From 723a82f61786735f236baaa75c70d045b1bd6c8f Mon Sep 17 00:00:00 2001
From: Philipp Fruck <dev@p-fruck.de>
Date: Thu, 23 Mar 2023 17:09:29 +0100
Subject: [PATCH] improve compatibility with other ilias instances

Running against a custom Ilias instance files, folders and forms were
not detected by the current implementation since the given Ilias
instance uses URL paths instead of a target parameter
---
 PFERD/crawl/ilias/kit_ilias_html.py | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/PFERD/crawl/ilias/kit_ilias_html.py b/PFERD/crawl/ilias/kit_ilias_html.py
index c0ebdc9..c4a0b46 100644
--- a/PFERD/crawl/ilias/kit_ilias_html.py
+++ b/PFERD/crawl/ilias/kit_ilias_html.py
@@ -43,6 +43,8 @@ class IliasPageElement:
         regexes = [
             r"eid=(?P<id>[0-9a-z\-]+)",
             r"file_(?P<id>\d+)",
+            r"fold_(?P<id>\d+)",
+            r"frm_(?P<id>\d+)",
             r"ref_id=(?P<id>\d+)",
             r"target=[a-z]+_(?P<id>\d+)"
         ]
@@ -773,6 +775,16 @@ class IliasPage:
         if "cmdClass=ilobjtestgui" in parsed_url.query:
             return IliasElementType.TEST
 
+        # other universities might have content type specified in URL path
+        if "_file_" in parsed_url.path:
+            return IliasElementType.FILE
+
+        if "_fold_" in parsed_url.path:
+            return IliasElementType.FOLDER
+
+        if "_frm_" in parsed_url.path:
+            return IliasElementType.FORUM
+
         # Booking and Meeting can not be detected based on the link. They do have a ref_id though, so
         # try to guess it from the image.