"""
Pydantic models for read_delta MCP tool input validation.

Validates Delta Lake table path, cloud provider, and optional query/limit.
"""

import re
from typing import Literal, Optional

from pydantic import BaseModel, Field, field_validator, model_validator

Vendor = Literal["aws", "alicloud", "gcp", "local"]

# Path scheme patterns per cloud (path part must be non-empty)
SCHEMES = {
    "local": (r"^file:///", "file:///absolute/path/to/delta-table"),
    "aws": (r"^s3://", "s3://bucket-name/path/to/delta-table"),
    "alicloud": (r"^s3://", "s3://bucket-name/path/to/delta-table"),
    "gcp": (r"^gs://", "gs://bucket-name/path/to/delta-table"),
}


class ReadDeltaRequest(BaseModel):
    """
    Input model for read_delta tool.

    Validates table_path format and optional limit/query.
    """

    table_path: str = Field(
        ...,
        description=(
            "URI to the Delta table "
            "(e.g. s3://bucket/path, gs://bucket/path, file:///local/path)"
        ),
        min_length=1,
    )
    query: Optional[str] = Field(
        default=None,
        description="Optional SQL over the table; must reference the table as `__t__`.",
    )
    limit: Optional[int] = Field(
        default=None,
        ge=1,
        le=100_000,
        description="Optional row limit (1–100000). Applied if query is not provided.",
    )
    cluster_id: str = Field(
        ...,
        description="Cluster identifier.",
    )
    region: Optional[str] = Field(
        default=None,
        description="Cloud region (e.g. us-west-2)",
    )
    vendor: Vendor = Field(
        ...,
        description="Cloud vendor (aws, alicloud, gcp, local)",
    )

    @field_validator("table_path")
    @classmethod
    def validate_table_path_not_empty(cls, v: str) -> str:
        v = v.strip()
        if not v:
            raise ValueError("table_path cannot be empty")
        return v

    @model_validator(mode="after")
    def validate_table_path_scheme(self) -> "ReadDeltaRequest":
        v = self.table_path
        vendor = self.vendor
        pattern, example = SCHEMES[vendor]

        if not re.search(pattern, v):
            raise ValueError(
                f"table_path for cloud '{vendor}' must match scheme. Example: {example}"
            )
        # Ensure path part exists (e.g. s3://bucket/ok, file:///a)
        no_trailing = v.rstrip("/")
        if no_trailing.endswith(":"):
            raise ValueError("table_path must include a path (e.g. bucket and key or path)")
        if vendor != "local" and no_trailing.count("/") < 3:
            raise ValueError("table_path must include a path (e.g. bucket and key)")
        return self