fix and improve the identification of ARD Mediethek identifiers

[h4b4n3r0]

This fixes #4478

Fix show ID extraction for ARD-Mediathek URLs with trailing numbers

Problem

The current code in ARDMediathekBridge.php assumes the show ID is always the last segment of the URL:

$pathComponents = explode('/', $this->getInput('path'));
$showID = $pathComponents[last_non_empty_segment];

However, some ARD-Mediathek URLs append a numeric segment (e.g., /1) after the actual show ID, which relates to the selected season.

Example URL:

https://www.ardmediathek.de/serie/tage-die-es-nicht-gab-oder-staffel-2/staffel-1/Y3JpZDovL2Rhc2Vyc3RlLmRlL3RhZ2UtZGllLWVzLW5pY2h0LWdhYg/1

explode('/') produces:

[
  "https:",
  "",
  "www.ardmediathek.de",
  "serie",
  "tage-die-es-nicht-gab-oder-staffel-2",
  "staffel-1",
  "Y3JpZDovL2Rhc2Vyc3RlLmRlL3RhZ2UtZGllLWVzLW5pY2h0LWdhYg",
  "1"
]

The current logic incorrectly picks the last segment 1 instead of the actual ID:

Y3JpZDovL2Rhc2Vyc3RlLmRlL3RhZ2UtZGllLWVzLW5pY2h0LWdhYg

---

Solution: Regex-based extraction

The ARD IDs are always base64 encoded strings, with the prefix 'crid://'

Thus, ARD IDs always start with the prefix Y3JpZDov, a regex can directly extract them.

preg_match('~(Y3JpZDov[^/]+)~', $this->getInput('path'), $matches);

if (empty($matches[1])) {
    throwClientException('Could not extract show ID');
}

$showID = $matches[1];

• Works with full URLs and plain IDs
• More robust and future-proof

---

Impact

• Fixes broken feeds caused by extra numeric URL segments
• Works for all ARD-Mediathek show URLs
• Prevents Could not extract show ID errors

[h4b4n3r0]

This was manually tested. It is ready to review.

[github-actions]

Pull request artifacts

Bridge	Context	Status
ARDMediathek	1 untitled (current)	❌ Bridge returned error 404! (20512) ❌ Type: HttpException ❌ Message: https://api.ardmediathek.de/page-gateway/widgets/ard/asset/None?pageSize=29 resulted in 404 Not Found
ARDMediathek	1 untitled (pr)	❌ Bridge returned error 400! (20512) ❌ Type: ClientException ❌ Message: Could not extract show ID

last change: Saturday 2026-02-28 16:15:11

[Bockiii]

The pr test failed and couldnt extract the ID. the test always uses the defaults for the bridge, in this case "https://www.ardmediathek.de/sendung/45-min/Y3JpZDovL25kci5kZS8xMzkx"

Can you manually check if it works for you with that show?

[h4b4n3r0]

Dear @Bockiii, thanks for the response. Strangely code is working perfectly on my end.

docker build -t rss-bridge .
docker create --name rss-bridge --publish 3000:80 --volume ./config:/config rss-bridg
docker start rss-bridge

For improved testing I also added var_dumps above and below which can be seen in my screenshot:

        $path = $this->getInput('path');
        var_dump($path);   
        preg_match('~(Y3JpZDov[^/]+)~', $this->getInput('path'), $matches);
        var_dump($matches);

Not sure, what is the problem here.

[h4b4n3r0]

I think I found the issue. I assume its in the testing end:

It says path=None on calling the endpoint.