Bump version to 3.8.3

We do not want a second path and the instruction field has an identical
link...

.github/dependabot.yml

@@ -0,0 +1,10 @@
+version: 2
+updates:
+  - package-ecosystem: github-actions
+    directory: /
+    schedule:
+      interval: monthly
+    groups:
+      gh-actions:
+        patterns:
+          - "*"

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

@@ -1,6 +1,6 @@
 name: build-and-release
 
-on: push
+on: [push, pull_request]
 
 defaults:
   run:
@@ -13,13 +13,12 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [ubuntu-latest, windows-latest, macos-latest]
+        os: [ubuntu-latest, windows-latest, macos-13, macos-latest]
-        python: ["3.9"]
+        python: ["3.11"]
     steps:
+      - uses: actions/checkout@v4
 
-      - uses: actions/checkout@v2
+      - uses: actions/setup-python@v5
-
-      - uses: actions/setup-python@v2
         with:
           python-version: ${{ matrix.python }}
 
@@ -34,7 +33,12 @@ jobs:
         run: ./scripts/setup --no-pip
 
       - name: Run checks
-        run: ./scripts/check
+        run: |
+          ./scripts/check
+          ./scripts/format
+
+      - name: Assert no changes
+        run: git diff --exit-code
 
       - name: Build
         run: ./scripts/build
@@ -45,9 +49,9 @@ jobs:
         run: mv dist/pferd* dist/pferd-${{ matrix.os }}
 
       - name: Upload binary
-        uses: actions/upload-artifact@v2
+        uses: actions/upload-artifact@v4
         with:
-          name: Binaries
+          name: pferd-${{ matrix.os }}
           path: dist/pferd-${{ matrix.os }}
 
   release:
@@ -57,18 +61,20 @@ jobs:
     steps:
 
       - name: Download binaries
-        uses: actions/download-artifact@v2
+        uses: actions/download-artifact@v4
         with:
-          name: Binaries
+          pattern: pferd-*
+          merge-multiple: true
 
       - name: Rename binaries
         run: |
           mv pferd-ubuntu-latest pferd-linux
           mv pferd-windows-latest pferd-windows.exe
+          mv pferd-macos-13 pferd-mac-x86_64
           mv pferd-macos-latest pferd-mac
 
       - name: Create release
-        uses: softprops/action-gh-release@v1
+        uses: softprops/action-gh-release@v2
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         with:
@@ -76,3 +82,4 @@ jobs:
             pferd-linux
             pferd-windows.exe
             pferd-mac
+            pferd-mac-x86_64

.gitignore

@@ -3,6 +3,7 @@
 /PFERD.egg-info/
 __pycache__/
 /.vscode/
+/.idea/
 
 # pyinstaller
 /pferd.spec

CHANGELOG.md

@@ -22,6 +22,148 @@ ambiguous situations.
 
 ## Unreleased
 
+## 3.8.3 - 2025-07-01
+
+## Added
+- Support for link collections.  
+  In "fancy" mode, a single HTML file with multiple links is generated.
+  In all other modes, PFERD creates a folder for the collection and a new file
+  for every link inside.
+
+## Fixed
+- Crawling of exercises with instructions
+- Don't download unavailable elements.  
+  Elements that are unavailable (for example, because their availability is
+  time restricted) will not download the HTML for the info page anymore.
+- `base_url` argument for `ilias-web` crawler causing crashes
+
+## 3.8.2 - 2025-04-29
+
+## Changed
+- Explicitly mention that wikis are not supported at the moment and ignore them
+
+## Fixed
+- Ilias-native login
+- Exercise crawling
+
+## 3.8.1 - 2025-04-17
+
+## Fixed
+- Description html files now specify at UTF-8 encoding
+- Images in descriptions now always have a white background
+
+## 3.8.0 - 2025-04-16
+
+### Added
+- Support for ILIAS 9
+
+### Changed
+- Added prettier CSS to forum threads
+- Downloaded forum threads now link to the forum instead of the ILIAS thread
+- Increase minimum supported Python version to 3.11
+- Do not crawl nested courses (courses linked in other courses)
+
+## Fixed
+- File links in report on Windows
+- TOTP authentication in KIT Shibboleth
+- Forum crawling only considering the first 20 entries
+
+## 3.7.0 - 2024-11-13
+
+### Added
+- Support for MOB videos in page descriptions
+- Clickable links in the report to directly open new/modified/not-deleted files
+- Support for non KIT shibboleth login
+
+### Changed
+- Remove videos from description pages
+- Perform ILIAS cycle detection after processing the transform to allow
+  ignoring duplicated elements
+- Parse headings (h1-h3) as folders in kit-ipd crawler
+
+### Fixed
+- Personal desktop/dashboard/favorites crawling
+- Crawling of nested courses
+- Downloading of links with no target URL
+- Handle row flex on description pages
+- Add `<!DOCTYPE html>` heading to forum threads to fix mime type detection
+- Handle groups in cards
+
+## 3.6.0 - 2024-10-23
+
+### Added
+- Generic `ilias-web` crawler and `ilias-web` CLI command
+- Support for the course overview page. Using this URL as a target might cause
+  duplication warnings, as subgroups are listed separately.
+- Support for named capture groups in regex transforms
+- Crawl custom item groups as folders
+
+### Fixed
+- Normalization of meeting names in cards
+- Sanitization of slashes in exercise container names
+
+## 3.5.2 - 2024-04-14
+
+### Fixed
+- Crawling of personal desktop with ILIAS 8
+- Crawling of empty personal desktops
+
+## 3.5.1 - 2024-04-09
+
+### Added
+- Support for ILIAS 8
+
+### Fixed
+- Video name deduplication
+
+## 3.5.0 - 2023-09-13
+
+### Added
+- `no-delete-prompt-override` conflict resolution strategy
+- Support for ILIAS learning modules
+- `show_not_deleted` option to stop printing the "Not Deleted" status or report
+  message. This combines nicely with the `no-delete-prompt-override` strategy,
+  causing PFERD to mostly ignore local-only files.
+- Support for mediacast video listings
+- Crawling of files in info tab
+
+### Changed
+- Remove size suffix for files in content pages
+
+### Fixed
+- Crawling of courses with the timeline view as the default tab
+- Crawling of file and custom opencast cards
+- Crawling of button cards without descriptions
+- Abort crawling when encountering an unexpected ilias root page redirect
+- Sanitize ascii control characters on Windows
+- Crawling of paginated past meetings
+- Ignore SCORM learning modules
+
+## 3.4.3 - 2022-11-29
+
+### Added
+- Missing documentation for `forums` option
+
+### Changed
+- Clear up error message shown when multiple paths are found to an element
+
+### Fixed
+- IPD crawler unnecessarily appending trailing slashes
+- Crawling opencast when ILIAS is set to English
+
+## 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

CONFIG.md

@@ -4,11 +4,11 @@ A config file consists of sections. A section begins with a `[section]` header,
 which is followed by a list of `key = value` pairs. Comments must be on their
 own line and start with `#`. Multiline values must be indented beyond their key.
 Boolean values can be `yes` or `no`. For more details and some examples on the
-format, see the [configparser documentation][1] ([interpolation][2] is
+format, see the [configparser documentation][cp-file]
-disabled).
+([interpolation][cp-interp] is disabled).
 
-[1]: <https://docs.python.org/3/library/configparser.html#supported-ini-file-structure> "Supported INI File Structure"
+[cp-file]: <https://docs.python.org/3/library/configparser.html#supported-ini-file-structure> "Supported INI File Structure"
-[2]: <https://docs.python.org/3/library/configparser.html#interpolation-of-values> "Interpolation of values"
+[cp-interp]: <https://docs.python.org/3/library/configparser.html#interpolation-of-values> "Interpolation of values"
 
 ## The `DEFAULT` section
 
@@ -26,6 +26,9 @@ default values for the other sections.
   `Added ...`) while running a crawler. (Default: `yes`)
 - `report`: Whether PFERD should print a report of added, changed and deleted
    local files for all crawlers before exiting. (Default: `yes`)
+- `show_not_deleted`: Whether PFERD should print messages in status and report
+   when a local-only file wasn't deleted. Combines nicely with the
+   `no-delete-prompt-override` conflict resolution strategy.
 - `share_cookies`: Whether crawlers should share cookies where applicable. For
   example, some crawlers share cookies if they crawl the same website using the
   same account. (Default: `yes`)
@@ -75,6 +78,9 @@ common to all crawlers:
       using `prompt` and always choosing "yes".
     - `no-delete`: Never delete local files, but overwrite local files if the
       remote file is different.
+    - `no-delete-prompt-overwrite`: Never delete local files, but prompt to
+      overwrite local files if the remote file is different. Combines nicely
+      with the `show_not_deleted` option.
 - `transform`: Rules for renaming and excluding certain files and directories.
   For more details, see [this section](#transformation-rules). (Default: empty)
 - `tasks`: The maximum number of concurrent tasks (such as crawling or
@@ -140,7 +146,7 @@ crawler simulate a slower, network-based crawler.
 
 This crawler crawls a KIT-IPD page by url. The root page can be crawled from
 outside the KIT network so you will be informed about any new/deleted files,
-but downloading files requires you to be within. Adding a show delay between
+but downloading files requires you to be within. Adding a short delay between
 requests is likely a good idea.
 
 - `target`: URL to a KIT-IPD page
@@ -148,6 +154,64 @@ requests is likely a good idea.
   matches, the given link is downloaded as a file. This is used to extract
   files from KIT-IPD pages. (Default: `^.*?[^/]+\.(pdf|zip|c|cpp|java)$`)
 
+### The `ilias-web` crawler
+
+This crawler crawls a generic ILIAS instance.
+
+Inspired by [this ILIAS downloader][ilias-dl], the following configurations should work
+out of the box for the corresponding universities:
+
+[ilias-dl]: https://github.com/V3lop5/ilias-downloader/blob/main/configs "ilias-downloader configs"
+
+| University      | `base_url`                              | `login_type` | `client_id`   |
+|-----------------|-----------------------------------------|--------------|---------------|
+| FH Aachen       | https://www.ili.fh-aachen.de            | local        | elearning     |
+| Uni Köln        | https://www.ilias.uni-koeln.de/ilias    | local        | uk            |
+| Uni Konstanz    | https://ilias.uni-konstanz.de           | local        | ILIASKONSTANZ |
+| Uni Stuttgart   | https://ilias3.uni-stuttgart.de         | local        | Uni_Stuttgart |
+| Uni Tübingen    | https://ovidius.uni-tuebingen.de/ilias3 | shibboleth   |               |
+| KIT ILIAS Pilot | https://pilot.ilias.studium.kit.edu     | shibboleth   | pilot         |
+
+If your university isn't listed, try navigating to your instance's login page.
+Assuming no custom login service is used, the URL will look something like this:
+
+```jinja
+{{ base_url }}/login.php?client_id={{ client_id }}&cmd=force_login&lang=
+```
+
+If the values work, feel free to submit a PR and add them to the table above.
+
+- `base_url`: The URL where the ILIAS instance is located. (Required)
+- `login_type`: How you authenticate. (Required)
+    - `local`: Use `client_id` for authentication.
+    - `shibboleth`: Use shibboleth for authentication.
+- `client_id`: An ID used for authentication if `login_type` is `local`. Is
+  ignored if `login_type` is `shibboleth`.
+- `target`: The ILIAS element to crawl. (Required)
+    - `desktop`: Crawl your personal desktop / dashboard
+    - `<course id>`: Crawl the course with the given id
+    - `<url>`: Crawl a given element by URL (preferably the permanent URL linked
+      at the bottom of its ILIAS page).  
+      This also supports the "My Courses" overview page to download *all*
+      courses. Note that this might produce confusing local directory layouts
+      and duplication warnings if you are a member of an ILIAS group. The
+      `desktop` target is generally preferable.
+- `auth`: Name of auth section to use for login. (Required)
+- `tfa_auth`: Name of auth section to use for two-factor authentication. Only
+  uses the auth section's password. (Default: Anonymous `tfa` authenticator)
+- `links`: How to represent external links. (Default: `fancy`)
+    - `ignore`: Don't download links.
+    - `plaintext`: A text file containing only the URL.
+    - `fancy`: A HTML file looking like the ILIAS link element.
+    - `internet-shortcut`: An internet shortcut file (`.url` file).
+- `link_redirect_delay`: Time (in seconds) until `fancy` link files will
+  redirect to the actual URL. Set to a negative value to disable the automatic
+  redirect. (Default: `-1`)
+- `videos`: Whether to download videos. (Default: `no`)
+- `forums`: Whether to download forum threads. (Default: `no`)
+- `http_timeout`: The timeout (in seconds) for all HTTP requests. (Default:
+  `20.0`)
+
 ### The `kit-ilias-web` crawler
 
 This crawler crawls the KIT ILIAS instance.
@@ -181,6 +245,7 @@ script once per day should be fine.
   redirect to the actual URL. Set to a negative value to disable the automatic
   redirect. (Default: `-1`)
 - `videos`: Whether to download videos. (Default: `no`)
+- `forums`: Whether to download forum threads. (Default: `no`)
 - `http_timeout`: The timeout (in seconds) for all HTTP requests. (Default:
   `20.0`)
 
@@ -225,10 +290,10 @@ is stored in the keyring.
 
 ### The `pass` authenticator
 
-This authenticator queries the [`pass` password manager][3] for a username and
+This authenticator queries the [`pass` password manager][pass] for a username
-password. It tries to be mostly compatible with [browserpass][4] and
+and password. It tries to be mostly compatible with [browserpass][browserpass]
-[passff][5], so see those links for an overview of the format. If PFERD fails
+and [passff][passff], so see those links for an overview of the format. If PFERD
-to load your password, you can use the `--explain` flag to see why.
+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
@@ -236,9 +301,9 @@ to load your password, you can use the `--explain` flag to see why.
 - `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"
+[pass]: <https://www.passwordstore.org/> "Pass: The Standard Unix Password Manager"
-[4]: <https://github.com/browserpass/browserpass-extension#organizing-password-store> "Organizing password store"
+[browserpass]: <https://github.com/browserpass/browserpass-extension#organizing-password-store> "Organizing password store"
-[5]: <https://github.com/passff/passff#multi-line-format> "Multi-line format"
+[passff]: <https://github.com/passff/passff#multi-line-format> "Multi-line format"
 
 ### The `tfa` authenticator
 
@@ -289,7 +354,7 @@ path matches `SOURCE`, it is renamed to `TARGET`.
 Example: `foo/bar --> baz`
 - Doesn't match `foo`, `a/foo/bar` or `foo/baz`
 - Converts `foo/bar` into `baz`
-- Converts `foo/bar/wargl` into `bar/wargl`
+- Converts `foo/bar/wargl` into `baz/wargl`
 
 Example: `foo/bar --> !`
 - Doesn't match `foo`, `a/foo/bar` or `foo/baz`
@@ -337,7 +402,8 @@ matches `SOURCE`, the output path is created using `TARGET` as template.
 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
-valid float, the float value is available as `{f<n>}` (e.g. `{f3}`). If a
+valid float, the float value is available as `{f<n>}` (e.g. `{f3}`). Named capture
+groups (e.g. `(?P<name>)`) are available by their name (e.g. `{name}`). If a
 capturing group is not present (e.g. when matching the string `cd` with the
 regex `(ab)?cd`), the corresponding variables are not defined.
 

LICENSE

@@ -1,5 +1,6 @@
-Copyright 2019-2021 Garmelon, I-Al-Istannen, danstooamerican, pavelzw,
+Copyright 2019-2024 Garmelon, I-Al-Istannen, danstooamerican, pavelzw,
-                    TheChristophe, Scriptim, thelukasprobst, Toorero
+                    TheChristophe, Scriptim, thelukasprobst, Toorero,
+                    Mr-Pine, p-fruck, PinieP
 
 Permission is hereby granted, free of charge, to any person obtaining a copy of
 this software and associated documentation files (the "Software"), to deal in

PFERD/__main__.py

@@ -47,6 +47,8 @@ def configure_logging_from_args(args: argparse.Namespace) -> None:
         log.output_explain = args.explain
     if args.status is not None:
         log.output_status = args.status
+    if args.show_not_deleted is not None:
+        log.output_not_deleted = args.show_not_deleted
     if args.report is not None:
         log.output_report = args.report
 
@@ -72,6 +74,8 @@ def configure_logging_from_config(args: argparse.Namespace, config: Config) -> N
             log.output_status = config.default_section.status()
         if args.report is None:
             log.output_report = config.default_section.report()
+        if args.show_not_deleted is None:
+            log.output_not_deleted = config.default_section.show_not_deleted()
     except ConfigOptionError as e:
         log.error(str(e))
         sys.exit(1)

PFERD/auth/keyring.py

@@ -1,4 +1,4 @@
-from typing import Optional, Tuple
+from typing import Optional, Tuple, cast
 
 import keyring
 
@@ -13,7 +13,7 @@ class KeyringAuthSection(AuthSection):
         return self.s.get("username")
 
     def keyring_name(self) -> str:
-        return self.s.get("keyring_name", fallback=NAME)
+        return cast(str, self.s.get("keyring_name", fallback=NAME))
 
 
 class KeyringAuthenticator(Authenticator):

PFERD/cli/__init__.py

@@ -8,6 +8,7 @@
 # well.
 
 from . import command_local  # noqa: F401 imported but unused
+from . import command_ilias_web  # noqa: F401 imported but unused
 from . import command_kit_ilias_web  # noqa: F401 imported but unused
 from . import command_kit_ipd  # noqa: F401 imported but unused
 from .parser import PARSER, ParserLoadError, load_default_section  # noqa: F401 imported but unused

PFERD/cli/command_ilias_web.py

@@ -0,0 +1,56 @@
+import argparse
+import configparser
+
+from ..logging import log
+from .common_ilias_args import configure_common_group_args, load_common
+from .parser import CRAWLER_PARSER, SUBPARSERS, load_crawler
+
+COMMAND_NAME = "ilias-web"
+
+SUBPARSER = SUBPARSERS.add_parser(
+    COMMAND_NAME,
+    parents=[CRAWLER_PARSER],
+)
+
+GROUP = SUBPARSER.add_argument_group(
+    title=f"{COMMAND_NAME} crawler arguments",
+    description=f"arguments for the '{COMMAND_NAME}' crawler",
+)
+
+GROUP.add_argument(
+    "--base-url",
+    type=str,
+    metavar="BASE_URL",
+    help="The base url of the ilias instance"
+)
+
+GROUP.add_argument(
+    "--client-id",
+    type=str,
+    metavar="CLIENT_ID",
+    help="The client id of the ilias instance"
+)
+
+configure_common_group_args(GROUP)
+
+
+def load(
+        args: argparse.Namespace,
+        parser: configparser.ConfigParser,
+) -> None:
+    log.explain(f"Creating config for command '{COMMAND_NAME}'")
+
+    parser["crawl:ilias"] = {}
+    section = parser["crawl:ilias"]
+    load_crawler(args, section)
+
+    section["type"] = COMMAND_NAME
+    if args.base_url is not None:
+        section["base_url"] = args.base_url
+    if args.client_id is not None:
+        section["client_id"] = args.client_id
+
+    load_common(section, args, parser)
+
+
+SUBPARSER.set_defaults(command=load)

PFERD/cli/command_kit_ilias_web.py

@@ -1,120 +1,37 @@
 import argparse
 import configparser
-from pathlib import Path
 
-from ..crawl.ilias.file_templates import Links
 from ..logging import log
-from .parser import (CRAWLER_PARSER, SUBPARSERS, BooleanOptionalAction, ParserLoadError, load_crawler,
+from .common_ilias_args import configure_common_group_args, load_common
-                     show_value_error)
+from .parser import CRAWLER_PARSER, SUBPARSERS, load_crawler
+
+COMMAND_NAME = "kit-ilias-web"
 
 SUBPARSER = SUBPARSERS.add_parser(
-    "kit-ilias-web",
+    COMMAND_NAME,
     parents=[CRAWLER_PARSER],
 )
 
 GROUP = SUBPARSER.add_argument_group(
-    title="kit-ilias-web crawler arguments",
+    title=f"{COMMAND_NAME} crawler arguments",
-    description="arguments for the 'kit-ilias-web' crawler",
+    description=f"arguments for the '{COMMAND_NAME}' crawler",
-)
-GROUP.add_argument(
-    "target",
-    type=str,
-    metavar="TARGET",
-    help="course id, 'desktop', or ILIAS URL to crawl"
-)
-GROUP.add_argument(
-    "output",
-    type=Path,
-    metavar="OUTPUT",
-    help="output directory"
-)
-GROUP.add_argument(
-    "--username", "-u",
-    type=str,
-    metavar="USERNAME",
-    help="user name for authentication"
-)
-GROUP.add_argument(
-    "--keyring",
-    action=BooleanOptionalAction,
-    help="use the system keyring to store and retrieve passwords"
-)
-GROUP.add_argument(
-    "--credential-file",
-    type=Path,
-    metavar="PATH",
-    help="read username and password from a credential file"
-)
-GROUP.add_argument(
-    "--links",
-    type=show_value_error(Links.from_string),
-    metavar="OPTION",
-    help="how to represent external links"
-)
-GROUP.add_argument(
-    "--link-redirect-delay",
-    type=int,
-    metavar="SECONDS",
-    help="time before 'fancy' links redirect to to their target (-1 to disable)"
-)
-GROUP.add_argument(
-    "--videos",
-    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,
-    metavar="SECONDS",
-    help="timeout for all HTTP requests"
 )
 
+configure_common_group_args(GROUP)
+
 
 def load(
         args: argparse.Namespace,
         parser: configparser.ConfigParser,
 ) -> None:
-    log.explain("Creating config for command 'kit-ilias-web'")
+    log.explain(f"Creating config for command '{COMMAND_NAME}'")
 
     parser["crawl:ilias"] = {}
     section = parser["crawl:ilias"]
     load_crawler(args, section)
 
-    section["type"] = "kit-ilias-web"
+    section["type"] = COMMAND_NAME
-    section["target"] = str(args.target)
+    load_common(section, args, parser)
-    section["output_dir"] = str(args.output)
-    section["auth"] = "auth:ilias"
-    if args.links is not None:
-        section["links"] = str(args.links.value)
-    if args.link_redirect_delay is not None:
-        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)
-
-    parser["auth:ilias"] = {}
-    auth_section = parser["auth:ilias"]
-    if args.credential_file is not None:
-        if args.username is not None:
-            raise ParserLoadError("--credential-file and --username can't be used together")
-        if args.keyring:
-            raise ParserLoadError("--credential-file and --keyring can't be used together")
-        auth_section["type"] = "credential-file"
-        auth_section["path"] = str(args.credential_file)
-    elif args.keyring:
-        auth_section["type"] = "keyring"
-    else:
-        auth_section["type"] = "simple"
-    if args.username is not None:
-        auth_section["username"] = args.username
 
 
 SUBPARSER.set_defaults(command=load)

PFERD/cli/common_ilias_args.py

@@ -0,0 +1,104 @@
+import argparse
+import configparser
+from pathlib import Path
+
+from ..crawl.ilias.file_templates import Links
+from .parser import BooleanOptionalAction, ParserLoadError, show_value_error
+
+
+def configure_common_group_args(group: argparse._ArgumentGroup) -> None:
+    """These arguments are shared between the KIT and generic Ilias web command."""
+    group.add_argument(
+        "target",
+        type=str,
+        metavar="TARGET",
+        help="course id, 'desktop', or ILIAS URL to crawl"
+    )
+    group.add_argument(
+        "output",
+        type=Path,
+        metavar="OUTPUT",
+        help="output directory"
+    )
+    group.add_argument(
+        "--username", "-u",
+        type=str,
+        metavar="USERNAME",
+        help="user name for authentication"
+    )
+    group.add_argument(
+        "--keyring",
+        action=BooleanOptionalAction,
+        help="use the system keyring to store and retrieve passwords"
+    )
+    group.add_argument(
+        "--credential-file",
+        type=Path,
+        metavar="PATH",
+        help="read username and password from a credential file"
+    )
+    group.add_argument(
+        "--links",
+        type=show_value_error(Links.from_string),
+        metavar="OPTION",
+        help="how to represent external links"
+    )
+    group.add_argument(
+        "--link-redirect-delay",
+        type=int,
+        metavar="SECONDS",
+        help="time before 'fancy' links redirect to to their target (-1 to disable)"
+    )
+    group.add_argument(
+        "--videos",
+        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,
+        metavar="SECONDS",
+        help="timeout for all HTTP requests"
+    )
+
+
+def load_common(
+    section: configparser.SectionProxy,
+    args: argparse.Namespace,
+    parser: configparser.ConfigParser,
+) -> None:
+    """Load common config between generic and KIT ilias web command"""
+    section["target"] = str(args.target)
+    section["output_dir"] = str(args.output)
+    section["auth"] = "auth:ilias"
+    if args.links is not None:
+        section["links"] = str(args.links.value)
+    if args.link_redirect_delay is not None:
+        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)
+
+    parser["auth:ilias"] = {}
+    auth_section = parser["auth:ilias"]
+    if args.credential_file is not None:
+        if args.username is not None:
+            raise ParserLoadError("--credential-file and --username can't be used together")
+        if args.keyring:
+            raise ParserLoadError("--credential-file and --keyring can't be used together")
+        auth_section["type"] = "credential-file"
+        auth_section["path"] = str(args.credential_file)
+    elif args.keyring:
+        auth_section["type"] = "keyring"
+    else:
+        auth_section["type"] = "simple"
+    if args.username is not None:
+        auth_section["username"] = args.username

PFERD/cli/parser.py

@@ -215,6 +215,11 @@ PARSER.add_argument(
     action=BooleanOptionalAction,
     help="whether crawlers should share cookies where applicable"
 )
+PARSER.add_argument(
+    "--show-not-deleted",
+    action=BooleanOptionalAction,
+    help="print messages in status and report when PFERD did not delete a local only file"
+)
 
 
 def load_default_section(
@@ -233,6 +238,8 @@ def load_default_section(
         section["report"] = "yes" if args.report else "no"
     if args.share_cookies is not None:
         section["share_cookies"] = "yes" if args.share_cookies else "no"
+    if args.show_not_deleted is not None:
+        section["show_not_deleted"] = "yes" if args.show_not_deleted else "no"
 
 
 SUBPARSERS = PARSER.add_subparsers(title="crawlers")

PFERD/config.py

@@ -82,6 +82,9 @@ class DefaultSection(Section):
     def report(self) -> bool:
         return self.s.getboolean("report", fallback=True)
 
+    def show_not_deleted(self) -> bool:
+        return self.s.getboolean("show_not_deleted", fallback=True)
+
     def share_cookies(self) -> bool:
         return self.s.getboolean("share_cookies", fallback=True)
 

PFERD/crawl/__init__.py

@@ -4,7 +4,7 @@ from typing import Callable, Dict
 from ..auth import Authenticator
 from ..config import Config
 from .crawler import Crawler, CrawlError, CrawlerSection  # noqa: F401
-from .ilias import KitIliasWebCrawler, KitIliasWebCrawlerSection
+from .ilias import IliasWebCrawler, IliasWebCrawlerSection, KitIliasWebCrawler, KitIliasWebCrawlerSection
 from .kit_ipd_crawler import KitIpdCrawler, KitIpdCrawlerSection
 from .local_crawler import LocalCrawler, LocalCrawlerSection
 
@@ -18,6 +18,8 @@ CrawlerConstructor = Callable[[
 CRAWLERS: Dict[str, CrawlerConstructor] = {
     "local": lambda n, s, c, a:
         LocalCrawler(n, LocalCrawlerSection(s), c),
+    "ilias-web": lambda n, s, c, a:
+        IliasWebCrawler(n, IliasWebCrawlerSection(s), c, a),
     "kit-ilias-web": lambda n, s, c, a:
         KitIliasWebCrawler(n, KitIliasWebCrawlerSection(s), c, a),
     "kit-ipd": lambda n, s, c, a:

PFERD/crawl/crawler.py

@@ -149,9 +149,7 @@ class CrawlerSection(Section):
         return self.s.getboolean("skip", fallback=False)
 
     def output_dir(self, name: str) -> Path:
-        # TODO Use removeprefix() after switching to 3.9
+        name = name.removeprefix("crawl:")
-        if name.startswith("crawl:"):
-            name = name[len("crawl:"):]
         return Path(self.s.get("output_dir", name)).expanduser()
 
     def redownload(self) -> Redownload:
@@ -258,6 +256,10 @@ class Crawler(ABC):
     def prev_report(self) -> Optional[Report]:
         return self._output_dir.prev_report
 
+    @property
+    def output_dir(self) -> OutputDirectory:
+        return self._output_dir
+
     @staticmethod
     async def gather(awaitables: Sequence[Awaitable[Any]]) -> List[Any]:
         """
@@ -290,9 +292,40 @@ class Crawler(ABC):
         log.explain("Answer: Yes")
         return CrawlToken(self._limiter, path)
 
+    def should_try_download(
+            self,
+            path: PurePath,
+            *,
+            etag_differs: Optional[bool] = None,
+            mtime: Optional[datetime] = None,
+            redownload: Optional[Redownload] = None,
+            on_conflict: Optional[OnConflict] = None,
+    ) -> bool:
+        log.explain_topic(f"Decision: Should Download {fmt_path(path)}")
+
+        if self._transformer.transform(path) is None:
+            log.explain("Answer: No (ignored)")
+            return False
+
+        should_download = self._output_dir.should_try_download(
+            path,
+            etag_differs=etag_differs,
+            mtime=mtime,
+            redownload=redownload,
+            on_conflict=on_conflict
+        )
+        if should_download:
+            log.explain("Answer: Yes")
+            return True
+        else:
+            log.explain("Answer: No")
+            return False
+
     async def download(
             self,
             path: PurePath,
+            *,
+            etag_differs: Optional[bool] = None,
             mtime: Optional[datetime] = None,
             redownload: Optional[Redownload] = None,
             on_conflict: Optional[OnConflict] = None,
@@ -307,7 +340,14 @@ class Crawler(ABC):
             log.status("[bold bright_black]", "Ignored", fmt_path(path))
             return None
 
-        fs_token = await self._output_dir.download(path, transformed_path, mtime, redownload, on_conflict)
+        fs_token = await self._output_dir.download(
+            path,
+            transformed_path,
+            etag_differs=etag_differs,
+            mtime=mtime,
+            redownload=redownload,
+            on_conflict=on_conflict
+        )
         if fs_token is None:
             log.explain("Answer: No")
             return None

PFERD/crawl/http_crawler.py

@@ -1,12 +1,14 @@
 import asyncio
 import http.cookies
 import ssl
+from datetime import datetime
 from pathlib import Path, PurePath
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, List, Optional, Tuple, cast
 
 import aiohttp
 import certifi
 from aiohttp.client import ClientTimeout
+from bs4 import Tag
 
 from ..auth import Authenticator
 from ..config import Config
@@ -15,10 +17,12 @@ from ..utils import fmt_real_path
 from ..version import NAME, VERSION
 from .crawler import Crawler, CrawlerSection
 
+ETAGS_CUSTOM_REPORT_VALUE_KEY = "etags"
+
 
 class HttpCrawlerSection(CrawlerSection):
     def http_timeout(self) -> float:
-        return self.s.getfloat("http_timeout", fallback=20)
+        return self.s.getfloat("http_timeout", fallback=30)
 
 
 class HttpCrawler(Crawler):
@@ -169,6 +173,79 @@ class HttpCrawler(Crawler):
             log.warn(f"Failed to save cookies to {fmt_real_path(self._cookie_jar_path)}")
             log.warn(str(e))
 
+    @staticmethod
+    def get_folder_structure_from_heading_hierarchy(file_link: Tag, drop_h1: bool = False) -> PurePath:
+        """
+        Retrieves the hierarchy of headings associated with the give file link and constructs a folder
+        structure from them.
+
+        <h1> level headings usually only appear once and serve as the page title, so they would introduce
+        redundant nesting. To avoid this, <h1> headings are ignored via the drop_h1 parameter.
+        """
+
+        def find_associated_headings(tag: Tag, level: int) -> PurePath:
+            if level == 0 or (level == 1 and drop_h1):
+                return PurePath()
+
+            level_heading = cast(Optional[Tag], tag.find_previous(name=f"h{level}"))
+
+            if level_heading is None:
+                return find_associated_headings(tag, level - 1)
+
+            folder_name = level_heading.get_text().strip()
+            return find_associated_headings(level_heading, level - 1) / folder_name
+
+        # start at level <h3> because paragraph-level headings are usually too granular for folder names
+        return find_associated_headings(file_link, 3)
+
+    def _get_previous_etag_from_report(self, path: PurePath) -> Optional[str]:
+        """
+        If available, retrieves the entity tag for a given path which was stored in the previous report.
+        """
+        if not self._output_dir.prev_report:
+            return None
+
+        etags = self._output_dir.prev_report.get_custom_value(ETAGS_CUSTOM_REPORT_VALUE_KEY) or {}
+        return etags.get(str(path))
+
+    def _add_etag_to_report(self, path: PurePath, etag: Optional[str]) -> None:
+        """
+        Adds an entity tag for a given path to the report's custom values.
+        """
+        if not etag:
+            return
+
+        etags = self._output_dir.report.get_custom_value(ETAGS_CUSTOM_REPORT_VALUE_KEY) or {}
+        etags[str(path)] = etag
+        self._output_dir.report.add_custom_value(ETAGS_CUSTOM_REPORT_VALUE_KEY, etags)
+
+    async def _request_resource_version(self, resource_url: str) -> Tuple[Optional[str], Optional[datetime]]:
+        """
+        Requests the ETag and Last-Modified headers of a resource via a HEAD request.
+        If no entity tag / modification date can be obtained, the according value will be None.
+        """
+        try:
+            async with self.session.head(resource_url) as resp:
+                if resp.status != 200:
+                    return None, None
+
+                etag_header = resp.headers.get("ETag")
+                last_modified_header = resp.headers.get("Last-Modified")
+                last_modified = None
+
+                if last_modified_header:
+                    try:
+                        # https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Last-Modified#directives
+                        datetime_format = "%a, %d %b %Y %H:%M:%S GMT"
+                        last_modified = datetime.strptime(last_modified_header, datetime_format)
+                    except ValueError:
+                        # last_modified remains None
+                        pass
+
+                return etag_header, last_modified
+        except aiohttp.ClientError:
+            return None, None
+
     async def run(self) -> None:
         self._request_count = 0
         self._cookie_jar = aiohttp.CookieJar()
@@ -186,7 +263,12 @@ class HttpCrawler(Crawler):
                     connect=self._http_timeout,
                     sock_connect=self._http_timeout,
                     sock_read=self._http_timeout,
-                )
+                ),
+                # See https://github.com/aio-libs/aiohttp/issues/6626
+                # Without this aiohttp will mangle the redirect header from Shibboleth, invalidating the
+                # passed signature. Shibboleth will not accept the broken signature and authentication will
+                # fail.
+                requote_redirect_url=False
         ) as session:
             self.session = session
             try:

PFERD/crawl/ilias/__init__.py

@@ -1,3 +1,9 @@
-from .kit_ilias_web_crawler import KitIliasWebCrawler, KitIliasWebCrawlerSection
+from .kit_ilias_web_crawler import (IliasWebCrawler, IliasWebCrawlerSection, KitIliasWebCrawler,
+                                    KitIliasWebCrawlerSection)
 
-__all__ = ["KitIliasWebCrawler", "KitIliasWebCrawlerSection"]
+__all__ = [
+    "IliasWebCrawler",
+    "IliasWebCrawlerSection",
+    "KitIliasWebCrawler",
+    "KitIliasWebCrawlerSection",
+]

PFERD/crawl/ilias/async_helper.py

@@ -0,0 +1,40 @@
+import asyncio
+from typing import Any, Callable, Optional
+
+import aiohttp
+
+from ...logging import log
+from ..crawler import AWrapped, CrawlError, CrawlWarning
+
+
+def _iorepeat(attempts: int, name: str, failure_is_error: bool = False) -> Callable[[AWrapped], AWrapped]:
+    def decorator(f: AWrapped) -> AWrapped:
+        async def wrapper(*args: Any, **kwargs: Any) -> Optional[Any]:
+            last_exception: Optional[BaseException] = None
+            for round in range(attempts):
+                try:
+                    return await f(*args, **kwargs)
+                except aiohttp.ContentTypeError:  # invalid content type
+                    raise CrawlWarning("ILIAS returned an invalid content type")
+                except aiohttp.TooManyRedirects:
+                    raise CrawlWarning("Got stuck in a redirect loop")
+                except aiohttp.ClientPayloadError as e:  # encoding or not enough bytes
+                    last_exception = e
+                except aiohttp.ClientConnectionError as e:  # e.g. timeout, disconnect, resolve failed, etc.
+                    last_exception = e
+                except asyncio.exceptions.TimeoutError as e:  # explicit http timeouts in HttpCrawler
+                    last_exception = e
+                log.explain_topic(f"Retrying operation {name}. Retries left: {attempts - 1 - round}")
+                log.explain(f"Last exception: {last_exception!r}")
+
+            if last_exception:
+                message = f"Error in I/O Operation: {last_exception!r}"
+                if failure_is_error:
+                    raise CrawlError(message) from last_exception
+                else:
+                    raise CrawlWarning(message) from last_exception
+            raise CrawlError("Impossible return in ilias _iorepeat")
+
+        return wrapper  # type: ignore
+
+    return decorator

PFERD/crawl/ilias/file_templates.py

@@ -1,5 +1,11 @@
+import dataclasses
+import re
 from enum import Enum
-from typing import Optional
+from typing import Optional, cast
+
+import bs4
+
+from PFERD.utils import soupify
 
 _link_template_plain = "{{link}}"
 _link_template_fancy = """
@@ -8,7 +14,9 @@ _link_template_fancy = """
     <head>
         <meta charset="UTF-8">
         <title>ILIAS - Link: {{name}}</title>
+        <!-- REPEAT REMOVE START -->
         <meta http-equiv = "refresh" content = "{{redirect_delay}}; url = {{link}}" />
+        <!-- REPEAT REMOVE END -->
     </head>
 
     <style>
@@ -19,6 +27,8 @@ _link_template_fancy = """
         display: flex;
         align-items: center;
         justify-content: center;
+        flex-direction: column;
+        gap: 4px;
     }
     body {
         padding: 0;
@@ -27,11 +37,16 @@ _link_template_fancy = """
         font-family: "Open Sans", Verdana, Arial, Helvetica, sans-serif;
         height: 100vh;
     }
-    .row {
+    .column {
-        background-color: white;
         min-width: 500px;
         max-width: 90vw;
         display: flex;
+        flex-direction: column;
+        row-gap: 5px;
+    }
+    .row {
+        background-color: white;
+        display: flex;
         padding: 1em;
     }
     .logo {
@@ -71,6 +86,8 @@ _link_template_fancy = """
     }
     </style>
     <body class="center-flex">
+        <div class="column">
+        <!-- REPEAT START -->
             <div class="row">
                 <div class="logo center-flex">
                     <svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24">
@@ -85,6 +102,8 @@ _link_template_fancy = """
                 </div>
                 <div class="menu-button center-flex"> ⯆ </div>
             </div>
+        <!-- REPEAT END -->
+        </div>
     </body>
 </html>
 """.strip()  # noqa: E501 line too long
@@ -94,6 +113,169 @@ _link_template_internet_shortcut = """
 URL={{link}}
 """.strip()
 
+_learning_module_template = """
+<!DOCTYPE html>
+<html lang="en">
+    <head>
+        <meta charset="UTF-8">
+        <title>{{name}}</title>
+    </head>
+
+    <style>
+    * {
+        box-sizing: border-box;
+    }
+    .center-flex {
+        display: flex;
+        align-items: center;
+        justify-content: center;
+    }
+    .nav {
+        display: flex;
+        justify-content: space-between;
+    }
+    </style>
+    <body class="center-flex">
+{{body}}
+    </body>
+</html>
+"""
+
+_forum_thread_template = """
+<!DOCTYPE html>
+<html lang="en">
+    <head>
+        <meta charset="UTF-8">
+        <title>ILIAS - Forum: {{name}}</title>
+        <style>
+            * {
+                box-sizing: border-box;
+            }
+            body {
+                font-family: 'Open Sans', Verdana, Arial, Helvetica, sans-serif;
+                padding: 8px;
+            }
+            ul, ol, p {
+                margin: 1.2em 0;
+            }
+            p {
+                margin-top: 8px;
+                margin-bottom: 8px;
+            }
+            a {
+                color: #00876c;
+                text-decoration: none;
+                cursor: pointer;
+            }
+            a:hover {
+                text-decoration: underline;
+            }
+            body > p:first-child > span:first-child {
+                font-size: 1.6em;
+            }
+            body > p:first-child > span:first-child ~ span.default {
+                display: inline-block;
+                font-size: 1.2em;
+                padding-bottom: 8px;
+            }
+            .ilFrmPostContent {
+                margin-top: 8px;
+                max-width: 64em;
+            }
+            .ilFrmPostContent > *:first-child {
+                margin-top: 0px;
+            }
+            .ilFrmPostTitle {
+                margin-top: 24px;
+                color: #00876c;
+                font-weight: bold;
+            }
+            #ilFrmPostList {
+                list-style: none;
+                padding-left: 0;
+            }
+            li.ilFrmPostRow {
+                padding: 3px 0 3px 3px;
+                margin-bottom: 24px;
+                border-left: 6px solid #dddddd;
+            }
+            .ilFrmPostRow > div {
+                display: flex;
+            }
+            .ilFrmPostImage img {
+                margin: 0 !important;
+                padding: 6px 9px 9px 6px;
+            }
+            .ilUserIcon {
+                width: 115px;
+            }
+            .small {
+                text-decoration: none;
+                font-size: 0.75rem;
+                color: #6f6f6f;
+            }
+        </style>
+    </head>
+    <body>
+    {{heading}}
+    {{content}}
+    </body>
+</html>
+""".strip()  # noqa: E501 line too long
+
+
+def learning_module_template(body: bs4.Tag, name: str, prev: Optional[str], next: Optional[str]) -> str:
+    # Seems to be comments, ignore those.
+    for elem in body.select(".il-copg-mob-fullscreen-modal"):
+        elem.decompose()
+
+    nav_template = """
+        <div class="nav">
+            {{left}}
+            {{right}}
+        </div>
+    """
+    if prev and body.select_one(".ilc_page_lnav_LeftNavigation"):
+        text = cast(bs4.Tag, body.select_one(".ilc_page_lnav_LeftNavigation")).get_text().strip()
+        left = f'<a href="{prev}">{text}</a>'
+    else:
+        left = "<span></span>"
+
+    if next and body.select_one(".ilc_page_rnav_RightNavigation"):
+        text = cast(bs4.Tag, body.select_one(".ilc_page_rnav_RightNavigation")).get_text().strip()
+        right = f'<a href="{next}">{text}</a>'
+    else:
+        right = "<span></span>"
+
+    if top_nav := body.select_one(".ilc_page_tnav_TopNavigation"):
+        top_nav.replace_with(
+            soupify(nav_template.replace("{{left}}", left).replace("{{right}}", right).encode())
+        )
+
+    if bot_nav := body.select_one(".ilc_page_bnav_BottomNavigation"):
+        bot_nav.replace_with(soupify(nav_template.replace(
+            "{{left}}", left).replace("{{right}}", right).encode())
+        )
+
+    body_str = cast(str, body.prettify())
+    return _learning_module_template.replace("{{body}}", body_str).replace("{{name}}", name)
+
+
+def forum_thread_template(name: str, url: str, heading: bs4.Tag, content: bs4.Tag) -> str:
+    if title := cast(Optional[bs4.Tag], heading.find(name="b")):
+        title.wrap(bs4.Tag(name="a", attrs={"href": url}))
+    return _forum_thread_template \
+        .replace("{{name}}", name) \
+        .replace("{{heading}}", cast(str, heading.prettify())) \
+        .replace("{{content}}", cast(str, content.prettify()))
+
+
+@dataclasses.dataclass
+class LinkData:
+    name: str
+    url: str
+    description: str
+
 
 class Links(Enum):
     IGNORE = "ignore"
@@ -102,31 +284,74 @@ class Links(Enum):
     INTERNET_SHORTCUT = "internet-shortcut"
 
     def template(self) -> Optional[str]:
-        if self == self.FANCY:
+        if self == Links.FANCY:
             return _link_template_fancy
-        elif self == self.PLAINTEXT:
+        elif self == Links.PLAINTEXT:
             return _link_template_plain
-        elif self == self.INTERNET_SHORTCUT:
+        elif self == Links.INTERNET_SHORTCUT:
             return _link_template_internet_shortcut
-        elif self == self.IGNORE:
+        elif self == Links.IGNORE:
             return None
         raise ValueError("Missing switch case")
 
+    def collection_as_one(self) -> bool:
+        if self == Links.FANCY:
+            return True
+        return False
+
     def extension(self) -> Optional[str]:
-        if self == self.FANCY:
+        if self == Links.FANCY:
             return ".html"
-        elif self == self.PLAINTEXT:
+        elif self == Links.PLAINTEXT:
             return ".txt"
-        elif self == self.INTERNET_SHORTCUT:
+        elif self == Links.INTERNET_SHORTCUT:
             return ".url"
-        elif self == self.IGNORE:
+        elif self == Links.IGNORE:
             return None
         raise ValueError("Missing switch case")
 
+    def interpolate(self, redirect_delay: int, collection_name: str, links: list[LinkData]) -> str:
+        template = self.template()
+        if template is None:
+            raise ValueError("Cannot interpolate ignored links")
+
+        if len(links) == 1:
+            link = links[0]
+            content = template
+            content = content.replace("{{link}}", link.url)
+            content = content.replace("{{name}}", link.name)
+            content = content.replace("{{description}}", link.description)
+            content = content.replace("{{redirect_delay}}", str(redirect_delay))
+            return content
+        if self == Links.PLAINTEXT or self == Links.INTERNET_SHORTCUT:
+            return "\n".join(f"{link.url}" for link in links)
+
+        # All others get coerced to fancy
+        content = cast(str, Links.FANCY.template())
+        repeated_content = cast(
+            re.Match[str],
+            re.search(r"<!-- REPEAT START -->([\s\S]+)<!-- REPEAT END -->", content)
+        ).group(1)
+
+        parts = []
+        for link in links:
+            instance = repeated_content
+            instance = instance.replace("{{link}}", link.url)
+            instance = instance.replace("{{name}}", link.name)
+            instance = instance.replace("{{description}}", link.description)
+            instance = instance.replace("{{redirect_delay}}", str(redirect_delay))
+            parts.append(instance)
+
+        content = content.replace(repeated_content, "\n".join(parts))
+        content = content.replace("{{name}}", collection_name)
+        content = re.sub(r"<!-- REPEAT REMOVE START -->[\s\S]+<!-- REPEAT REMOVE END -->", "", content)
+
+        return content
+
     @staticmethod
     def from_string(string: str) -> "Links":
         try:
             return Links(string)
         except ValueError:
-            raise ValueError("must be one of 'ignore', 'plaintext',"
+            options = [f"'{option.value}'" for option in Links]
-                             " 'html', 'internet-shortcut'")
+            raise ValueError(f"must be one of {', '.join(options)}")

PFERD/crawl/ilias/ilias_html_cleaner.py

@@ -1,3 +1,5 @@
+from typing import cast
+
 from bs4 import BeautifulSoup, Comment, Tag
 
 _STYLE_TAG_CONTENT = """
@@ -12,6 +14,13 @@ _STYLE_TAG_CONTENT = """
       font-weight: bold;
     }
 
+    .row-flex {
+      display: flex;
+    }
+    .row-flex-wrap {
+      flex-wrap: wrap;
+    }
+
     .accordion-head {
       background-color: #f5f7fa;
       padding: 0.5rem 0;
@@ -30,6 +39,10 @@ _STYLE_TAG_CONTENT = """
       margin: 0.5rem 0;
     }
 
+    img {
+        background-color: white;
+    }
+
     body {
       padding: 1em;
       grid-template-columns: 1fr min(60rem, 90%) 1fr;
@@ -47,12 +60,11 @@ _ARTICLE_WORTHY_CLASSES = [
 def insert_base_markup(soup: BeautifulSoup) -> BeautifulSoup:
     head = soup.new_tag("head")
     soup.insert(0, head)
+    # Force UTF-8 encoding
+    head.append(soup.new_tag("meta", charset="utf-8"))
 
-    simplecss_link: Tag = soup.new_tag("link")
     # <link rel="stylesheet" href="https://cdn.simplecss.org/simple.css">
-    simplecss_link["rel"] = "stylesheet"
+    head.append(soup.new_tag("link", rel="stylesheet", href="https://cdn.simplecss.org/simple.css"))
-    simplecss_link["href"] = "https://cdn.simplecss.org/simple.css"
-    head.append(simplecss_link)
 
     # Basic style tags for compat
     style: Tag = soup.new_tag("style")
@@ -63,18 +75,18 @@ def insert_base_markup(soup: BeautifulSoup) -> BeautifulSoup:
 
 
 def clean(soup: BeautifulSoup) -> BeautifulSoup:
-    for block in soup.find_all(class_=lambda x: x in _ARTICLE_WORTHY_CLASSES):
+    for block in cast(list[Tag], soup.find_all(class_=lambda x: x in _ARTICLE_WORTHY_CLASSES)):
         block.name = "article"
 
-    for block in soup.find_all("h3"):
+    for block in cast(list[Tag], soup.find_all("h3")):
         block.name = "div"
 
-    for block in soup.find_all("h1"):
+    for block in cast(list[Tag], soup.find_all("h1")):
         block.name = "h3"
 
-    for block in soup.find_all(class_="ilc_va_ihcap_VAccordIHeadCap"):
+    for block in cast(list[Tag], soup.find_all(class_="ilc_va_ihcap_VAccordIHeadCap")):
         block.name = "h3"
-        block["class"] += ["accordion-head"]
+        block["class"] += ["accordion-head"]  # type: ignore
 
     for dummy in soup.select(".ilc_text_block_Standard.ilc_Paragraph"):
         children = list(dummy.children)
@@ -82,10 +94,15 @@ def clean(soup: BeautifulSoup) -> BeautifulSoup:
             dummy.decompose()
         if len(children) > 1:
             continue
-        if type(children[0]) == Comment:
+        if isinstance(type(children[0]), Comment):
             dummy.decompose()
 
-    for hrule_imposter in soup.find_all(class_="ilc_section_Separator"):
+    # Delete video figures, as they can not be internalized anyway
+    for video in soup.select(".ilc_media_cont_MediaContainerHighlighted .ilPageVideo"):
+        if figure := video.find_parent("figure"):
+            figure.decompose()
+
+    for hrule_imposter in cast(list[Tag], soup.find_all(class_="ilc_section_Separator")):
         hrule_imposter.insert(0, soup.new_tag("hr"))
 
     return soup

PFERD/crawl/ilias/kit_ilias_web_crawler.py

@@ -1,933 +1,37 @@
-import asyncio
+from typing import Dict, Literal
-import re
-from collections.abc import Awaitable, Coroutine
-from pathlib import PurePath
-from typing import Any, Callable, Dict, List, Optional, Set, Union, cast
 
-import aiohttp
+from ...auth import Authenticator
-import yarl
-from aiohttp import hdrs
-from bs4 import BeautifulSoup, Tag
-
-from ...auth import Authenticator, TfaAuthenticator
 from ...config import Config
-from ...logging import ProgressBar, log
+from .ilias_web_crawler import IliasWebCrawler, IliasWebCrawlerSection
-from ...output_dir import FileSink, Redownload
+from .shibboleth_login import ShibbolethLogin
-from ...utils import fmt_path, soupify, url_set_query_param
-from ..crawler import AWrapped, CrawlError, CrawlToken, CrawlWarning, DownloadToken, anoncritical
-from ..http_crawler import HttpCrawler, HttpCrawlerSection
-from .file_templates import Links
-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 KitIliasWebCrawlerSection(HttpCrawlerSection):
+class KitShibbolethBackgroundLoginSuccessful:
-    def target(self) -> TargetType:
+    pass
-        target = self.s.get("target")
-        if not target:
-            self.missing_value("target")
-
-        if re.fullmatch(r"\d+", target):
-            # Course id
-            return int(target)
-        if target == "desktop":
-            # Full personal desktop
-            return target
-        if target.startswith("https://ilias.studium.kit.edu"):
-            # ILIAS URL
-            return target
-
-        self.invalid_value("target", target, "Should be <course id | desktop | kit ilias URL>")
-
-    def tfa_auth(self, authenticators: Dict[str, Authenticator]) -> Optional[Authenticator]:
-        value: Optional[str] = self.s.get("tfa_auth")
-        if value is None:
-            return None
-        auth = authenticators.get(value)
-        if auth is None:
-            self.invalid_value("tfa_auth", value, "No such auth section exists")
-        return auth
-
-    def links(self) -> Links:
-        type_str: Optional[str] = self.s.get("links")
-
-        if type_str is None:
-            return Links.FANCY
-
-        try:
-            return Links.from_string(type_str)
-        except ValueError as e:
-            self.invalid_value("links", type_str, str(e).capitalize())
-
-    def link_redirect_delay(self) -> int:
-        return self.s.getint("link_redirect_delay", fallback=-1)
-
-    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([
+class KitIliasWebCrawlerSection(IliasWebCrawlerSection):
-    IliasElementType.EXERCISE,
+    def base_url(self) -> str:
-    IliasElementType.EXERCISE_FILES,
+        return _ILIAS_URL
-    IliasElementType.FOLDER,
-    IliasElementType.MEETING,
-    IliasElementType.VIDEO_FOLDER,
-    IliasElementType.VIDEO_FOLDER_MAYBE_PAGINATED,
-])
 
-_VIDEO_ELEMENTS: Set[IliasElementType] = set([
+    def login(self) -> Literal["shibboleth"]:
-    IliasElementType.VIDEO,
+        return "shibboleth"
-    IliasElementType.VIDEO_PLAYER,
-    IliasElementType.VIDEO_FOLDER,
-    IliasElementType.VIDEO_FOLDER_MAYBE_PAGINATED,
-])
 
 
-def _iorepeat(attempts: int, name: str, failure_is_error: bool = False) -> Callable[[AWrapped], AWrapped]:
+class KitIliasWebCrawler(IliasWebCrawler):
-    def decorator(f: AWrapped) -> AWrapped:
-        async def wrapper(*args: Any, **kwargs: Any) -> Optional[Any]:
-            last_exception: Optional[BaseException] = None
-            for round in range(attempts):
-                try:
-                    return await f(*args, **kwargs)
-                except aiohttp.ContentTypeError:  # invalid content type
-                    raise CrawlWarning("ILIAS returned an invalid content type")
-                except aiohttp.TooManyRedirects:
-                    raise CrawlWarning("Got stuck in a redirect loop")
-                except aiohttp.ClientPayloadError as e:  # encoding or not enough bytes
-                    last_exception = e
-                except aiohttp.ClientConnectionError as e:  # e.g. timeout, disconnect, resolve failed, etc.
-                    last_exception = e
-                except asyncio.exceptions.TimeoutError as e:  # explicit http timeouts in HttpCrawler
-                    last_exception = e
-                log.explain_topic(f"Retrying operation {name}. Retries left: {attempts - 1 - round}")
-
-            if last_exception:
-                message = f"Error in I/O Operation: {last_exception}"
-                if failure_is_error:
-                    raise CrawlError(message) from last_exception
-                else:
-                    raise CrawlWarning(message) from last_exception
-            raise CrawlError("Impossible return in ilias _iorepeat")
-
-        return wrapper  # type: ignore
-    return decorator
-
-
-def _wrap_io_in_warning(name: str) -> Callable[[AWrapped], AWrapped]:
-    """
-    Wraps any I/O exception in a CrawlWarning.
-    """
-    return _iorepeat(1, name)
-
-
-# Crawler control flow:
-#
-#     crawl_desktop -+
-#                    |
-#     crawl_course --+
-#                    |
-#     @_io_repeat    |        # retries internally (before the bar)
-#  +- crawl_url    <-+
-#  |
-#  |
-#  |  @_wrap_io_exception     # does not need to retry as children acquire bars
-#  +> crawl_ilias_element -+
-#  ^                       |
-#  |  @_io_repeat          |  # retries internally (before the bar)
-#  +- crawl_ilias_page <---+
-#  |                       |
-#  +> get_page             |  # Handles and retries authentication
-#                          |
-#     @_io_repeat          |  # retries internally (before the bar)
-#  +- download_link    <---+
-#  |                       |
-#  +> resolve_target       |  # Handles and retries authentication
-#                          |
-#     @_io_repeat          |  # retries internally (before the bar)
-#  +- download_video   <---+
-#  |                       |
-#  |  @_io_repeat          |  # retries internally (before the bar)
-#  +- download_file    <---+
-#  |
-#  +> stream_from_url         # Handles and retries authentication
-
-class KitIliasWebCrawler(HttpCrawler):
     def __init__(
         self,
         name: str,
         section: KitIliasWebCrawlerSection,
         config: Config,
-            authenticators: Dict[str, Authenticator]
+        authenticators: Dict[str, Authenticator],
     ):
-        # Setting a main authenticator for cookie sharing
+        super().__init__(name, section, config, authenticators)
-        auth = section.auth(authenticators)
-        super().__init__(name, section, config, shared_auth=auth)
 
-        if section.tasks() > 1:
+        self._shibboleth_login = ShibbolethLogin(
-            log.warn("""
+            _ILIAS_URL,
-Please avoid using too many parallel requests as these are the KIT ILIAS
+            self._auth,
-instance's greatest bottleneck.
-            """.strip())
-
-        self._shibboleth_login = KitShibbolethLogin(
-            auth,
             section.tfa_auth(authenticators),
         )
-
-        self._base_url = "https://ilias.studium.kit.edu"
-
-        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:
-        if isinstance(self._target, int):
-            log.explain_topic(f"Inferred crawl target: Course with id {self._target}")
-            await self._crawl_course(self._target)
-        elif self._target == "desktop":
-            log.explain_topic("Inferred crawl target: Personal desktop")
-            await self._crawl_desktop()
-        else:
-            log.explain_topic(f"Inferred crawl target: URL {self._target}")
-            await self._crawl_url(self._target)
-
-    async def _crawl_course(self, course_id: int) -> None:
-        # Start crawling at the given course
-        root_url = url_set_query_param(
-            self._base_url + "/goto.php", "target", f"crs_{course_id}"
-        )
-
-        await self._crawl_url(root_url, expected_id=course_id)
-
-    async def _crawl_desktop(self) -> None:
-        appendix = r"ILIAS\PersonalDesktop\PDMainBarProvider|mm_pd_sel_items"
-        appendix = appendix.encode("ASCII").hex()
-        await self._crawl_url(self._base_url + "/gs_content.php?item=" + appendix)
-
-    async def _crawl_url(self, url: str, expected_id: Optional[int] = None) -> None:
-        maybe_cl = await self.crawl(PurePath("."))
-        if not maybe_cl:
-            return
-        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)
-
-                if 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("Parsing root HTML page")
-                log.explain(f"URL: {url}")
-                page = IliasPage(soup, 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]] = []
-        for element in elements:
-            if handle := await self._handle_ilias_element(PurePath("."), element):
-                tasks.append(asyncio.create_task(handle))
-
-        # And execute them
-        await self.gather(tasks)
-
-    async def _handle_ilias_page(
-        self,
-        url: str,
-        parent: IliasPageElement,
-        path: PurePath,
-    ) -> Optional[Coroutine[Any, Any, None]]:
-        maybe_cl = await self.crawl(path)
-        if not maybe_cl:
-            return None
-        return self._crawl_ilias_page(url, parent, maybe_cl)
-
-    @anoncritical
-    async def _crawl_ilias_page(
-        self,
-        url: str,
-        parent: IliasPageElement,
-        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:
-            elements.clear()
-            async with cl:
-                next_stage_url: Optional[str] = url
-                current_parent = parent
-
-                while next_stage_url:
-                    soup = await self._get_page(next_stage_url)
-                    log.explain_topic(f"Parsing HTML page for {fmt_path(cl.path)}")
-                    log.explain(f"URL: {next_stage_url}")
-                    page = IliasPage(soup, next_stage_url, current_parent)
-                    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(cl.path, description[0])
-
-        elements.sort(key=lambda e: e.id())
-
-        tasks: List[Awaitable[None]] = []
-        for element in elements:
-            if handle := await self._handle_ilias_element(cl.path, element):
-                tasks.append(asyncio.create_task(handle))
-
-        # And execute them
-        await self.gather(tasks)
-
-    # These decorators only apply *to this method* and *NOT* to the returned
-    # awaitables!
-    # This method does not await the handlers but returns them instead.
-    # This ensures one level is handled at a time and name deduplication
-    # works correctly.
-    @anoncritical
-    async def _handle_ilias_element(
-        self,
-        parent_path: PurePath,
-        element: IliasPageElement,
-    ) -> 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"
-            )
-        self._visited_urls.add(element.url)
-
-        element_path = PurePath(parent_path, element.name)
-
-        if element.type in _VIDEO_ELEMENTS:
-            if not self._videos:
-                log.status(
-                    "[bold bright_black]",
-                    "Ignored",
-                    fmt_path(element_path),
-                    "[bright_black](enable with option 'videos')"
-                )
-                return None
-
-        if element.type == IliasElementType.FILE:
-            return await self._handle_file(element, element_path)
-        elif element.type == IliasElementType.FORUM:
-            if not self._forums:
-                log.status(
-                    "[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.explain("Tests contain no relevant files")
-            log.explain("Answer: No")
-            return None
-        elif element.type == IliasElementType.LINK:
-            return await self._handle_link(element, element_path)
-        elif element.type == IliasElementType.BOOKING:
-            return await self._handle_booking(element, element_path)
-        elif element.type == IliasElementType.VIDEO:
-            return await self._handle_file(element, element_path)
-        elif element.type == IliasElementType.VIDEO_PLAYER:
-            return await self._handle_video(element, element_path)
-        elif element.type in _DIRECTORY_PAGES:
-            return await self._handle_ilias_page(element.url, element, element_path)
-        else:
-            # This will retry it a few times, failing everytime. It doesn't make any network
-            # requests, so that's fine.
-            raise CrawlWarning(f"Unknown element type: {element.type!r}")
-
-    async def _handle_link(
-        self,
-        element: IliasPageElement,
-        element_path: PurePath,
-    ) -> Optional[Coroutine[Any, Any, None]]:
-        log.explain_topic(f"Decision: Crawl Link {fmt_path(element_path)}")
-        log.explain(f"Links type is {self._links}")
-
-        link_template_maybe = self._links.template()
-        link_extension = self._links.extension()
-        if not link_template_maybe or not link_extension:
-            log.explain("Answer: No")
-            return None
-        else:
-            log.explain("Answer: Yes")
-        element_path = element_path.with_name(element_path.name + link_extension)
-
-        maybe_dl = await self.download(element_path, mtime=element.mtime)
-        if not maybe_dl:
-            return None
-
-        return self._download_link(element, link_template_maybe, maybe_dl)
-
-    @anoncritical
-    @_iorepeat(3, "resolving link")
-    async def _download_link(self, element: IliasPageElement, link_template: str, dl: DownloadToken) -> None:
-        async with dl as (bar, sink):
-            export_url = element.url.replace("cmd=calldirectlink", "cmd=exportHTML")
-            real_url = await self._resolve_link_target(export_url)
-            self._write_link_content(link_template, real_url, element.name, element.description, sink)
-
-    def _write_link_content(
-        self,
-        link_template: str,
-        url: str,
-        name: str,
-        description: Optional[str],
-        sink: FileSink,
-    ) -> None:
-        content = link_template
-        content = content.replace("{{link}}", url)
-        content = content.replace("{{name}}", name)
-        content = content.replace("{{description}}", str(description))
-        content = content.replace("{{redirect_delay}}", str(self._link_file_redirect_delay))
-        sink.file.write(content.encode("utf-8"))
-        sink.done()
-
-    async def _handle_booking(
-        self,
-        element: IliasPageElement,
-        element_path: PurePath,
-    ) -> 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}")
-
-        link_template_maybe = self._links.template()
-        link_extension = self._links.extension()
-        if not link_template_maybe or not link_extension:
-            log.explain("Answer: No")
-            return None
-        else:
-            log.explain("Answer: Yes")
-        element_path = element_path.with_name(element_path.name + link_extension)
-
-        maybe_dl = await self.download(element_path, mtime=element.mtime)
-        if not maybe_dl:
-            return None
-
-        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(
-        self,
-        element: IliasPageElement,
-        link_template: str,
-        dl: DownloadToken,
-    ) -> None:
-        async with dl as (bar, sink):
-            self._write_link_content(link_template, element.url, element.name, element.description, sink)
-
-    async def _resolve_link_target(self, export_url: str) -> str:
-        async with self.session.get(export_url, allow_redirects=False) as resp:
-            # No redirect means we were authenticated
-            if hdrs.LOCATION not in resp.headers:
-                return soupify(await resp.read()).select_one("a").get("href").strip()
-
-        await self._authenticate()
-
-        async with self.session.get(export_url, allow_redirects=False) as resp:
-            # No redirect means we were authenticated
-            if hdrs.LOCATION not in resp.headers:
-                return soupify(await resp.read()).select_one("a").get("href").strip()
-
-        raise CrawlError("resolve_link_target failed even after authenticating")
-
-    async def _handle_video(
-        self,
-        element: IliasPageElement,
-        element_path: PurePath,
-    ) -> Optional[Coroutine[Any, Any, None]]:
-        # Copy old mapping as it is likely still relevant
-        if self.prev_report:
-            self.report.add_custom_value(
-                str(element_path),
-                self.prev_report.get_custom_value(str(element_path))
-            )
-
-        # A video might contain other videos, so let's "crawl" the video first
-        # to ensure rate limits apply. This must be a download as *this token*
-        # is re-used if the video consists of a single stream. In that case the
-        # file name is used and *not* the stream name the ilias html parser reported
-        # to ensure backwards compatibility.
-        maybe_dl = await self.download(element_path, mtime=element.mtime, redownload=Redownload.ALWAYS)
-
-        # If we do not want to crawl it (user filter) or we have every file
-        # from the cached mapping already, we can ignore this and bail
-        if not maybe_dl or self._all_videos_locally_present(element_path):
-            # Mark all existing cideos as known so they do not get deleted
-            # during dleanup. We "downloaded" them, just without actually making
-            # a network request as we assumed they did not change.
-            for video in self._previous_contained_videos(element_path):
-                await self.download(video)
-
-            return None
-
-        return self._download_video(element_path, element, maybe_dl)
-
-    def _previous_contained_videos(self, video_path: PurePath) -> List[PurePath]:
-        if not self.prev_report:
-            return []
-        custom_value = self.prev_report.get_custom_value(str(video_path))
-        if not custom_value:
-            return []
-        names = cast(List[str], custom_value)
-        folder = video_path.parent
-        return [PurePath(folder, name) for name in names]
-
-    def _all_videos_locally_present(self, video_path: PurePath) -> bool:
-        if contained_videos := self._previous_contained_videos(video_path):
-            log.explain_topic(f"Checking local cache for video {video_path.name}")
-            all_found_locally = True
-            for video in contained_videos:
-                transformed_path = self._to_local_video_path(video)
-                if transformed_path:
-                    exists_locally = self._output_dir.resolve(transformed_path).exists()
-                    all_found_locally = all_found_locally and exists_locally
-            if all_found_locally:
-                log.explain("Found all videos locally, skipping enumeration request")
-                return True
-            log.explain("Missing at least one video, continuing with requests!")
-        return False
-
-    def _to_local_video_path(self, path: PurePath) -> Optional[PurePath]:
-        if transformed := self._transformer.transform(path):
-            return self._deduplicator.fixup_path(transformed)
-        return None
-
-    @anoncritical
-    @_iorepeat(3, "downloading video")
-    async def _download_video(
-        self,
-        original_path: PurePath,
-        element: IliasPageElement,
-        dl: DownloadToken
-    ) -> None:
-        stream_elements: List[IliasPageElement] = []
-        async with dl as (bar, sink):
-            page = IliasPage(await self._get_page(element.url), element.url, element)
-            stream_elements = page.get_child_elements()
-
-            if len(stream_elements) > 1:
-                log.explain(f"Found multiple video streams for {element.name}")
-            else:
-                log.explain(f"Using single video mode for {element.name}")
-                stream_element = stream_elements[0]
-
-                transformed_path = self._to_local_video_path(original_path)
-                if not transformed_path:
-                    raise CrawlError(f"Download returned a path but transform did not for {original_path}")
-
-                # We do not have a local cache yet
-                if self._output_dir.resolve(transformed_path).exists():
-                    log.explain(f"Video for {element.name} existed locally")
-                else:
-                    await self._stream_from_url(stream_element.url, sink, bar, is_video=True)
-                self.report.add_custom_value(str(original_path), [original_path.name])
-                return
-
-        contained_video_paths: List[str] = []
-
-        for stream_element in stream_elements:
-            video_path = original_path.parent / stream_element.name
-            contained_video_paths.append(str(video_path))
-
-            maybe_dl = await self.download(video_path, mtime=element.mtime, redownload=Redownload.NEVER)
-            if not maybe_dl:
-                continue
-            async with maybe_dl as (bar, sink):
-                log.explain(f"Streaming video from real url {stream_element.url}")
-                await self._stream_from_url(stream_element.url, sink, bar, is_video=True)
-
-        self.report.add_custom_value(str(original_path), contained_video_paths)
-
-    async def _handle_file(
-        self,
-        element: IliasPageElement,
-        element_path: PurePath,
-    ) -> Optional[Coroutine[Any, Any, None]]:
-        maybe_dl = await self.download(element_path, mtime=element.mtime)
-        if not maybe_dl:
-            return None
-        return self._download_file(element, maybe_dl)
-
-    @anoncritical
-    @_iorepeat(3, "downloading file")
-    async def _download_file(self, element: IliasPageElement, dl: DownloadToken) -> None:
-        assert dl  # The function is only reached when dl is not None
-        async with dl as (bar, sink):
-            await self._stream_from_url(element.url, sink, bar, is_video=False)
-
-    async def _stream_from_url(self, url: str, sink: FileSink, bar: ProgressBar, is_video: bool) -> None:
-        async def try_stream() -> bool:
-            async with self.session.get(url, allow_redirects=is_video) as resp:
-                if not is_video:
-                    # Redirect means we weren't authenticated
-                    if hdrs.LOCATION in resp.headers:
-                        return False
-                # we wanted a video but got HTML
-                if is_video and "html" in resp.content_type:
-                    return False
-
-                if resp.content_length:
-                    bar.set_total(resp.content_length)
-
-                async for data in resp.content.iter_chunked(1024):
-                    sink.file.write(data)
-                    bar.advance(len(data))
-
-                sink.done()
-            return True
-
-        auth_id = await self._current_auth_id()
-        if await try_stream():
-            return
-
-        await self.authenticate(auth_id)
-
-        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 = []
-
-        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")
-            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:
-            soup = soupify(await request.read())
-            if self._is_logged_in(soup):
-                return soup
-
-        # 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.get(url) as request:
-            soup = soupify(await request.read())
-            if self._is_logged_in(soup):
-                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)
-    async def _authenticate(self) -> None:
-        await self._shibboleth_login.login(self.session)
-
-    @ staticmethod
-    def _is_logged_in(soup: BeautifulSoup) -> bool:
-        # Normal ILIAS pages
-        mainbar: Optional[Tag] = soup.find(class_="il-maincontrols-metabar")
-        if mainbar is not None:
-            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
-
-        # Personal Desktop
-        if soup.find("a", attrs={"href": lambda x: x and "block_type=pditems" in x}):
-            return True
-
-        # Video listing embeds do not have complete ILIAS html. Try to match them by
-        # their video listing table
-        video_table = soup.find(
-            recursive=True,
-            name="table",
-            attrs={"id": lambda x: x is not None and x.startswith("tbl_xoct")}
-        )
-        if video_table is not None:
-            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:
-            return True
-        return False
-
-
-class KitShibbolethLogin:
-    """
-    Login via KIT's shibboleth system.
-    """
-
-    def __init__(self, authenticator: Authenticator, tfa_authenticator: Optional[Authenticator]) -> None:
-        self._auth = authenticator
-        self._tfa_auth = tfa_authenticator
-
-    async def login(self, sess: aiohttp.ClientSession) -> None:
-        """
-        Performs the ILIAS Shibboleth authentication dance and saves the login
-        cookies it receieves.
-
-        This function should only be called whenever it is detected that you're
-        not logged in. The cookies obtained should be good for a few minutes,
-        maybe even an hour or two.
-        """
-
-        # 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"
-        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)
-
-        # Attempt to login using credentials, if necessary
-        while not self._login_successful(soup):
-            # Searching the form here so that this fails before asking for
-            # credentials rather than after asking.
-            form = soup.find("form", {"class": "full content", "method": "post"})
-            action = form["action"]
-
-            csrf_token = form.find("input", {"name": "csrf_token"})["value"]
-
-            # Equivalent: Enter credentials in
-            # https://idp.scc.kit.edu/idp/profile/SAML2/Redirect/SSO
-            url = "https://idp.scc.kit.edu" + action
-            username, password = await self._auth.credentials()
-            data = {
-                "_eventId_proceed": "",
-                "j_username": username,
-                "j_password": password,
-                "csrf_token": csrf_token
-            }
-            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)
-
-            if not self._login_successful(soup):
-                self._auth.invalidate_credentials()
-
-        # Equivalent: Being redirected via JS automatically
-        # (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"
-        data = {  # using the info obtained in the while loop above
-            "RelayState": relay_state["value"],
-            "SAMLResponse": saml_response["value"],
-        }
-        await sess.post(url, data=data)
-
-    async def _authenticate_tfa(
-            self,
-            session: aiohttp.ClientSession,
-            soup: BeautifulSoup
-    ) -> BeautifulSoup:
-        if not self._tfa_auth:
-            self._tfa_auth = TfaAuthenticator("ilias-anon-tfa")
-
-        tfa_token = await self._tfa_auth.password()
-
-        # Searching the form here so that this fails before asking for
-        # credentials rather than after asking.
-        form = soup.find("form", {"method": "post"})
-        action = form["action"]
-        csrf_token = form.find("input", {"name": "csrf_token"})["value"]
-
-        # Equivalent: Enter token in
-        # https://idp.scc.kit.edu/idp/profile/SAML2/Redirect/SSO
-        url = "https://idp.scc.kit.edu" + action
-        data = {
-            "_eventId_proceed": "",
-            "j_tokenNumber": tfa_token,
-            "csrf_token": csrf_token
-        }
-        return await _post(session, url, data)
-
-    @staticmethod
-    def _login_successful(soup: BeautifulSoup) -> bool:
-        relay_state = soup.find("input", {"name": "RelayState"})
-        saml_response = soup.find("input", {"name": "SAMLResponse"})
-        return relay_state is not None and saml_response is not None
-
-    @staticmethod
-    def _tfa_required(soup: BeautifulSoup) -> bool:
-        return soup.find(id="j_tokenNumber") is not None
-
-
-async def _post(session: aiohttp.ClientSession, url: str, data: Any) -> BeautifulSoup:
-    async with session.post(url, data=data) as response:
-        return soupify(await response.read())
-
-
-async def _shib_post(session: aiohttp.ClientSession, url: str, data: Any) -> BeautifulSoup:
-    """
-    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??
-    """
-    async with session.post(url, data=data, allow_redirects=False) as response:
-        location = response.headers.get("location")
-        if not location:
-            raise CrawlWarning(f"Login failed (1), no location header present at {url}")
-        correct_url = yarl.URL(location, encoded=True)
-
-        async with session.get(correct_url, allow_redirects=False) as response:
-            location = response.headers.get("location")
-            # If shib still still has a valid session, it will directly respond to the request
-            if location is None:
-                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 (2), no location header present at {correct_url}")
-
-            correct_url = yarl.URL.build(
-                scheme=as_yarl.scheme,
-                host=as_yarl.host,
-                path=location,
-                encoded=True
-            )
-
-            async with session.get(correct_url, allow_redirects=False) as response:
-                return soupify(await response.read())

PFERD/crawl/ilias/shibboleth_login.py

@@ -0,0 +1,129 @@
+from typing import Any, Optional, cast
+
+import aiohttp
+import yarl
+from bs4 import BeautifulSoup, Tag
+
+from ...auth import Authenticator, TfaAuthenticator
+from ...logging import log
+from ...utils import soupify
+from ..crawler import CrawlError
+
+
+class ShibbolethLogin:
+    """
+    Login via shibboleth system.
+    """
+
+    def __init__(
+        self, ilias_url: str, authenticator: Authenticator, tfa_authenticator: Optional[Authenticator]
+    ) -> None:
+        self._ilias_url = ilias_url
+        self._auth = authenticator
+        self._tfa_auth = tfa_authenticator
+
+    async def login(self, sess: aiohttp.ClientSession) -> None:
+        """
+        Performs the ILIAS Shibboleth authentication dance and saves the login
+        cookies it receieves.
+
+        This function should only be called whenever it is detected that you're
+        not logged in. The cookies obtained should be good for a few minutes,
+        maybe even an hour or two.
+        """
+
+        # Equivalent: Click on "Mit KIT-Account anmelden" button in
+        # https://ilias.studium.kit.edu/login.php
+        url = f"{self._ilias_url}/shib_login.php"
+        async with sess.get(url) as response:
+            shib_url = response.url
+            if str(shib_url).startswith(self._ilias_url):
+                log.explain(
+                    "ILIAS recognized our shib token and logged us in in the background, returning"
+                )
+                return
+            soup: BeautifulSoup = soupify(await response.read())
+
+        # Attempt to login using credentials, if necessary
+        while not self._login_successful(soup):
+            # Searching the form here so that this fails before asking for
+            # credentials rather than after asking.
+            form = cast(Tag, soup.find("form", {"method": "post"}))
+            action = cast(str, form["action"])
+
+            # Equivalent: Enter credentials in
+            # https://idp.scc.kit.edu/idp/profile/SAML2/Redirect/SSO
+            url = str(shib_url.origin()) + action
+            username, password = await self._auth.credentials()
+            data = {
+                "_eventId_proceed": "",
+                "j_username": username,
+                "j_password": password,
+                "fudis_web_authn_assertion_input": "",
+            }
+            if csrf_token_input := form.find("input", {"name": "csrf_token"}):
+                data["csrf_token"] = csrf_token_input["value"]  # type: ignore
+            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, shib_url)
+
+            if not self._login_successful(soup):
+                self._auth.invalidate_credentials()
+
+        # Equivalent: Being redirected via JS automatically
+        # (or clicking "Continue" if you have JS disabled)
+        relay_state = cast(Tag, soup.find("input", {"name": "RelayState"}))
+        saml_response = cast(Tag, soup.find("input", {"name": "SAMLResponse"}))
+        url = form = soup.find("form", {"method": "post"})["action"]  # type: ignore
+        data = {  # using the info obtained in the while loop above
+            "RelayState": cast(str, relay_state["value"]),
+            "SAMLResponse": cast(str, saml_response["value"]),
+        }
+        await sess.post(cast(str, url), data=data)
+
+    async def _authenticate_tfa(
+        self, session: aiohttp.ClientSession, soup: BeautifulSoup, shib_url: yarl.URL
+    ) -> BeautifulSoup:
+        if not self._tfa_auth:
+            self._tfa_auth = TfaAuthenticator("ilias-anon-tfa")
+
+        tfa_token = await self._tfa_auth.password()
+
+        # Searching the form here so that this fails before asking for
+        # credentials rather than after asking.
+        form = cast(Tag, soup.find("form", {"method": "post"}))
+        action = cast(str, form["action"])
+
+        # Equivalent: Enter token in
+        # https://idp.scc.kit.edu/idp/profile/SAML2/Redirect/SSO
+        url = str(shib_url.origin()) + action
+        username, password = await self._auth.credentials()
+        data = {
+            "_eventId_proceed": "",
+            "fudis_otp_input": tfa_token,
+        }
+        if csrf_token_input := form.find("input", {"name": "csrf_token"}):
+            data["csrf_token"] = csrf_token_input["value"]  # type: ignore
+        return await _post(session, url, data)
+
+    @staticmethod
+    def _login_successful(soup: BeautifulSoup) -> bool:
+        relay_state = soup.find("input", {"name": "RelayState"})
+        saml_response = soup.find("input", {"name": "SAMLResponse"})
+        return relay_state is not None and saml_response is not None
+
+    @staticmethod
+    def _tfa_required(soup: BeautifulSoup) -> bool:
+        return soup.find(id="fudiscr-form") is not None
+
+
+async def _post(session: aiohttp.ClientSession, url: str, data: Any) -> BeautifulSoup:
+    async with session.post(url, data=data) as response:
+        return soupify(await response.read())

PFERD/crawl/kit_ipd_crawler.py

@@ -1,8 +1,9 @@
 import os
 import re
 from dataclasses import dataclass
+from datetime import datetime
 from pathlib import PurePath
-from typing import Awaitable, List, Optional, Pattern, Set, Union
+from typing import Any, Awaitable, Generator, Iterable, List, Optional, Pattern, Tuple, Union, cast
 from urllib.parse import urljoin
 
 from bs4 import BeautifulSoup, Tag
@@ -31,24 +32,24 @@ class KitIpdCrawlerSection(HttpCrawlerSection):
         return re.compile(regex)
 
 
-@dataclass(unsafe_hash=True)
+@dataclass
 class KitIpdFile:
     name: str
     url: str
 
+    def explain(self) -> None:
+        log.explain(f"File {self.name!r} (href={self.url!r})")
+
 
 @dataclass
 class KitIpdFolder:
     name: str
-    files: List[KitIpdFile]
+    entries: List[Union[KitIpdFile, "KitIpdFolder"]]
 
     def explain(self) -> None:
         log.explain_topic(f"Folder {self.name!r}")
-        for file in self.files:
+        for entry in self.entries:
-            log.explain(f"File {file.name!r} (href={file.url!r})")
+            entry.explain()
-
-    def __hash__(self) -> int:
-        return self.name.__hash__()
 
 
 class KitIpdCrawler(HttpCrawler):
@@ -72,81 +73,96 @@ class KitIpdCrawler(HttpCrawler):
 
         async with maybe_cl:
             for item in await self._fetch_items():
+                item.explain()
                 if isinstance(item, KitIpdFolder):
-                    tasks.append(self._crawl_folder(item))
+                    tasks.append(self._crawl_folder(PurePath("."), item))
                 else:
-                    # Orphan files are placed in the root folder
+                    log.explain_topic(f"Orphan file {item.name!r} (href={item.url!r})")
-                    tasks.append(self._download_file(PurePath("."), item))
+                    log.explain("Attributing it to root folder")
+                    # do this here to at least be sequential and not parallel (rate limiting is hard, as the
+                    # crawl abstraction does not hold for these requests)
+                    etag, mtime = await self._request_resource_version(item.url)
+                    tasks.append(self._download_file(PurePath("."), item, etag, mtime))
 
         await self.gather(tasks)
 
-    async def _crawl_folder(self, folder: KitIpdFolder) -> None:
+    async def _crawl_folder(self, parent: PurePath, folder: KitIpdFolder) -> None:
-        path = PurePath(folder.name)
+        path = parent / folder.name
         if not await self.crawl(path):
             return
 
-        tasks = [self._download_file(path, file) for file in folder.files]
+        tasks = []
+        for entry in folder.entries:
+            if isinstance(entry, KitIpdFolder):
+                tasks.append(self._crawl_folder(path, entry))
+            else:
+                # do this here to at least be sequential and not parallel (rate limiting is hard, as the crawl
+                # abstraction does not hold for these requests)
+                etag, mtime = await self._request_resource_version(entry.url)
+                tasks.append(self._download_file(path, entry, etag, mtime))
 
         await self.gather(tasks)
 
-    async def _download_file(self, parent: PurePath, file: KitIpdFile) -> None:
+    async def _download_file(
+        self,
+        parent: PurePath,
+        file: KitIpdFile,
+        etag: Optional[str],
+        mtime: Optional[datetime]
+    ) -> None:
         element_path = parent / file.name
-        maybe_dl = await self.download(element_path)
+
+        prev_etag = self._get_previous_etag_from_report(element_path)
+        etag_differs = None if prev_etag is None else prev_etag != etag
+
+        maybe_dl = await self.download(element_path, etag_differs=etag_differs, mtime=mtime)
         if not maybe_dl:
+            # keep storing the known file's etag
+            if prev_etag:
+                self._add_etag_to_report(element_path, prev_etag)
             return
 
         async with maybe_dl as (bar, sink):
-            await self._stream_from_url(file.url, sink, bar)
+            await self._stream_from_url(file.url, element_path, sink, bar)
 
-    async def _fetch_items(self) -> Set[Union[KitIpdFile, KitIpdFolder]]:
+    async def _fetch_items(self) -> Iterable[Union[KitIpdFile, KitIpdFolder]]:
-        page = await self.get_page()
+        page, url = await self.get_page()
         elements: List[Tag] = self._find_file_links(page)
-        items: Set[Union[KitIpdFile, KitIpdFolder]] = set()
 
+        # do not add unnecessary nesting for a single <h1> heading
+        drop_h1: bool = len(page.find_all(name="h1")) <= 1
+
+        folder_tree: KitIpdFolder = KitIpdFolder(".", [])
         for element in elements:
-            folder_label = self._find_folder_label(element)
+            parent = HttpCrawler.get_folder_structure_from_heading_hierarchy(element, drop_h1)
-            if folder_label:
+            file = self._extract_file(element, url)
-                folder = self._extract_folder(folder_label)
-                if folder not in items:
-                    items.add(folder)
-                    folder.explain()
-            else:
-                file = self._extract_file(element)
-                items.add(file)
-                log.explain_topic(f"Orphan file {file.name!r} (href={file.url!r})")
-                log.explain("Attributing it to root folder")
 
-        return items
+            current_folder: KitIpdFolder = folder_tree
+            for folder_name in parent.parts:
+                # helps the type checker to verify that current_folder is indeed a folder
+                def subfolders() -> Generator[KitIpdFolder, Any, None]:
+                    return (entry for entry in current_folder.entries if isinstance(entry, KitIpdFolder))
 
-    def _extract_folder(self, folder_tag: Tag) -> KitIpdFolder:
+                if not any(entry.name == folder_name for entry in subfolders()):
-        files: List[KitIpdFile] = []
+                    current_folder.entries.append(KitIpdFolder(folder_name, []))
-        name = folder_tag.getText().strip()
+                current_folder = next(entry for entry in subfolders() if entry.name == folder_name)
 
-        container: Tag = folder_tag.findNextSibling(name="table")
+            current_folder.entries.append(file)
-        for link in self._find_file_links(container):
-            files.append(self._extract_file(link))
 
-        return KitIpdFolder(name, files)
+        return folder_tree.entries
 
-    @staticmethod
+    def _extract_file(self, link: Tag, url: str) -> KitIpdFile:
-    def _find_folder_label(file_link: Tag) -> Optional[Tag]:
+        url = self._abs_url_from_link(url, link)
-        enclosing_table: Tag = file_link.findParent(name="table")
-        if enclosing_table is None:
-            return None
-        return enclosing_table.findPreviousSibling(name=re.compile("^h[1-6]$"))
-
-    def _extract_file(self, link: Tag) -> KitIpdFile:
-        url = self._abs_url_from_link(link)
         name = os.path.basename(url)
         return KitIpdFile(name, url)
 
-    def _find_file_links(self, tag: Union[Tag, BeautifulSoup]) -> List[Tag]:
+    def _find_file_links(self, tag: Union[Tag, BeautifulSoup]) -> list[Tag]:
-        return tag.findAll(name="a", attrs={"href": self._file_regex})
+        return cast(list[Tag], tag.find_all(name="a", attrs={"href": self._file_regex}))
 
-    def _abs_url_from_link(self, link_tag: Tag) -> str:
+    def _abs_url_from_link(self, url: str, link_tag: Tag) -> str:
-        return urljoin(self._url, link_tag.get("href"))
+        return urljoin(url, cast(str, link_tag.get("href")))
 
-    async def _stream_from_url(self, url: str, sink: FileSink, bar: ProgressBar) -> None:
+    async def _stream_from_url(self, url: str, path: PurePath, sink: FileSink, bar: ProgressBar) -> None:
         async with self.session.get(url, allow_redirects=False) as resp:
             if resp.status == 403:
                 raise CrawlError("Received a 403. Are you within the KIT network/VPN?")
@@ -159,7 +175,9 @@ class KitIpdCrawler(HttpCrawler):
 
             sink.done()
 
-    async def get_page(self) -> BeautifulSoup:
+            self._add_etag_to_report(path, resp.headers.get("ETag"))
+
+    async def get_page(self) -> Tuple[BeautifulSoup, str]:
         async with self.session.get(self._url) as request:
             # The web page for Algorithmen für Routenplanung contains some
             # weird comments that beautifulsoup doesn't parse correctly. This
@@ -167,4 +185,4 @@ class KitIpdCrawler(HttpCrawler):
             # cause issues on other pages.
             content = (await request.read()).decode("utf-8")
             content = re.sub(r"<!--.*?-->", "", content)
-            return soupify(content.encode("utf-8"))
+            return soupify(content.encode("utf-8")), str(request.url)

PFERD/deduplicator.py

@@ -14,7 +14,7 @@ def name_variants(path: PurePath) -> Iterator[PurePath]:
 
 
 class Deduplicator:
-    FORBIDDEN_CHARS = '<>:"/\\|?*'
+    FORBIDDEN_CHARS = '<>:"/\\|?*' + "".join([chr(i) for i in range(0, 32)])
     FORBIDDEN_NAMES = {
         "CON", "PRN", "AUX", "NUL",
         "COM1", "COM2", "COM3", "COM4", "COM5", "COM6", "COM7", "COM8", "COM9",

PFERD/logging.py

@@ -1,9 +1,8 @@
 import asyncio
 import sys
 import traceback
-from contextlib import asynccontextmanager, contextmanager
+from contextlib import AbstractContextManager, asynccontextmanager, contextmanager
-# TODO In Python 3.9 and above, ContextManager is deprecated
+from typing import AsyncIterator, Iterator, List, Optional
-from typing import AsyncIterator, ContextManager, Iterator, List, Optional
 
 from rich.console import Console, Group
 from rich.live import Live
@@ -59,6 +58,7 @@ class Log:
         # Whether different parts of the output are enabled or disabled
         self.output_explain = False
         self.output_status = True
+        self.output_not_deleted = True
         self.output_report = True
 
     def _update_live(self) -> None:
@@ -207,6 +207,17 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
             action = escape(f"{action:<{self.STATUS_WIDTH}}")
             self.print(f"{style}{action}[/] {escape(text)} {suffix}")
 
+    def not_deleted(self, style: str, action: str, text: str, suffix: str = "") -> None:
+        """
+        Print a message for a local only file that wasn't
+        deleted while crawling. Allows markup in the "style"
+        argument which will be applied to the "action" string.
+        """
+
+        if self.output_status and self.output_not_deleted:
+            action = escape(f"{action:<{self.STATUS_WIDTH}}")
+            self.print(f"{style}{action}[/] {escape(text)} {suffix}")
+
     def report(self, text: str) -> None:
         """
         Print a report after crawling. Allows markup.
@@ -215,6 +226,14 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
         if self.output_report:
             self.print(text)
 
+    def report_not_deleted(self, text: str) -> None:
+        """
+        Print a report for a local only file that wasn't deleted after crawling. Allows markup.
+        """
+
+        if self.output_report and self.output_not_deleted:
+            self.print(text)
+
     @contextmanager
     def _bar(
             self,
@@ -241,7 +260,7 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
             action: str,
             text: str,
             total: Optional[float] = None,
-    ) -> ContextManager[ProgressBar]:
+    ) -> AbstractContextManager[ProgressBar]:
         """
         Allows markup in the "style" argument which will be applied to the
         "action" string.
@@ -257,7 +276,7 @@ directly or as a GitHub issue: https://github.com/Garmelon/PFERD/issues/new
             action: str,
             text: str,
             total: Optional[float] = None,
-    ) -> ContextManager[ProgressBar]:
+    ) -> AbstractContextManager[ProgressBar]:
         """
         Allows markup in the "style" argument which will be applied to the
         "action" string.

PFERD/output_dir.py

@@ -44,6 +44,7 @@ class OnConflict(Enum):
     LOCAL_FIRST = "local-first"
     REMOTE_FIRST = "remote-first"
     NO_DELETE = "no-delete"
+    NO_DELETE_PROMPT_OVERWRITE = "no-delete-prompt-overwrite"
 
     @staticmethod
     def from_string(string: str) -> "OnConflict":
@@ -51,11 +52,12 @@ class OnConflict(Enum):
             return OnConflict(string)
         except ValueError:
             raise ValueError("must be one of 'prompt', 'local-first',"
-                             " 'remote-first', 'no-delete'")
+                             " 'remote-first', 'no-delete', 'no-delete-prompt-overwrite'")
 
 
 @dataclass
 class Heuristics:
+    etag_differs: Optional[bool]
     mtime: Optional[datetime]
 
 
@@ -232,8 +234,16 @@ class OutputDirectory:
 
         remote_newer = None
 
+        # ETag should be a more reliable indicator than mtime, so we check it first
+        if heuristics.etag_differs is not None:
+            remote_newer = heuristics.etag_differs
+            if remote_newer:
+                log.explain("Remote file's entity tag differs")
+            else:
+                log.explain("Remote file's entity tag is the same")
+
         # Python on Windows crashes when faced with timestamps around the unix epoch
-        if heuristics.mtime and (os.name != "nt" or heuristics.mtime.year > 1970):
+        if remote_newer is None and heuristics.mtime and (os.name != "nt" or heuristics.mtime.year > 1970):
             mtime = heuristics.mtime
             remote_newer = mtime.timestamp() > stat.st_mtime
             if remote_newer:
@@ -264,7 +274,7 @@ class OutputDirectory:
             on_conflict: OnConflict,
             path: PurePath,
     ) -> bool:
-        if on_conflict == OnConflict.PROMPT:
+        if on_conflict in {OnConflict.PROMPT, OnConflict.NO_DELETE_PROMPT_OVERWRITE}:
             async with log.exclusive_output():
                 prompt = f"Replace {fmt_path(path)} with remote file?"
                 return await prompt_yes_no(prompt, default=False)
@@ -283,7 +293,7 @@ class OutputDirectory:
             on_conflict: OnConflict,
             path: PurePath,
     ) -> bool:
-        if on_conflict == OnConflict.PROMPT:
+        if on_conflict in {OnConflict.PROMPT, OnConflict.NO_DELETE_PROMPT_OVERWRITE}:
             async with log.exclusive_output():
                 prompt = f"Recursively delete {fmt_path(path)} and replace with remote file?"
                 return await prompt_yes_no(prompt, default=False)
@@ -303,7 +313,7 @@ class OutputDirectory:
             path: PurePath,
             parent: PurePath,
     ) -> bool:
-        if on_conflict == OnConflict.PROMPT:
+        if on_conflict in {OnConflict.PROMPT, OnConflict.NO_DELETE_PROMPT_OVERWRITE}:
             async with log.exclusive_output():
                 prompt = f"Delete {fmt_path(parent)} so remote file {fmt_path(path)} can be downloaded?"
                 return await prompt_yes_no(prompt, default=False)
@@ -330,7 +340,7 @@ class OutputDirectory:
             return False
         elif on_conflict == OnConflict.REMOTE_FIRST:
             return True
-        elif on_conflict == OnConflict.NO_DELETE:
+        elif on_conflict in {OnConflict.NO_DELETE, OnConflict.NO_DELETE_PROMPT_OVERWRITE}:
             return False
 
         # This should never be reached
@@ -361,10 +371,28 @@ class OutputDirectory:
 
         raise OutputDirError("Failed to create temporary file")
 
+    def should_try_download(
+        self,
+        path: PurePath,
+        *,
+        etag_differs: Optional[bool] = None,
+        mtime: Optional[datetime] = None,
+        redownload: Optional[Redownload] = None,
+        on_conflict: Optional[OnConflict] = None,
+    ) -> bool:
+        heuristics = Heuristics(etag_differs, mtime)
+        redownload = self._redownload if redownload is None else redownload
+        on_conflict = self._on_conflict if on_conflict is None else on_conflict
+        local_path = self.resolve(path)
+
+        return self._should_download(local_path, heuristics, redownload, on_conflict)
+
     async def download(
             self,
             remote_path: PurePath,
             path: PurePath,
+            *,
+            etag_differs: Optional[bool] = None,
             mtime: Optional[datetime] = None,
             redownload: Optional[Redownload] = None,
             on_conflict: Optional[OnConflict] = None,
@@ -374,7 +402,7 @@ class OutputDirectory:
         MarkConflictError.
         """
 
-        heuristics = Heuristics(mtime)
+        heuristics = Heuristics(etag_differs, mtime)
         redownload = self._redownload if redownload is None else redownload
         on_conflict = self._on_conflict if on_conflict is None else on_conflict
         local_path = self.resolve(path)
@@ -495,7 +523,7 @@ class OutputDirectory:
             except OSError:
                 pass
         else:
-            log.status("[bold bright_magenta]", "Not deleted", fmt_path(pure))
+            log.not_deleted("[bold bright_magenta]", "Not deleted", fmt_path(pure))
             self._report.not_delete_file(pure)
 
     def load_prev_report(self) -> None:

PFERD/pferd.py

@@ -1,4 +1,4 @@
-from pathlib import Path
+from pathlib import Path, PurePath
 from typing import Dict, List, Optional
 
 from rich.markup import escape
@@ -168,19 +168,24 @@ class Pferd:
             log.report("")
             log.report(f"[bold bright_cyan]Report[/] for {escape(name)}")
 
+            def fmt_path_link(relative_path: PurePath) -> str:
+                # We need to URL-encode the path because it might contain spaces or special characters
+                link = crawler.output_dir.resolve(relative_path).absolute().as_uri()
+                return f"[link={link}]{fmt_path(relative_path)}[/link]"
+
             something_changed = False
             for path in sorted(crawler.report.added_files):
                 something_changed = True
-                log.report(f"  [bold bright_green]Added[/] {fmt_path(path)}")
+                log.report(f"  [bold bright_green]Added[/] {fmt_path_link(path)}")
             for path in sorted(crawler.report.changed_files):
                 something_changed = True
-                log.report(f"  [bold bright_yellow]Changed[/] {fmt_path(path)}")
+                log.report(f"  [bold bright_yellow]Changed[/] {fmt_path_link(path)}")
             for path in sorted(crawler.report.deleted_files):
                 something_changed = True
                 log.report(f"  [bold bright_magenta]Deleted[/] {fmt_path(path)}")
             for path in sorted(crawler.report.not_deleted_files):
                 something_changed = True
-                log.report(f"  [bold bright_magenta]Not deleted[/] {fmt_path(path)}")
+                log.report_not_deleted(f"  [bold bright_magenta]Not deleted[/] {fmt_path_link(path)}")
 
             for warning in crawler.report.encountered_warnings:
                 something_changed = True

PFERD/report.py

@@ -34,15 +34,6 @@ class MarkConflictError(Exception):
         self.collides_with = collides_with
 
 
-# TODO Use PurePath.is_relative_to when updating to 3.9
-def is_relative_to(a: PurePath, b: PurePath) -> bool:
-    try:
-        a.relative_to(b)
-        return True
-    except ValueError:
-        return False
-
-
 class Report:
     """
     A report of a synchronization. Includes all files found by the crawler, as
@@ -173,7 +164,7 @@ class Report:
             if path == other:
                 raise MarkDuplicateError(path)
 
-            if is_relative_to(path, other) or is_relative_to(other, path):
+            if path.is_relative_to(other) or other.is_relative_to(path):
                 raise MarkConflictError(path, other)
 
         self.known_files.add(path)

PFERD/transformer.py

@@ -110,6 +110,10 @@ class ExactReTf(Transformation):
             except ValueError:
                 pass
 
+        named_groups: Dict[str, str] = match.groupdict()
+        for name, capture in named_groups.items():
+            locals_dir[name] = capture
+
         result = eval(f"f{right!r}", {}, locals_dir)
         return Transformed(PurePath(result))
 

PFERD/version.py

@@ -1,2 +1,2 @@
 NAME = "PFERD"
-VERSION = "3.4.1"
+VERSION = "3.8.3"

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 3.9 installed. Run the following command to
+Ensure you have at least Python 3.11 installed. Run the following command to
 install PFERD or upgrade it to the latest version:
 
 ```
@@ -30,7 +30,10 @@ The use of [venv](https://docs.python.org/3/library/venv.html) is recommended.
 
 Unofficial packages are available for:
 - [AUR](https://aur.archlinux.org/packages/pferd)
+- [brew](https://formulae.brew.sh/formula/pferd)
+- [conda-forge](https://github.com/conda-forge/pferd-feedstock)
 - [nixpkgs](https://github.com/NixOS/nixpkgs/blob/master/pkgs/tools/misc/pferd/default.nix)
+- [PyPi](https://pypi.org/project/pferd)
 
 See also PFERD's [repology page](https://repology.org/project/pferd/versions).
 
@@ -53,6 +56,17 @@ Also, you can download most ILIAS pages directly like this:
 $ pferd kit-ilias-web <url> <output_directory>
 ```
 
+PFERD supports other ILIAS instances as well, using the `ilias-web` crawler (see
+the [config section on `ilias-web`](CONFIG.md#the-ilias-web-crawler) for more
+detail on the `base-url` and `client-id` parameters):
+
+```
+$ pferd ilias-web \
+    --base-url https://ilias.my-university.example \
+    --client-id My_University desktop \
+    <output_directory>
+```
+
 However, the CLI only lets you download a single thing at a time, and the
 resulting command can grow long quite quickly. Because of this, PFERD can also
 be used with a config file.

flake.lock

@@ -0,0 +1,27 @@
+{
+  "nodes": {
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1744440957,
+        "narHash": "sha256-FHlSkNqFmPxPJvy+6fNLaNeWnF1lZSgqVCl/eWaJRc4=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "26d499fc9f1d567283d5d56fcf367edd815dba1d",
+        "type": "github"
+      },
+      "original": {
+        "owner": "NixOS",
+        "ref": "nixos-24.11",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "root": {
+      "inputs": {
+        "nixpkgs": "nixpkgs"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}

flake.nix

@@ -0,0 +1,41 @@
+{
+  description = "Tool for downloading course-related files from ILIAS";
+
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-24.11";
+  };
+
+  outputs = { self, nixpkgs }:
+    let
+      # Helper function to generate an attrset '{ x86_64-linux = f "x86_64-linux"; ... }'.
+      forAllSystems = nixpkgs.lib.genAttrs nixpkgs.lib.systems.flakeExposed;
+    in
+    {
+      packages = forAllSystems (system:
+        let pkgs = import nixpkgs { inherit system; };
+        in
+        rec {
+          default = pkgs.python3Packages.buildPythonApplication rec {
+            pname = "pferd";
+            # Performing black magic
+            # Don't worry, I sacrificed enough goats for the next few years
+            version = (pkgs.lib.importTOML ./PFERD/version.py).VERSION;
+            format = "pyproject";
+
+            src = ./.;
+
+            nativeBuildInputs = with pkgs.python3Packages; [
+              setuptools
+            ];
+
+            propagatedBuildInputs = with pkgs.python3Packages; [
+              aiohttp
+              beautifulsoup4
+              rich
+              keyring
+              certifi
+            ];
+          };
+        });
+    };
+}

mypy.ini

@@ -1,11 +0,0 @@
-[mypy]
-disallow_any_generics = True
-disallow_untyped_defs = True
-disallow_incomplete_defs = True
-no_implicit_optional = True
-warn_unused_ignores = True
-warn_unreachable = True
-show_error_context = True
-
-[mypy-rich.*,bs4,keyring]
-ignore_missing_imports = True

pyproject.toml

@@ -1,3 +1,42 @@
 [build-system]
 requires = ["setuptools", "wheel"]
 build-backend = "setuptools.build_meta"
+
+[project]
+name = "PFERD"
+dependencies = [
+  "aiohttp>=3.8.1",
+  "beautifulsoup4>=4.10.0",
+  "rich>=11.0.0",
+  "keyring>=23.5.0",
+  "certifi>=2021.10.8"
+]
+dynamic = ["version"]
+requires-python = ">=3.11"
+
+[project.scripts]
+pferd = "PFERD.__main__:main"
+
+[tool.setuptools.dynamic]
+version = {attr = "PFERD.version.VERSION"}
+
+[tool.flake8]
+max-line-length = 110
+
+[tool.isort]
+line_length = 110
+
+[tool.autopep8]
+max_line_length = 110
+in-place = true
+recursive = true
+
+[tool.mypy]
+disallow_any_generics = true
+disallow_untyped_defs = true
+disallow_incomplete_defs = true
+no_implicit_optional = true
+warn_unused_ignores = true
+warn_unreachable = true
+show_error_context = true
+ignore_missing_imports = true

scripts/bump-version

@@ -1,8 +1,8 @@
 #!/usr/bin/env python3
 
 import argparse
-import time
 import re
+import time
 from subprocess import run
 
 

scripts/check

@@ -2,5 +2,5 @@
 
 set -e
 
-mypy PFERD
+mypy .
 flake8 PFERD

scripts/format

@@ -2,5 +2,5 @@
 
 set -e
 
-autopep8 --recursive --in-place PFERD
+autopep8 .
-isort PFERD
+isort .

scripts/setup

@@ -13,5 +13,5 @@ pip install --upgrade setuptools
 pip install --editable .
 
 # Installing tools and type hints
-pip install --upgrade mypy flake8 autopep8 isort pyinstaller
+pip install --upgrade mypy flake8 flake8-pyproject autopep8 isort pyinstaller
 pip install --upgrade types-chardet types-certifi

setup.cfg

@@ -1,23 +0,0 @@
-[metadata]
-name = PFERD
-version = attr: PFERD.version.VERSION
-
-[options]
-packages = find:
-python_requires = >=3.9
-install_requires =
-  aiohttp>=3.8.1
-  beautifulsoup4>=4.10.0
-  rich>=11.0.0
-  keyring>=23.5.0
-  certifi>=2021.10.8
-
-[options.entry_points]
-console_scripts =
-  pferd = PFERD.__main__:main
-
-[flake8]
-max_line_length = 110
-
-[isort]
-line_length = 110