DealDocumentScreening/src/contract_check/core/config.py

"""Typed configuration via pydantic-settings (12-factor).

Base `Settings` carries everything shared across services. A service imports
this and reads only the fields it needs. All values come from environment
variables (or `.env`); nothing is hardcoded.

Per-service entrypoints may subclass `Settings` to add their own fields
(e.g. the bot adds `BOT_TOKEN`, `API_URL`). See docs/ARCHITECTURE.md §11 for the
full env reference.
"""

from __future__ import annotations

from functools import lru_cache
from typing import Literal

from pydantic import Field
from pydantic_settings import BaseSettings, SettingsConfigDict

Env = Literal["dev", "staging", "prod"]
RefundPolicy = Literal["all", "infra_only"]


class Settings(BaseSettings):
    """Application configuration. See docs/ARCHITECTURE.md §11."""

    model_config = SettingsConfigDict(
        env_file=".env",
        env_file_encoding="utf-8",
        extra="ignore",
        case_sensitive=False,
    )

    # --- runtime ---
    env: Env = "dev"
    log_level: str = "INFO"

    # --- datastores / brokers ---
    database_url: str = Field(..., description="async DSN: postgresql+asyncpg://...")
    redis_url: str = "redis://redis:6379/0"
    rabbitmq_url: str = Field(..., description="amqp://user:pass@host:5672//")

    # --- RabbitMQ tuning ---
    mq_prefetch_extract: int = 1
    mq_prefetch_analyze: int = 3
    mq_prefetch_notify: int = 5
    mq_max_attempts: int = 5
    mq_retry_base_ms: int = 2000

    # --- object storage (MinIO) ---
    s3_endpoint_url: str
    s3_access_key: str
    s3_secret_key: str
    s3_bucket: str = "contract-check-docs"
    s3_region: str = "us-east-1"
    s3_server_side_encryption: bool = False
    doc_retention_days: int = 7
    text_retention_days: int = 30

    # --- billing ---
    refund_policy: RefundPolicy = "all"

    # --- observability (empty disables) ---
    sentry_dsn: str = ""
    otel_exporter_otlp_endpoint: str = ""
    otel_service_name: str = "contract-check"

    # --- LLM provider ---
    llm_provider: str = "ollama_cloud"
    ollama_host: str = ""
    ollama_api_key: str = ""
    ollama_model: str = "qwen2.5:14b"
    ollama_fallback_model: str = "qwen2.5:7b"
    ollama_temperature: float = 0.2
    ollama_num_predict: int = 3072
    ollama_timeout: float = 120.0
    ollama_max_concurrency: int = 3

    # --- API ---
    api_host: str = "0.0.0.0"
    api_port: int = 8000
    api_metrics_port: int = 9100
    b2b_default_rate_limit_rps: int = 3

    # --- auth (JWT + Telegram identity verification) ---
    telegram_bot_token: str = Field(
        "", description="Telegram bot token; used to verify Login Widget / Mini App signatures"
    )
    jwt_secret: str = Field(..., description="HS256 secret for signing user JWTs")
    jwt_algorithm: str = "HS256"
    jwt_access_ttl_minutes: int = 24 * 60  # 24 hours default; tune per env
    jwt_refresh_ttl_days: int = 30  # refresh-token lifetime for webUI auth

    # --- webUI auth (email + password) ---
    web_auth_enabled: bool = Field(
        default=True,
        description="Toggle for /api/v1/auth/{register,login,...} email-password routes",
    )
    password_reset_ttl_minutes: int = 60  # reset-token validity window
    web_app_base_url: str = Field(
        default="http://localhost:5173",
        description="Base URL of the webUI SPA — used to build password-reset links",
    )
    # Minimum password length enforced at register / reset.
    password_min_length: int = 8

    # --- admin panel (server-rendered, mounted at /admin in the api) ---
    web_admin_enabled: bool = Field(
        default=True,
        description="Toggle for the /admin management UI (users, future subscriptions)",
    )
    admin_required_role: str = Field(
        default="admin",
        description="users.role value required to enter /admin (must match the DB CHECK constraint)",
    )
    admin_default_email: str = Field(
        default="admin@contract-check.local",
        description="Default admin panel login email (auto-created if no admin exists)",
    )
    admin_default_password: str = Field(
        default="",
        description="Default admin panel password. Empty disables auto-admin creation.",
    )

    # --- SMTP (notification transport; empty host disables sending) ---
    smtp_host: str = ""
    smtp_port: int = 587
    smtp_username: str = ""
    smtp_password: str = ""
    smtp_from: str = Field(
        default="no-reply@contract-check.local",
        description="From: address used by the notify worker",
    )
    smtp_use_tls: bool = True  # STARTTLS on port 587; set false for plain SMTP

    # --- logging ---
    log_format: str = "json"  # json | console

    # --- analysis ---
    chunk_size_chars: int = 10000

    @property
    def json_logs(self) -> bool:
        """JSON logs in staging/prod, pretty console in dev.

        Override with LOG_FORMAT=json|console.
        """
        fmt = self.log_format.lower()
        if fmt == "console":
            return False
        if fmt == "json":
            return True
        return self.env != "dev"

    @property
    def log_format_value(self) -> str:
        """Resolved log format: json in prod/staging unless explicitly console."""
        fmt = self.log_format.lower()
        if fmt in ("json", "console"):
            return fmt
        return "json" if self.env != "dev" else "console"


@lru_cache
def get_settings() -> Settings:
    """Cached settings singleton. Call `get_settings.cache_clear()` to reset."""
    return Settings()  # type: ignore[call-arg]