models.py

from pydantic import BaseModel, Field
from typing import List, Dict, Optional, Any


class CompetitorText(BaseModel):
    id: int
    text: str

class AnalysisRequest(BaseModel):
    target_text: str                # Текст пользователя
    competitors: List[str]          # Список текстов конкурентов
    keywords: List[str]             # Список ключевых фраз (сырых)
    language: str = "en"            # en, ru, de, es, it, pl, pt
    target_title: str = ""          # Title пользователя
    competitor_titles: List[str] = Field(default_factory=list)  # Title конкурентов

class AnalysisResponse(BaseModel):
    ngram_stats: dict               # Статистика униграм/биграм
    bm25_recommendations: List[dict] # Рекомендации "добавить/убрать"
    bert_analysis: Dict[str, Any]   # Векторный анализ
    word_counts: Dict[str, Any]     # {'target': 500, 'competitors': [600, 450], 'avg': 525}
    title_analysis: Dict[str, Any] = Field(default_factory=dict)  # Анализ Title


class SemanticAnalyzeRequest(BaseModel):
    text: str
    competitors: List[str] = Field(default_factory=list)
    language: str = "ru"
    threshold: int = 50
    compression_ratio: float = 0.1


class SemanticAnalyzeResponse(BaseModel):
    target: Dict[str, Any]
    competitors: List[Dict[str, Any]]
    comparison: Dict[str, Any]


class SemanticSearchRequest(BaseModel):
    query_text: str
    text: str
    language: str = "ru"
    top_n: int = 20


class SemanticSearchResponse(BaseModel):
    results: List[Dict[str, Any]]


class UrlFetchRequest(BaseModel):
    url: str
    user_agent: str = "chrome_desktop"
    timeout_seconds: int = 15


class UrlFetchResponse(BaseModel):
    ok: bool = False
    url: str = ""
    final_url: str = ""
    status_code: int = 0
    user_agent_key: str = ""
    user_agent_value: str = ""
    title: str = ""
    text: str = ""
    error: str = ""


class UserAgentInfo(BaseModel):
    key: str
    name: str
    value: str


class UserAgentsResponse(BaseModel):
    user_agents: List[UserAgentInfo] = Field(default_factory=list)


class OptimizerRequest(BaseModel):
    target_text: str
    competitors: List[str] = Field(default_factory=list)
    keywords: List[str] = Field(default_factory=list)
    language: str = "en"
    target_title: str = ""
    competitor_titles: List[str] = Field(default_factory=list)
    # Base for highlighting what changed in this optimization run.
    # - diff_from_input: compare with `target_text` passed in this request (snapshot before optimization)
    # - diff_from_original: compare with `original_target_text` from the first snapshot in session
    diff_mode: str = "diff_from_input"  # diff_from_input | diff_from_original
    original_target_text: Optional[str] = None
    original_target_title: Optional[str] = None

    api_key: str
    api_base_url: str = "https://api.deepseek.com/v1"
    model: str = "deepseek-chat"

    max_iterations: int = 2
    candidates_per_iteration: int = 2
    temperature: float = 0.25
    optimization_mode: str = "balanced"
    phrase_strategy_mode: str = "auto"  # auto | exact_preferred | distributed_preferred | ensemble
    bert_stage_target: float = 0.70
    # Optional stage control. If empty -> default full pipeline order.
    enabled_stages: List[str] = Field(default_factory=list)  # bert|bm25|ngram|semantic|title
    # Per-stage manual goal selection and custom additions.
    # Example:
    # {
    #   "bm25": {"mode":"mixed","selected":["canadian online casino"],"custom_add":["online casinos canada"]},
    #   "bert": {"mode":"manual","selected":["best payout casinos"],"custom_add":[]}
    # }
    stage_goal_overrides: Dict[str, Dict[str, Any]] = Field(default_factory=dict)


class OptimizerResponse(BaseModel):
    ok: bool = True
    optimized_text: str = ""
    optimized_title: str = ""
    baseline_metrics: Dict[str, Any] = Field(default_factory=dict)
    final_metrics: Dict[str, Any] = Field(default_factory=dict)
    iterations: List[Dict[str, Any]] = Field(default_factory=list)
    applied_changes: int = 0
    optimization_mode: str = "balanced"
    phrase_strategy_mode: str = "auto"
    bert_stage_target: float = 0.70
    diff_mode: str = ""
    # HTML with <mark class="diff-changed"> around changed parts.
    diff_body_html: str = ""
    diff_title_html: str = ""
    # List of (type/from/to) blocks for "что именно поменять".
    diff_changes: List[Dict[str, str]] = Field(default_factory=list)
    diff_title_changes: List[Dict[str, str]] = Field(default_factory=list)
    error: str = ""
    stopped_early: bool = False
    stop_reason: str = ""


class OptimizerCancelRequest(BaseModel):
    job_id: str = Field(..., min_length=8)