"""Prompt generation for LLM consultation.

This module generates prompts suitable for consulting LLMs (ChatGPT, Claude, etc.)
about appropriate CLI options for datamodel-code-generator.
"""

from __future__ import annotations

from collections import defaultdict
from typing import TYPE_CHECKING

from datamodel_code_generator.cli_options import CLI_OPTION_META, OptionCategory
from datamodel_code_generator.prompt_data import OPTION_DESCRIPTIONS

if TYPE_CHECKING:
    from argparse import Namespace

# Options to exclude from prompt output
PROMPT_EXCLUDED_OPTIONS: frozenset[str] = frozenset({
    "generate_prompt",
    "generate_pyproject_config",
    "generate_cli_command",
    "version",
    "help",
    "no_color",
})


def _format_current_options(args: Namespace) -> str:
    """Format currently set CLI options."""
    lines: list[str] = []
    args_dict = vars(args)

    for key, value in sorted(args_dict.items()):
        if value is None:
            continue
        if key in PROMPT_EXCLUDED_OPTIONS:
            continue

        cli_key = key.replace("_", "-")

        if isinstance(value, bool):
            if value:
                lines.append(f"--{cli_key}")
        elif isinstance(value, list):
            formatted = " ".join(str(v) for v in value)
            lines.append(f"--{cli_key} {formatted}")
        else:
            lines.append(f"--{cli_key} {value}")

    return "\n".join(lines) if lines else "(No options specified)"


def _format_options_by_category() -> str:
    """Format options grouped by category with descriptions."""
    by_category: dict[OptionCategory, list[tuple[str, str]]] = defaultdict(list)

    for option, meta in CLI_OPTION_META.items():
        desc = OPTION_DESCRIPTIONS.get(option, "")
        by_category[meta.category].append((option, desc))

    lines: list[str] = []
    for category in OptionCategory:
        options = by_category[category]
        lines.append(f"### {category.value}")
        for opt, desc in sorted(options):
            if desc:
                lines.append(f"- `{opt}`: {desc}")
            else:  # pragma: no cover
                lines.append(f"- `{opt}`")
        lines.append("")

    return "\n".join(lines)


def generate_prompt(args: Namespace, help_text: str) -> str:
    """Generate LLM consultation prompt.

    Args:
        args: Parsed command-line arguments.
        help_text: Full help text from arg_parser.format_help().

    Returns:
        Formatted prompt string suitable for LLM consultation.
    """
    lines: list[str] = []

    # Header
    lines.extend(("# datamodel-code-generator CLI Options Consultation", ""))

    # User's question (if provided)
    question = getattr(args, "generate_prompt", "") or ""
    if question:
        lines.extend(("## Your Question", "", question, ""))

    # Current configuration
    lines.extend((
        "## Current CLI Options",
        "",
        "```",
        _format_current_options(args),
        "```",
        "",
        # Options by category with descriptions
        "## Options by Category",
        "",
        _format_options_by_category(),
        # Full help text
        "## All Available Options (Full Help)",
        "",
        "```",
        help_text,
        "```",
        "",
        # Instructions for LLM
        "## Instructions",
        "",
        "Based on the above information, please help with the question or suggest",
        "appropriate CLI options for the use case. Consider:",
        "",
        "1. The current options already set",
        "2. Option descriptions and their purposes",
        "3. Potential conflicts between options",
        "4. Best practices for the target output format",
        "",
    ))

    return "\n".join(lines)