Fix crawling of exercises with instructions

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

CHANGELOG.md

@@ -22,6 +22,9 @@ ambiguous situations.
 
 ## Unreleased
 
+## Fixed
+- Crawling of exercises with instructions
+
 ## 3.8.2 - 2025-04-29
 
 ## Changed

PFERD/cli/command_ilias_web.py

@@ -45,8 +45,8 @@ def load(
     load_crawler(args, section)
 
     section["type"] = COMMAND_NAME
-    if args.base_url is not None:
+    if args.ilias_url is not None:
-        section["base_url"] = args.base_url
+        section["base_url"] = args.ilias_url
     if args.client_id is not None:
         section["client_id"] = args.client_id
 

PFERD/crawl/ilias/ilias_web_crawler.py

@@ -983,8 +983,6 @@ instance's greatest bottleneck.
             soup = IliasSoup(soupify(await request.read()), str(request.url))
             if IliasPage.is_logged_in(soup):
                 return self._verify_page(soup, url, root_page_allowed)
-            with open("/tmp/ilias_debug.html", "w") as f:
-                f.write(str(soup.soup.prettify()))
         raise CrawlError(f"get_page failed even after authenticating on {url!r}")
 
     @staticmethod

PFERD/crawl/ilias/kit_ilias_html.py

@@ -975,16 +975,17 @@ class IliasPage:
             _unexpected_html_warning()
             return []
 
-        individual_exercises = content_tab.find_all(
+        exercise_links = content_tab.select(".il-item-title a")
-            name="a",
+
-            attrs={
+        for exercise in cast(list[Tag], exercise_links):
-                "href": lambda x: x is not None
+            if "href" not in exercise.attrs:
-                and "ass_id=" in x
+                continue
-                and "cmdClass=ilAssignmentPresentationGUI" in x
+            href = exercise.attrs["href"]
-            }
+            if type(href) is not str:
-        )
+                continue
+            if "ass_id=" not in href or "cmdclass=ilassignmentpresentationgui" not in href.lower():
+                continue
 
-        for exercise in cast(list[Tag], individual_exercises):
             name = _sanitize_path_name(exercise.get_text().strip())
             results.append(IliasPageElement.create_new(
                 IliasElementType.EXERCISE,