tool-call: fix llama 3.x and functionary 3.2, play nice w/ pydantic_ai package, update readme

[ochafik]

Makes more models to play nice w/ pydantic agent (follow up from #9639)

• Always return tool call id (default = empty) even if the model / template doesn't support it
• Allow missing tool call name for backfill of templates that don't support tool calls (sync Don't require tool name in backfill behaviour google/minja#36)
• Fix parsing of non-tool-calling outputs for Functionary v3.2
• Make Llama 3.x a bit more compliant by adding more grammar triggers (tested 1B & 8B, still not perfect)
• Updated examples/server/README
• Log Chat format: even w/o --verbose

# Was already working (tool call id was generated):
llama-server --jinja -fa -hf bartowski/Mistral-Nemo-Instruct-2407-GGUF:Q6_K_L

# Newly working:

## Native format
llama-server --jinja -fa -hf bartowski/Qwen2.5-7B-Instruct-GGUF:Q4_K_M
llama-server --jinja -fa -hf bartowski/Mistral-Nemo-Instruct-2407-GGUF:Q4_K_M
llama-server --jinja -fa -hf bartowski/Llama-3.2-3B-Instruct-GGUF:Q6_K
llama-server --jinja -fa -hf bartowski/functionary-small-v3.2-GGUF:Q4_K_M
llama-server --jinja -fa -hf bartowski/Hermes-3-Llama-3.1-8B-GGUF:Q4_K_M \
  --chat-template-file <( python scripts/get_chat_template.py NousResearch/Hermes-3-Llama-3.1-8B tool_use )
llama-server --jinja -fa -hf bartowski/firefunction-v2-GGUF -hff firefunction-v2-IQ1_M.gguf \
  --chat-template-file <( python scripts/get_chat_template.py fireworks-ai/firellama-3-firefunction-v2 )
## Generic format:
llama-server --jinja -fa -hf bartowski/Phi-3.5-mini-instruct-GGUF:Q4_K_M

Example test code from @brucepro (ref):

import random
from pydantic_ai import Agent, RunContext
from pydantic_ai.models.openai import OpenAIModel

model = OpenAIModel(
    'llama3.3-70B',
    base_url='http://127.0.0.1:8082',
    api_key='123',
)

agent = Agent(
    model,
    deps_type=str,
    system_prompt=(
        "You're a dice game, you should roll the die and see if the number "
        "you get back matches the user's guess. If so, tell them they're a winner. "
        "Use the player's name in the response."
    ),
)

@agent.tool_plain
def roll_die() -> str:
    """Roll a six-sided die and return the result."""
    return str(random.randint(1, 6))

@agent.tool
def get_player_name(ctx: RunContext[str]) -> str:
    """Get the player's name."""
    return ctx.deps

dice_result = agent.run_sync('My guess is 4', deps='Anne')
print(dice_result.data)

[brucepro]

I have added a pr to start the discussion of how MCP should be added to the client. Please share your thoughts if you don't mind. #11853