from dataclasses import dataclass, field
from pathlib import Path


@dataclass
class Config:
    project_name: str = ""
    copybook_paths: list = field(default_factory=lambda: ["./copybooks"])
    dialect: str = "ibm"
    llm_model: str = "gpt-4o-mini"
    llm_timeout: int = 15
    llm_cache_dir: str = ".cache/llm"
    coverage_default: str = "boundary"
    rounding_mode: str = "TRUNCATE"
    tolerance: float = 0.01
    runner_mode: str = "native"
    spark_master: str = "local[*]"
    spark_input_format: str = "json"
    num_records: int = 1000
    branch_pass: float = 0.80
    max_llm_cost: float = 0.50

    @classmethod
    def from_toml(cls, path="aurak.toml"):
        import tomllib
        try:
            with open(path, "rb") as f:
                d = tomllib.load(f)
        except:
            return cls()
        c = cls()
        p = d.get("project", {})
        c.project_name = p.get("name", "")
        c.copybook_paths = p.get("copybook_paths", c.copybook_paths)
        c.dialect = p.get("dialect", "ibm")
        ll = d.get("llm", {})
        c.llm_model = ll.get("model", c.llm_model)
        co = d.get("coverage", {})
        c.coverage_default = co.get("default_target", "boundary")
        cp = d.get("comparison", {})
        c.rounding_mode = cp.get("rounding_mode", "TRUNCATE")
        c.tolerance = cp.get("default_tolerance", c.tolerance)
        r = d.get("runner", {})
        c.runner_mode = r.get("mode", "native")
        s = d.get("spark", {})
        c.spark_master = s.get("master", "local[*]")
        c.num_records = s.get("num_records", c.num_records)
        return c