Merge 77c1f1516c into 533bc27439

CHANGELOG.md

@@ -22,9 +22,6 @@ ambiguous situations.
 
 ## Unreleased
 
-### Fixed
-- Video name deduplication
-
 ## 3.5.0 - 2023-09-13
 
 ### Added

CONFIG.md

@@ -92,6 +92,9 @@ common to all crawlers:
   load for the crawl target. (Default: `0.0`)
 - `windows_paths`: Whether PFERD should find alternative names for paths that
   are invalid on Windows. (Default: `yes` on Windows, `no` otherwise)
+- `aliases`: List of strings that are considered as an alias when invoking with
+  the `--crawler` or `-C` flag. If there is more than one crawl section with
+  the same aliases all are selected. Thereby, you can group different crawlers.
 
 Some crawlers may also require credentials for authentication. To configure how
 the crawler obtains its credentials, the `auth` option is used. It is set to the
@@ -106,6 +109,7 @@ username = foo
 password = bar
 
 [crawl:something]
+aliases = [sth, some]
 type = some-complex-crawler
 auth = auth:example
 on_conflict = no-delete

PFERD/crawl/ilias/kit_ilias_html.py

@@ -1067,34 +1067,6 @@ class IliasPage:
         rest_of_name = split_delimiter.join(meeting_name.split(split_delimiter)[1:])
         return datetime.strftime(date_portion, "%Y-%m-%d") + split_delimiter + rest_of_name
 
-    @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
-
     def _abs_url_from_link(self, link_tag: Tag) -> str:
         """
         Create an absolute url from an <a> tag.

PFERD/crawl/ilias/kit_ilias_web_crawler.py

@@ -81,7 +81,7 @@ class KitIliasWebCrawlerSection(HttpCrawlerSection):
         return self.s.getboolean("forums", fallback=False)
 
 
-_DIRECTORY_PAGES: Set[IliasElementType] = {
+_DIRECTORY_PAGES: Set[IliasElementType] = set([
     IliasElementType.EXERCISE,
     IliasElementType.EXERCISE_FILES,
     IliasElementType.FOLDER,
@@ -90,16 +90,16 @@ _DIRECTORY_PAGES: Set[IliasElementType] = {
     IliasElementType.MEDIACAST_VIDEO_FOLDER,
     IliasElementType.OPENCAST_VIDEO_FOLDER,
     IliasElementType.OPENCAST_VIDEO_FOLDER_MAYBE_PAGINATED,
-}
+])
 
-_VIDEO_ELEMENTS: Set[IliasElementType] = {
+_VIDEO_ELEMENTS: Set[IliasElementType] = set([
     IliasElementType.MEDIACAST_VIDEO_FOLDER,
     IliasElementType.MEDIACAST_VIDEO,
     IliasElementType.OPENCAST_VIDEO,
     IliasElementType.OPENCAST_VIDEO_PLAYER,
     IliasElementType.OPENCAST_VIDEO_FOLDER,
     IliasElementType.OPENCAST_VIDEO_FOLDER_MAYBE_PAGINATED,
-}
+])
 
 
 def _iorepeat(attempts: int, name: str, failure_is_error: bool = False) -> Callable[[AWrapped], AWrapped]:
@@ -140,10 +140,6 @@ def _wrap_io_in_warning(name: str) -> Callable[[AWrapped], AWrapped]:
     return _iorepeat(1, name)
 
 
-def _get_video_cache_key(element: IliasPageElement) -> str:
-    return f"ilias-video-cache-{element.id()}"
-
-
 # Crawler control flow:
 #
 #     crawl_desktop -+
@@ -551,8 +547,8 @@ instance's greatest bottleneck.
         # Copy old mapping as it is likely still relevant
         if self.prev_report:
             self.report.add_custom_value(
-                _get_video_cache_key(element),
+                str(element_path),
-                self.prev_report.get_custom_value(_get_video_cache_key(element))
+                self.prev_report.get_custom_value(str(element_path))
             )
 
         # A video might contain other videos, so let's "crawl" the video first
@@ -562,69 +558,58 @@ instance's greatest bottleneck.
         # 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), we can move on
+        # If we do not want to crawl it (user filter) or we have every file
-        if not maybe_dl:
+        # from the cached mapping already, we can ignore this and bail
-            return None
+        if not maybe_dl or self._all_opencast_videos_locally_present(element_path):
-
+            # Mark all existing cideos as known so they do not get deleted
-        # If we have every file from the cached mapping already, we can ignore this and bail
+            # during dleanup. We "downloaded" them, just without actually making
-        if self._all_opencast_videos_locally_present(element, maybe_dl.path):
+            # a network request as we assumed they did not change.
-            # Mark all existing videos as known to ensure they do not get deleted during cleanup.
+            for video in self._previous_contained_opencast_videos(element_path):
-            # We "downloaded" them, just without actually making a network request as we assumed
+                await self.download(video)
-            # they did not change.
-            contained = self._previous_contained_opencast_videos(element, maybe_dl.path)
-            if len(contained) > 1:
-                # Only do this if we threw away the original dl token,
-                # to not download single-stream videos twice
-                for video in contained:
-                    await self.download(video)
 
             return None
 
-        return self._download_opencast_video(element, maybe_dl)
+        return self._download_opencast_video(element_path, element, maybe_dl)
 
-    def _previous_contained_opencast_videos(
+    def _previous_contained_opencast_videos(self, video_path: PurePath) -> List[PurePath]:
-        self, element: IliasPageElement, element_path: PurePath
-    ) -> List[PurePath]:
         if not self.prev_report:
             return []
-        custom_value = self.prev_report.get_custom_value(_get_video_cache_key(element))
+        custom_value = self.prev_report.get_custom_value(str(video_path))
         if not custom_value:
             return []
-        cached_value = cast(dict[str, Any], custom_value)
+        names = cast(List[str], custom_value)
-        if "known_paths" not in cached_value or "own_path" not in cached_value:
+        folder = video_path.parent
-            log.explain(f"'known_paths' or 'own_path' missing from cached value: {cached_value}")
+        return [PurePath(folder, name) for name in names]
-            return []
-        transformed_own_path = self._transformer.transform(element_path)
-        if cached_value["own_path"] != str(transformed_own_path):
-            log.explain(
-                f"own_path '{transformed_own_path}' does not match cached value: '{cached_value['own_path']}"
-            )
-            return []
-        return [PurePath(name) for name in cached_value["known_paths"]]
 
-    def _all_opencast_videos_locally_present(self, element: IliasPageElement, element_path: PurePath) -> bool:
+    def _all_opencast_videos_locally_present(self, video_path: PurePath) -> bool:
-        log.explain_topic(f"Checking local cache for video {fmt_path(element_path)}")
+        if contained_videos := self._previous_contained_opencast_videos(video_path):
-        if contained_videos := self._previous_contained_opencast_videos(element, element_path):
+            log.explain_topic(f"Checking local cache for video {video_path.name}")
-            log.explain(
+            all_found_locally = True
-                f"The following contained videos are known: {','.join(map(fmt_path, contained_videos))}"
+            for video in contained_videos:
-            )
+                transformed_path = self._to_local_opencast_video_path(video)
-            if all(self._output_dir.resolve(path).exists() for path in contained_videos):
+                if transformed_path:
-                log.explain("Found all known videos locally, skipping enumeration request")
+                    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!")
-        else:
-            log.explain("No local cache present")
         return False
 
+    def _to_local_opencast_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_opencast_video(self, element: IliasPageElement, dl: DownloadToken) -> None:
+    async def _download_opencast_video(
-        def add_to_report(paths: list[str]) -> None:
+        self,
-            self.report.add_custom_value(
+        original_path: PurePath,
-                _get_video_cache_key(element),
+        element: IliasPageElement,
-                {"known_paths": paths, "own_path": str(self._transformer.transform(dl.path))}
+        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()
@@ -635,25 +620,32 @@ instance's greatest bottleneck.
                 log.explain(f"Using single video mode for {element.name}")
                 stream_element = stream_elements[0]
 
+                transformed_path = self._to_local_opencast_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
-                await self._stream_from_url(stream_element.url, sink, bar, is_video=True)
+                if self._output_dir.resolve(transformed_path).exists():
-                add_to_report([str(self._transformer.transform(dl.path))])
+                    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 = dl.path.parent / stream_element.name
+            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}")
-                contained_video_paths.append(str(self._transformer.transform(maybe_dl.path)))
                 await self._stream_from_url(stream_element.url, sink, bar, is_video=True)
 
-        add_to_report(contained_video_paths)
+        self.report.add_custom_value(str(original_path), contained_video_paths)
 
     async def _handle_file(
         self,
@@ -665,8 +657,8 @@ instance's greatest bottleneck.
             return None
         return self._download_file(element, maybe_dl)
 
-    @_iorepeat(3, "downloading file")
     @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):
@@ -736,6 +728,7 @@ instance's greatest bottleneck.
                 raise CrawlWarning("Failed to extract forum data")
             if download_data.empty:
                 log.explain("Forum had no threads")
+                elements = []
                 return
             html = await self._post_authenticated(download_data.url, download_data.form_data)
             elements = parse_ilias_forum_export(soupify(html))
@@ -901,7 +894,7 @@ instance's greatest bottleneck.
         auth_id = await self._current_auth_id()
         async with self.session.get(url) as request:
             soup = soupify(await request.read())
-            if IliasPage.is_logged_in(soup):
+            if self._is_logged_in(soup):
                 return self._verify_page(soup, url, root_page_allowed)
 
         # We weren't authenticated, so try to do that
@@ -910,12 +903,11 @@ instance's greatest bottleneck.
         # Retry once after authenticating. If this fails, we will die.
         async with self.session.get(url) as request:
             soup = soupify(await request.read())
-            if IliasPage.is_logged_in(soup):
+            if self._is_logged_in(soup):
                 return self._verify_page(soup, url, root_page_allowed)
         raise CrawlError(f"get_page failed even after authenticating on {url!r}")
 
-    @staticmethod
+    def _verify_page(self, soup: BeautifulSoup, url: str, root_page_allowed: bool) -> BeautifulSoup:
-    def _verify_page(soup: BeautifulSoup, url: str, root_page_allowed: bool) -> BeautifulSoup:
         if IliasPage.is_root_page(soup) and not root_page_allowed:
             raise CrawlError(
                 "Unexpectedly encountered ILIAS root page. "
@@ -969,10 +961,38 @@ instance's greatest bottleneck.
 
     # 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)
+    @ _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:
     """
@@ -1119,7 +1139,7 @@ async def _shib_post(
         async with session.get(correct_url, allow_redirects=False) as response:
             location = response.headers.get("location")
             log.explain(f"Redirected to {location!r} with status {response.status}")
-            # If shib still has a valid session, it will directly respond to the request
+            # If shib still still has a valid session, it will directly respond to the request
             if location is None:
                 log.explain("Shib recognized us, returning its response directly")
                 return soupify(await response.read())

PFERD/pferd.py

@@ -1,5 +1,5 @@
 from pathlib import Path
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Set
 
 from rich.markup import escape
 
@@ -43,16 +43,24 @@ class Pferd:
 
         crawl_sections = [name for name, _ in config.crawl_sections()]
 
-        crawlers_to_run = []  # With crawl: prefix
+        crawlers_to_run = set()  # With crawl: prefix
         unknown_names = []  # Without crawl: prefix
 
         for name in cli_crawlers:
             section_name = f"crawl:{name}"
             if section_name in crawl_sections:
                 log.explain(f"Crawler section named {section_name!r} exists")
-                crawlers_to_run.append(section_name)
+                crawlers_to_run.add(section_name)
-            else:
+            # interprete name as alias of a crawler
-                log.explain(f"There's no crawler section named {section_name!r}")
+            alias_names = self._find_crawlers_by_alias(name, config)
+            if alias_names:
+                crawlers_to_run.update(alias_names)
+                log.explain_topic(f"Crawler alias {name!r} found corresponding crawler sections:")
+                for alias_name in alias_names:
+                    log.explain(f"Crawler section named {alias_name!r} with alias {name!r} exists")
+
+            if not section_name in crawl_sections and not alias_names:
+                log.explain(f"There's neither a crawler section named {section_name!r} nor does a crawler with alias {name!r} exist.")
                 unknown_names.append(name)
 
         if unknown_names:
@@ -65,6 +73,14 @@ class Pferd:
 
         return crawlers_to_run
 
+    def _find_crawlers_by_alias(self, alias: str, config: Config) -> Set[str]:
+        alias_names = set()
+        for (section_name, section) in config.crawl_sections():
+            section_aliases = section.get("aliases", [])
+            if alias in section_aliases:
+                alias_names.add(section_name)
+        return alias_names
+
     def _find_crawlers_to_run(
             self,
             config: Config,