Expose model capabilities

[sarthak-19]

Summary

Closes #226 Expose model input/output modality capabilities at runtime.

Currently there is no ergonomic way to ask a model instance which input and output modalities it accepts. Developers had to guess from model names ("transcribe", "audio", etc.) or dig through raw SupportedOption objects — both fragile approaches.

This PR adds three first-class methods to ModelInterface and implements them in AbstractApiBasedModel:

// Returns all supported input/output modality enums
$model->getCapabilities();
/*
[
  'input'  => [ModalityEnum::text(), ModalityEnum::image()],
  'output' => [ModalityEnum::text()],
]
*/

// Simple boolean helpers
$model->supportsInput(ModalityEnum::text());   // true
$model->supportsInput(ModalityEnum::audio());  // false
$model->supportsOutput(ModalityEnum::text());  // true
$model->supportsOutput(ModalityEnum::image()); // false

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.21%. Comparing base (6317042) to head (3655406).

Additional details and impacted files

@@             Coverage Diff              @@
##              trunk     #227      +/-   ##
============================================
+ Coverage     88.12%   88.21%   +0.08%     
- Complexity     1213     1227      +14     
============================================
  Files            60       60              
  Lines          3934     3964      +30     
============================================
+ Hits           3467     3497      +30     
  Misses          467      467              

Flag	Coverage Δ
unit	88.21% <100.00%> (+0.08%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.