api/config.py

"""
Configurações para para.AI v3.0 - VERSÃO TESTE INTERNO (SQLite)
Otimizado para desenvolvimento, testes e deploys leves sem PostgreSQL
"""
import os
from typing import Optional, List
from pydantic_settings import BaseSettings
from pydantic import Field
from functools import lru_cache
from pathlib import Path


class Settings(BaseSettings):
    """
    Configurações para ambiente de TESTE INTERNO com SQLite.

    Características:
    - SQLite em vez de PostgreSQL (sem container DB)
    - Configurações simplificadas
    - Otimizado para desenvolvimento e CI/CD
    - Build rápido (~2min vs 8min)
    """

    # ========================================================================
    # API SETTINGS
    # ========================================================================
    APP_NAME: str = "para.AI API - SQLite Test"
    APP_VERSION: str = "3.0.0-sqlite"
    APP_DESCRIPTION: str = "API de teste para análise de acórdãos (SQLite)"
    APP_ENV: str = Field(default="development", env="APP_ENV")
    DEBUG: bool = Field(default=True, env="DEBUG")

    # ========================================================================
    # SERVER SETTINGS
    # ========================================================================
    HOST: str = Field(default="0.0.0.0", env="HOST")
    PORT: int = Field(default=7860, env="PORT")
    WORKERS: int = Field(default=1, env="WORKERS")  # 1 worker para SQLite
    RELOAD: bool = Field(default=True, env="RELOAD")

    # ========================================================================
    # DATABASE SETTINGS - SQLITE
    # ========================================================================
    DATABASE_TYPE: str = Field(default="sqlite", env="DATABASE_TYPE")

    SQLITE_DB_PATH: str = Field(
        default="./data/para_ai.db",
        env="SQLITE_DB_PATH",
        description="Caminho do arquivo SQLite"
    )

    DATABASE_URL: str = Field(
        default="sqlite:///./data/para_ai.db",
        env="DATABASE_URL",
        description="SQLite connection string"
    )

    SQLITE_TIMEOUT: int = Field(default=30, env="SQLITE_TIMEOUT")
    SQLITE_CHECK_SAME_THREAD: bool = Field(
        default=False, 
        env="SQLITE_CHECK_SAME_THREAD"
    )

    # Pool settings (reduzidos)
    DB_POOL_SIZE: int = Field(default=5, env="DB_POOL_SIZE")
    DB_MAX_OVERFLOW: int = Field(default=10, env="DB_MAX_OVERFLOW")
    DB_POOL_TIMEOUT: int = Field(default=30, env="DB_POOL_TIMEOUT")
    DB_POOL_RECYCLE: int = Field(default=3600, env="DB_POOL_RECYCLE")
    SQL_ECHO: bool = Field(default=False, env="SQL_ECHO")

    # ========================================================================
    # FILES & STORAGE
    # ========================================================================
    FILES_BASE_PATH: str = Field(default="./data/files", env="FILES_BASE_PATH")
    UPLOAD_PATH: str = Field(default="./data/uploads", env="UPLOAD_PATH")
    OUTPUT_PATH: str = Field(default="./data/outputs", env="OUTPUT_PATH")
    TEMP_PATH: str = Field(default="./data/temp", env="TEMP_PATH")
    BACKUP_PATH: str = Field(default="./data/backups", env="BACKUP_PATH")
    MAX_UPLOAD_SIZE_MB: int = Field(default=100, env="MAX_UPLOAD_SIZE_MB")

    # ========================================================================
    # LLM PROVIDERS
    # ========================================================================
    GROQ_API_KEY: Optional[str] = Field(default=None, env="GROQ_API_KEY")
    OPENAI_API_KEY: Optional[str] = Field(default=None, env="OPENAI_API_KEY")
    ANTHROPIC_API_KEY: Optional[str] = Field(default=None, env="ANTHROPIC_API_KEY")

    DEFAULT_LLM_PROVIDER: str = Field(default="groq", env="DEFAULT_LLM_PROVIDER")
    DEFAULT_MODEL_TYPE: str = Field(default="openai/gpt-oss-120b", env="DEFAULT_MODEL_TYPE")

    LLM_DEFAULT_TEMPERATURE: float = Field(default=1.4, env="LLM_DEFAULT_TEMPERATURE")
    LLM_DEFAULT_MAX_TOKENS: int = Field(default=10024, env="LLM_DEFAULT_MAX_TOKENS")
    LLM_TIMEOUT_SECONDS: int = Field(default=3000, env="LLM_TIMEOUT_SECONDS")

    # ========================================================================
    # PROCESSING
    # ========================================================================
    MAX_CONCURRENT_PROCESSES: int = Field(default=4, env="MAX_CONCURRENT_PROCESSES")
    PROCESS_TIMEOUT_SECONDS: int = Field(default=3000, env="PROCESS_TIMEOUT_SECONDS")
    ENABLE_PARALLEL_PROCESSING: bool = Field(default=True, env="ENABLE_PARALLEL")
    DEFAULT_MAX_WORKERS: int = Field(default=1, env="DEFAULT_MAX_WORKERS")

    BATCH_SIZE: int = Field(default=10, env="BATCH_SIZE")
    BATCH_DELAY_MS: int = Field(default=200, env="BATCH_DELAY_MS")

    # ========================================================================
    # LOGGING
    # ========================================================================
    LOG_LEVEL: str = Field(default="DEBUG", env="LOG_LEVEL")
    LOG_FORMAT: str = "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
    LOG_FILE_ENABLED: bool = Field(default=True, env="LOG_FILE_ENABLED")
    LOG_FILE_PATH: str = Field(default="./logs", env="LOG_FILE_PATH")
    LOG_FILE_MAX_BYTES: int = Field(default=5242880, env="LOG_FILE_MAX_BYTES")
    LOG_FILE_BACKUP_COUNT: int = Field(default=3, env="LOG_FILE_BACKUP_COUNT")

    # ========================================================================
    # SECURITY
    # ========================================================================
    CORS_ORIGINS: str = Field(default="*", env="CORS_ORIGINS")
    CORS_ALLOW_CREDENTIALS: bool = True

    API_KEY_HEADER: str = Field(default="X-API-Key", env="API_KEY_HEADER")
    REQUIRE_API_KEY: bool = Field(default=False, env="REQUIRE_API_KEY")
    VALID_API_KEYS: str = Field(default="test-key-123,dev-key-456", env="VALID_API_KEYS")

    RATE_LIMIT_ENABLED: bool = Field(default=False, env="RATE_LIMIT_ENABLED")
    RATE_LIMIT_PER_MINUTE: int = Field(default=1000, env="RATE_LIMIT_PER_MINUTE")

    # ========================================================================
    # CACHE
    # ========================================================================
    ENABLE_CACHE: bool = Field(default=True, env="ENABLE_CACHE")
    CACHE_TTL_SECONDS: int = Field(default=300, env="CACHE_TTL_SECONDS")
    CACHE_BACKEND: str = Field(default="memory", env="CACHE_BACKEND")
    REDIS_URL: Optional[str] = Field(default=None, env="REDIS_URL")

    # ========================================================================
    # METRICS
    # ========================================================================
    ENABLE_METRICS: bool = Field(default=True, env="ENABLE_METRICS")
    METRICS_EXPORT_INTERVAL: int = Field(default=60, env="METRICS_EXPORT_INTERVAL")
    PROMETHEUS_ENABLED: bool = Field(default=False, env="PROMETHEUS_ENABLED")
    PROMETHEUS_PORT: int = Field(default=9090, env="PROMETHEUS_PORT")

    # ========================================================================
    # PROCESSORS
    # ========================================================================
    MIN_CONFIDENCE_THRESHOLD: float = Field(default=0.5, env="MIN_CONFIDENCE_THRESHOLD")

    # ========================================================================
    # DEVELOPMENT
    # ========================================================================
    DEV_MODE: bool = Field(default=True, env="DEV_MODE")
    MOCK_LLM_RESPONSES: bool = Field(default=False, env="MOCK_LLM_RESPONSES")
    SAVE_DEBUG_FILES: bool = Field(default=True, env="SAVE_DEBUG_FILES")

    # ========================================================================
    # ADVANCED
    # ========================================================================
    ENABLE_REQUEST_ID: bool = Field(default=True, env="ENABLE_REQUEST_ID")
    REQUEST_ID_HEADER: str = "X-Request-ID"
    ENABLE_GZIP: bool = Field(default=True, env="ENABLE_GZIP")
    GZIP_MIN_SIZE: int = Field(default=1000, env="GZIP_MIN_SIZE")
    HTTP_TIMEOUT_SECONDS: int = Field(default=60, env="HTTP_TIMEOUT_SECONDS")
    TASK_RETENTION_HOURS: int = Field(default=6, env="TASK_RETENTION_HOURS")
    AUTO_CLEANUP_ENABLED: bool = Field(default=True, env="AUTO_CLEANUP_ENABLED")

    # ========================================================================
    # SQLITE-SPECIFIC
    # ========================================================================
    SQLITE_JOURNAL_MODE: str = Field(default="WAL", env="SQLITE_JOURNAL_MODE")
    SQLITE_SYNCHRONOUS: str = Field(default="NORMAL", env="SQLITE_SYNCHRONOUS")
    SQLITE_CACHE_SIZE: int = Field(default=10000, env="SQLITE_CACHE_SIZE")
    SQLITE_AUTO_VACUUM: bool = Field(default=True, env="SQLITE_AUTO_VACUUM")

    class Config:
        env_file = ".env"
        env_file_encoding = "utf-8"
        case_sensitive = True
        extra = "ignore"

    @property
    def cors_origins_list(self) -> List[str]:
        if isinstance(self.CORS_ORIGINS, str):
            if self.CORS_ORIGINS == "*":
                return ["*"]
            return [origin.strip() for origin in self.CORS_ORIGINS.split(",") if origin.strip()]
        return [self.CORS_ORIGINS]

    @property
    def valid_api_keys_list(self) -> List[str]:
        if not self.VALID_API_KEYS:
            return []
        if isinstance(self.VALID_API_KEYS, str):
            return [key.strip() for key in self.VALID_API_KEYS.split(",") if key.strip()]
        return []

    @property
    def is_production(self) -> bool:
        return self.APP_ENV.lower() == "production"

    @property
    def is_development(self) -> bool:
        return self.APP_ENV.lower() in ["development", "dev", "test"]

    @property
    def is_sqlite(self) -> bool:
        return self.DATABASE_TYPE.lower() == "sqlite" or self.DATABASE_URL.startswith("sqlite")

    @property
    def database_url_masked(self) -> str:
        if self.is_sqlite:
            return self.DATABASE_URL
        if "@" in self.DATABASE_URL:
            parts = self.DATABASE_URL.split("@")
            return f"***@{parts[1]}"
        return self.DATABASE_URL

    @property
    def sqlite_db_file_path(self) -> Path:
        if self.DATABASE_URL.startswith("sqlite:///"):
            path_str = self.DATABASE_URL.replace("sqlite:///", "")
            return Path(path_str)
        return Path(self.SQLITE_DB_PATH)

    def get_llm_providers_status(self) -> dict:
        return {
            "groq": bool(self.GROQ_API_KEY),
            "openai": bool(self.OPENAI_API_KEY),
            "anthropic": bool(self.ANTHROPIC_API_KEY)
        }

    def get_sqlite_connection_args(self) -> dict:
        return {
            "check_same_thread": self.SQLITE_CHECK_SAME_THREAD,
            "timeout": self.SQLITE_TIMEOUT
        }

    def get_sqlite_pragma_statements(self) -> List[str]:
        return [
            f"PRAGMA journal_mode={self.SQLITE_JOURNAL_MODE}",
            f"PRAGMA synchronous={self.SQLITE_SYNCHRONOUS}",
            f"PRAGMA cache_size={self.SQLITE_CACHE_SIZE}",
            "PRAGMA foreign_keys=ON",
            f"PRAGMA auto_vacuum={'FULL' if self.SQLITE_AUTO_VACUUM else 'NONE'}"
        ]

    def validate_paths(self) -> None:
        paths = [
            self.FILES_BASE_PATH,
            self.UPLOAD_PATH,
            self.OUTPUT_PATH,
            self.TEMP_PATH,
            self.BACKUP_PATH,
            self.LOG_FILE_PATH
        ]

        for path_str in paths:
            path = Path(path_str)
            path.mkdir(parents=True, exist_ok=True)

        db_path = self.sqlite_db_file_path
        db_path.parent.mkdir(parents=True, exist_ok=True)

    def to_dict(self) -> dict:
        data = self.model_dump()

        sensitive_keys = [
            "GROQ_API_KEY",
            "OPENAI_API_KEY", 
            "ANTHROPIC_API_KEY",
            "VALID_API_KEYS",
            "REDIS_URL"
        ]

        for key in sensitive_keys:
            if key in data and data[key]:
                data[key] = "***HIDDEN***"

        data["_runtime_info"] = {
            "is_sqlite": self.is_sqlite,
            "is_development": self.is_development,
            "database_file_exists": self.sqlite_db_file_path.exists() if self.is_sqlite else None,
            "database_file_size_mb": (
                self.sqlite_db_file_path.stat().st_size / (1024**2)
                if self.is_sqlite and self.sqlite_db_file_path.exists()
                else None
            )
        }

        return data


@lru_cache()
def get_settings() -> Settings:
    settings = Settings()
    settings.validate_paths()
    return settings


def get_env(key: str, default: any = None) -> any:
    return os.getenv(key, default)


def is_production() -> bool:
    return get_settings().is_production


def is_development() -> bool:
    return get_settings().is_development


def is_sqlite() -> bool:
    return get_settings().is_sqlite


def get_database_info() -> dict:
    settings = get_settings()

    info = {
        "type": "sqlite" if settings.is_sqlite else "postgresql",
        "url_masked": settings.database_url_masked
    }

    if settings.is_sqlite:
        db_path = settings.sqlite_db_file_path
        info.update({
            "file_path": str(db_path),
            "file_exists": db_path.exists(),
            "file_size_mb": (
                db_path.stat().st_size / (1024**2)
                if db_path.exists()
                else 0
            ),
            "journal_mode": settings.SQLITE_JOURNAL_MODE,
            "synchronous": settings.SQLITE_SYNCHRONOUS
        })

    return info