Merge remote-tracking branch 'origin/main' into 1983-fix-la-blazor

# Conflicts:
#	CHANGES.md

Author: quevon24

CHANGES.md

@@ -26,6 +26,10 @@ Fixes:
   opinions, and the legacy http:// host now returns 521. The scraper now reads
   the RSS feed for recent Opinions sub-pages and drives the Blazor SignalR
   circuit to retrieve the server-rendered news release. #1983
+- `ca9` oral argument audio URLs now point to `cdn.ca9.uscourts.gov`; the old
+  `www` host started returning 404 pages after the site redesign (#1987)
+- `uscfc` was crashing ingestion because some opinions (attorney discipline
+  orders) now have relative download links. Resolve them with `urljoin` #1986
 
 ## 3.0.21 - 2026-05-29
 

juriscraper/opinions/united_states/federal_special/uscfc.py

@@ -9,6 +9,7 @@
 
 import json
 import re
+from urllib.parse import urljoin
 
 from juriscraper.lib.string_utils import titlecase
 from juriscraper.OpinionSiteLinear import OpinionSiteLinear
@@ -68,7 +69,9 @@ def _process_html(self):
                 status = "Unknown"
 
             parsed_case = {
-                "url": opinion["link"],
+                # Some opinions have relative links,
+                # e.g. "/cofc/opinions/25-11281-parks-....pdf"
+                "url": urljoin(self.url, opinion["link"]),
                 "date": opinion["date"],
                 "other_date": other_date,
                 "status": status,

juriscraper/oral_args/united_states/federal_appellate/ca9.py

@@ -1,6 +1,10 @@
 """Scraper for Ninth Circuit of Appeals
 CourtID: ca9
 Court Short Name: ca9
+History:
+    - 2026-06-04: Audio files moved from www.ca9.uscourts.gov to
+      cdn.ca9.uscourts.gov after the site redesign; same migration that
+      moved the opinion feeds (#1987).
 """
 
 import json
@@ -22,9 +26,15 @@ def __init__(self, *args, **kwargs):
 
         self.court_id = self.__module__
         self.table = "media"
-        self.base_url = "https://www.ca9.uscourts.gov/datastore/media/"
+        self.base_url = "https://cdn.ca9.uscourts.gov/datastore/media/"
+        # Recent files are served as "application/octet-stream; charset=UTF-8";
+        # older files (relevant for backscrapes) as "binary/octet-stream",
+        # "audio/mpeg" or "audio/x-ms-wma" depending on the year
         self.expected_content_types = [
-            "application/octet-stream; charset=UTF-8"
+            "application/octet-stream; charset=UTF-8",
+            "binary/octet-stream",
+            "audio/mpeg",
+            "audio/x-ms-wma",
         ]
         # AWS Cognito creds step:
         self.headers = {