethod
     def load(cls, path: Path) -> "Report":
         """
-        May raise OSError, JsonDecodeError, ReportLoadError.
+        May raise OSError, UnicodeDecodeError, JsonDecodeError, ReportLoadError.
         """
 
-        with open(path) as f:
+        with open(path, encoding="utf-8") as f:
             data = json.load(f)
 
         if not isinstance(data, dict):
@@ -148,7 +148,7 @@ class Report:
             "encountered_errors": self.encountered_errors,
         }
 
-        with open(path, "w") as f:
+        with open(path, "w", encoding="utf-8") as f:
             json.dump(data, f, indent=2, sort_keys=True)
             f.write("\n")  # json.dump doesn't do this
 

PFERD/version.py

@@ -1,2 +1,2 @@
 NAME = "PFERD"
-VERSION = "3.3.1"
+VERSION = "3.4.2"

README.md

@@ -17,7 +17,7 @@ Binaries for Linux, Windows and Mac can be downloaded directly from the
 
 ### With pip
 
-Ensure you have at least Python 3.8 installed. Run the following command to
+Ensure you have at least Python 3.9 installed. Run the following command to
 install PFERD or upgrade it to the latest version:
 
 ```
@@ -26,6 +26,14 @@ $ pip install --upgrade git+https://github.com/Garmelon/PFERD@latest
 
 The use of [venv](https://docs.python.org/3/library/venv.html) is recommended.
 
+### With package managers
+
+Unofficial packages are available for:
+- [AUR](https://aur.archlinux.org/packages/pferd)
+- [nixpkgs](https://github.com/NixOS/nixpkgs/blob/master/pkgs/tools/misc/pferd/default.nix)
+
+See also PFERD's [repology page](https://repology.org/project/pferd/versions).
+
 ## Basic usage
 
 PFERD can be run directly from the command line with no config file. Run `pferd

setup.cfg

@@ -4,7 +4,7 @@ version = attr: PFERD.version.VERSION
 
 [options]
 packages = find:
-python_requires = >=3.8
+python_requires = >=3.9
 install_requires =
   aiohttp>=3.8.1
   beautifulsoup4>=4.10.0