# Trouve-KA — types partagés du pipeline
# Author: Simon-Pierre Boucher
# Contact: contact@spboucher.ai

"""Modèles Pydantic partagés entre crawler, parser, classifier, indexer et API."""

from datetime import datetime
from enum import StrEnum

from pydantic import BaseModel, Field


class ErrorCode(StrEnum):
    DNS = "dns"
    TIMEOUT = "timeout"
    TLS = "tls"
    HTTP_4XX = "http_4xx"
    HTTP_5XX = "http_5xx"
    ROBOTS_DENIED = "robots_denied"
    PARSE_FAILED = "parse_failed"
    UNSUPPORTED_CONTENT = "unsupported_content"
    TOO_LARGE = "too_large"
    DUPLICATE = "duplicate"
    SPAM = "spam"
    NOT_QUEBEC = "not_quebec"
    SSRF_BLOCKED = "ssrf_blocked"
    TOO_MANY_REDIRECTS = "too_many_redirects"
    CONNECTION = "connection"
    OK = "ok"


class Outcome(StrEnum):
    INDEXED = "indexed"
    DUPLICATE = "duplicate"
    UNCHANGED = "unchanged"
    ERROR = "error"
    NOT_QUEBEC = "not_quebec"
    ROBOTS_BLOCKED = "robots_blocked"
    REDIRECT = "redirect"


class FetchResult(BaseModel):
    url: str
    final_url: str
    status_code: int | None = None
    error_code: ErrorCode = ErrorCode.OK
    content_type: str | None = None
    charset: str | None = None
    body: bytes | None = None
    etag: str | None = None
    last_modified: str | None = None
    duration_ms: int = 0
    redirect_chain: list[str] = Field(default_factory=list)


class ExtractedLink(BaseModel):
    url: str
    anchor: str = ""
    nofollow: bool = False


class ParsedPage(BaseModel):
    url: str
    canonical_url: str | None = None
    title: str = ""
    description: str = ""
    body: str = ""
    headings: list[str] = Field(default_factory=list)
    language: str | None = None
    links: list[ExtractedLink] = Field(default_factory=list)
    published_at: datetime | None = None
    modified_at: datetime | None = None
    author: str | None = None
    image_url: str | None = None
    noindex: bool = False
    nofollow_page: bool = False
    structured_hints: list[str] = Field(default_factory=list)  # textes JSON-LD/OG utiles au scoring


class QuebecSignals(BaseModel):
    score: float = 0.0
    locations: list[str] = Field(default_factory=list)
    reasons: list[str] = Field(default_factory=list)


class SearchResult(BaseModel):
    title: str
    url: str
    display_url: str
    snippet: str
    domain: str
    language: str | None = None
    quebec_score: float = 0.0
    badges: list[str] = Field(default_factory=list)
    published_at: datetime | None = None


class SearchResponse(BaseModel):
    query: str
    total: int
    took_ms: int
    page: int
    limit: int
    results: list[SearchResult]