From f770fa37657cd9d897598f282e5df02d529da578 Mon Sep 17 00:00:00 2001
From: dgtlmoon <dgtlmoon@gmail.com>
Date: Wed, 8 Oct 2025 23:00:40 +0200
Subject: [PATCH] WIP

---
 changedetectionio/processors/magic.py         | 31 ++++++++++---------
 .../processors/text_json_diff/processor.py    |  5 ++-
 2 files changed, 20 insertions(+), 16 deletions(-)

diff --git a/changedetectionio/processors/magic.py b/changedetectionio/processors/magic.py
index e3ce066d1..8c2f919a7 100644
--- a/changedetectionio/processors/magic.py
+++ b/changedetectionio/processors/magic.py
@@ -87,21 +87,8 @@ class guess_stream_type():
         # Check for HTML patterns first - if found, override magic's text/plain
         has_html_patterns = any(p in test_content_normalized for p in HTML_PATTERNS)
 
-        if has_html_patterns or content_header == 'text/html':
-            self.is_html = True
-        # If magic says text/plain and we found no HTML patterns, trust it
-        elif magic_result == 'text/plain':
-            self.is_plaintext = True
-            logger.debug(f"Trusting magic's text/plain result (no HTML patterns detected)")
-        elif '<rss' in test_content_normalized or '<feed' in test_content_normalized:
-            self.is_rss = True
-        elif test_content_normalized.startswith('<?xml'):
-            # Generic XML that's not RSS/Atom (RSS/Atom checked above)
-            self.is_xml = True
-        elif '%pdf-1' in test_content:
-            self.is_pdf = True
-        # Check headers for types we didn't detect by content
-        elif any(s in content_header for s in RSS_XML_CONTENT_TYPES) or any(s in magic_content_header for s in RSS_XML_CONTENT_TYPES):
+        # Always trust headers first
+        if any(s in content_header for s in RSS_XML_CONTENT_TYPES) or any(s in magic_content_header for s in RSS_XML_CONTENT_TYPES):
             self.is_rss = True
         elif any(s in content_header for s in JSON_CONTENT_TYPES) or any(s in magic_content_header for s in JSON_CONTENT_TYPES):
             self.is_json = True
@@ -115,6 +102,20 @@ class guess_stream_type():
             self.is_yaml = True
         elif 'pdf' in magic_content_header:
             self.is_pdf = True
+###
+        elif has_html_patterns or content_header == 'text/html':
+            self.is_html = True
+        # If magic says text/plain and we found no HTML patterns, trust it
+        elif magic_result == 'text/plain':
+            self.is_plaintext = True
+            logger.debug(f"Trusting magic's text/plain result (no HTML patterns detected)")
+        elif '<rss' in test_content_normalized or '<feed' in test_content_normalized:
+            self.is_rss = True
+        elif test_content_normalized.startswith('<?xml'):
+            # Generic XML that's not RSS/Atom (RSS/Atom checked above)
+            self.is_xml = True
+        elif '%pdf-1' in test_content:
+            self.is_pdf = True
         # Only trust magic for 'text' if no other patterns matched
         elif 'text' in magic_content_header:
             self.is_plaintext = True
diff --git a/changedetectionio/processors/text_json_diff/processor.py b/changedetectionio/processors/text_json_diff/processor.py
index 9a82b62be..7690da4c7 100644
--- a/changedetectionio/processors/text_json_diff/processor.py
+++ b/changedetectionio/processors/text_json_diff/processor.py
@@ -143,9 +143,12 @@ class perform_site_check(difference_detection_processor):
                 for filter in include_filters_rule:
                     if any(prefix in filter for prefix in json_filter_prefixes):
                         stripped_text_from_html += html_tools.extract_json_as_string(content=self.fetcher.content, json_filter=filter)
+                        if stripped_text_from_html:
+                            stream_content_type.is_json = True
+                            stream_content_type.is_html = False
 
         # We have 'watch.is_source_type_url' because we should be able to use selectors on the raw HTML but return just that selected HTML
-        if stream_content_type.is_html or watch.is_source_type_url or stream_content_type.is_plaintext:
+        if stream_content_type.is_html or watch.is_source_type_url or stream_content_type.is_plaintext or stream_content_type.is_rss or stream_content_type.is_xml:
 
             # CSS Filter, extract the HTML that matches and feed that into the existing inscriptis::get_text
             self.fetcher.content = html_tools.workarounds_for_obfuscations(self.fetcher.content)