[Bug] OpenAI Audio handler hardcodes filename without extension, causing "Unrecognized file format" error

[pleasantbalogun]

Description

The OpenAI Audio handler hardcodes the filename as 'audio' without a file extension when uploading to the Whisper API. OpenAI uses the filename extension to determine the audio format, so this causes an "Unrecognized file format" error.

Location

https://github.com/prism-php/prism/blob/main/src/Providers/OpenAI/Handlers/Audio.php#L56

->attach(
    'file',
    $request->input()->resource(),
    'audio',  // <-- Missing extension
    ['Content-Type' => $request->input()->mimeType()]
)

Expected Behavior
The filename should include the extension based on MIME type (e.g., audio.webm, audio.mp3), similar to how the Groq handler does it: https://github.com/prism-php/prism/blob/main/src/Providers/Groq/Handlers/Audio.php#L92-L110

Regression
This worked in v0.82 which used `SpeechToTextRequestMapper`. The mapper was removed in v0.99+ and replaced with inline code that hardcodes the filename.

Suggested Fix
Add a `generateFilename()` method (like Groq has) and use it instead of the hardcoded 'audio' string.