llama-cli: add support for reasoning

[bandoti]

This change adds a "partial formatter" that processes partially collected messages (like the server streaming logic) in order to render reasoning logic prior to EOG token arrival.

In addition, the chat_add_and_format lambda has been moved to a functor, and this now calls common_chat_templates_apply directly to allow more robust template-application options.

Logic has been put in place to suppress the system/prompt tags to clean up output. I am thinking in a separate PR I can work on fixing the colorization (e.g. blue for system prompt).

Example output :

./build/bin/llama-cli.exe -m ./models/gpt-oss-20b-mxfp4.gguf -c 2048 -sys "you are a wizard" -p "please recite me a haiku about llamas" --jinja

...

== Running in interactive mode. ==
 - Press Ctrl+C to interject at any time.
 - Press Return to return control to the AI.
 - To return control without starting a new line, end your input with '/'.
 - If you want to submit another line, end your input with '\'.

you are a wizard
please recite me a haiku about llamas
The user wants a haiku about llamas. The developer instruction: "you are a wizard". That seems to be a role instruction to the assistant. The user wants a haiku. We should comply. There's no conflict. We can produce a haiku, 5-7-5 syllable structure, about llamas. Also maybe incorporate wizard? But user didn't ask wizard. The developer says "you are a wizard" - maybe we need to incorporate that in style. But the user didn't request anything about wizard. We could just produce a haiku. It's allowed. There's no policy conflict. So just produce a haiku: "Llama steps, soft hum / desert wind whispers, calm, wise / moonlit paws echo". Let's check syllable counts: Llama steps (2) soft (1) hum (1) = 4? "Llama steps, soft hum" -> Lla-ma (2), steps (1), soft (1), hum (1) = 5. Next line: "desert wind whispers, calm, wise" -> de-sert (2), wind (1), whis-pers (2), calm (1), wise (1) = 7? de-sert (2) wind (1) whis-pers (2) calm (1) wise (1) = 7. Third: "moonlit paws echo" -> moon-lit (2) paws (1) e-cho (2) = 5. Good.

We can present that. Possibly add a wizard vibe: "Wizardly eyes gleam". But we must keep structure. We could do: "Wizard llama roams" but okay.

Let's produce a haiku.
Llama steps, soft hum,
Desert wind whispers, calm, wise—
Moonlit paws echo.

>

[bandoti]

I just updated to clean up the system/prompt tags (see description changes), but I will await feedback before changing anything else! 😊

One thing I was contemplating was splitting the display block into a separate abstraction. The display could be a new type because more state was added here, it might be a good time to do refactors like this to encapsulate functionality incrementally.