Added support for input devices/streams and tests

[Omar-Riaz]

I created a subclass called FFMpegInputBuilder which extends the AbstractFFmpegStreamBuilder. Seeing that the AbstractFFmpegInputBuilder was designed for outputs, I re-implemented the filename property and build method to retain all previous functionality.

I also added tests to see that the inputs are built properly, modelled after #156 . I added a convenience method to parse commands separated by spaces for assertion purposes.

From the FFmpegBuilder, you simply call addInputStream(String inputStream) to build the input.

[Omar-Riaz]

Updated test to support device_number. Is it ok to include the parameter if there are no duplicates and it is set to zero?

[coveralls]

Coverage increased (+2.3%) to 84.106% when pulling 3da1165 on Omar-Riaz:master into 3819b15 on bramp:master.

[bramp]

Thanks for the PR. I'm on vacation for the next week, so I won't look at this until later next week. Feel free to ping me if I don't respond next week.

[Omar-Riaz]

Hi @bramp , did you get a chance to review?

[bramp]

Hi @Omar-Riaz, thank you for this long-standing PR! We are currently modernizing the library's builder architecture and have decided to implement support for input devices/streams using a new architectural approach based on your foundation.

We will be creating a specialized FFmpegInputBuilder to support options (like -f x11grab) applied before the input flag. We'll be closing this PR once the new implementation is merged into main. Thanks for starting this conversation!