app/api/v1/system.py

from __future__ import annotations

import io
import os
import platform
import tarfile
import time
from datetime import datetime, timezone
from pathlib import Path

from fastapi import APIRouter, Depends, File, HTTPException, UploadFile
from fastapi.responses import StreamingResponse

from app.api.deps import get_extraction_service
from app.config import get_settings
from app.core.constants import (
    ARCHIVE_EXTENSIONS,
    AUDIO_EXTENSIONS,
    DOCUMENT_EXTENSIONS,
    IMAGE_EXTENSIONS,
    OFFICE_EXTENSIONS,
    SUPPORTED_EXTENSIONS,
    TEXT_EXTENSIONS,
    WEB_EXTENSIONS,
)
from app.core.logger import get_logger
from app.core.thread_pool import run_in_executor
from app.models.schemas import (
    HealthResponse,
    InfoResponse,
    SpacyLabelsResponse,
    SupportedFormatsResponse,
)
from app.services.extraction_service import ExtractionService

router = APIRouter()  # system routes
_settings = get_settings()
_START_TIME = time.time()
_logger = get_logger(__name__)

_MAINTENANCE_MODE: bool = False


def _maintenance_lock_path() -> Path:
    return _get_data_dir() / ".maintenance.lock"


def is_maintenance() -> bool:
    if _MAINTENANCE_MODE:
        return True
    if os.environ.get("APP_UNDER_MAINTENANCE", "").lower() in ("1", "true", "yes"):
        return True
    lock_path = _maintenance_lock_path()
    return lock_path.is_file()


@router.get("/health", response_model=HealthResponse, summary="Health check")
async def health():
    return HealthResponse(
        success=True,
        status="ok",
        version=_settings.app_version,
        uptime_seconds=round(time.time() - _START_TIME, 2),
        timestamp=datetime.now(timezone.utc).isoformat(),
    )


@router.get("/info", response_model=InfoResponse, summary="Server and environment information")
async def info():
    return InfoResponse(
        success=True,
        app=_settings.app_name,
        version=_settings.app_version,
        python_version=platform.python_version(),
        platform=platform.system(),
        uptime_seconds=round(time.time() - _START_TIME, 2),
        max_upload_mb=_settings.max_upload_mb,
        supported_extensions=len(SUPPORTED_EXTENSIONS),
        timestamp=datetime.now(timezone.utc).isoformat(),
    )


@router.get("/formats", response_model=SupportedFormatsResponse, summary="List supported file formats")
async def list_formats():
    by_category = {
        "documents": [e for e in SUPPORTED_EXTENSIONS if e in DOCUMENT_EXTENSIONS],
        "office": [e for e in SUPPORTED_EXTENSIONS if e in OFFICE_EXTENSIONS],
        "data": [e for e in SUPPORTED_EXTENSIONS if e in {".csv", ".json", ".xml"}],
        "web": [e for e in SUPPORTED_EXTENSIONS if e in WEB_EXTENSIONS],
        "text": [e for e in SUPPORTED_EXTENSIONS if e in TEXT_EXTENSIONS],
        "images": [e for e in SUPPORTED_EXTENSIONS if e in IMAGE_EXTENSIONS],
        "audio": [e for e in SUPPORTED_EXTENSIONS if e in AUDIO_EXTENSIONS],
        "archives": [e for e in SUPPORTED_EXTENSIONS if e in ARCHIVE_EXTENSIONS],
    }
    return SupportedFormatsResponse(
        success=True,
        total_count=len(SUPPORTED_EXTENSIONS),
        all_extensions=sorted(SUPPORTED_EXTENSIONS),
        by_category={k: sorted(v) for k, v in by_category.items()},
    )


@router.get("/spacy-labels", response_model=SpacyLabelsResponse, summary="List available spaCy NER labels")
async def list_spacy_labels(
    extraction_service: ExtractionService = Depends(get_extraction_service),
):
    return SpacyLabelsResponse(
        success=True,
        spacy_labels=extraction_service.get_spacy_labels(),
        source_types={
            "entity": "Extract using spaCy NER labels (ORG, PERSON, DATE, etc.)",
            "regex": "Extract using custom regular expressions",
            "token_attr": "Extract using token attributes (text, pos_, tag_, etc.)",
        },
        example_mappings={
            "company": {"source_type": "entity", "label": "ORG"},
            "person": {"source_type": "entity", "label": "PERSON"},
            "date": {"source_type": "entity", "label": "DATE"},
            "money": {"source_type": "entity", "label": "MONEY"},
            "email": {
                "source_type": "regex",
                "pattern": r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b",
            },
            "phone": {"source_type": "regex", "pattern": r"\b\d{3}-\d{3}-\d{4}\b"},
        },
    )


def _build_archive(data_dir: Path) -> io.BytesIO:
    buf = io.BytesIO()
    with tarfile.open(fileobj=buf, mode="w:gz") as tar:
        for path in sorted(data_dir.rglob("*")):
            if path.is_file():
                arcname = path.relative_to(data_dir.parent)
                tar.add(str(path), arcname=str(arcname))
    buf.seek(0)
    return buf


def _extract_archive(content: bytes, data_dir: Path) -> int:
    restored = 0
    with tarfile.open(fileobj=io.BytesIO(content), mode="r:gz") as tar:
        for member in tar.getmembers():
            if member.isfile():
                tar.extract(member, path=data_dir.parent)
                restored += 1
    return restored


def _get_data_dir() -> Path:
    raw = _settings.data_dir
    p = Path(raw)
    if not p.is_absolute():
        p = Path.cwd() / p
    return p.resolve()


@router.get("/backup", summary="Download a full data backup archive")
async def download_backup():
    data_dir = _get_data_dir()
    if not data_dir.is_dir():
        raise HTTPException(status_code=404, detail="Data directory not found")
    try:
        archive = await run_in_executor(_build_archive, data_dir)
    except Exception as exc:
        _logger.error("Backup creation failed: %s", exc)
        raise HTTPException(status_code=500, detail=f"Backup failed: {exc}")

    return StreamingResponse(
        archive,
        media_type="application/gzip",
        headers={"Content-Disposition": "attachment; filename=backup.tar.gz"},
    )


@router.post("/backup/restore", summary="Upload and restore a data backup archive")
async def upload_and_restore(
    file: UploadFile = File(...),
):
    data_dir = _get_data_dir()
    data_dir.mkdir(parents=True, exist_ok=True)

    try:
        content = await file.read()
    except Exception as exc:
        raise HTTPException(status_code=400, detail=f"Failed to read upload: {exc}")

    if not content:
        raise HTTPException(status_code=400, detail="Empty file")

    restored = 0
    try:
        restored = await run_in_executor(_extract_archive, content, data_dir)
    except tarfile.TarError as exc:
        raise HTTPException(status_code=400, detail=f"Invalid archive: {exc}")

    return {
        "success": True,
        "message": f"Restored {restored} files to {data_dir}",
        "files_restored": restored,
        "data_dir": str(data_dir),
    }


@router.post("/maintenance/enable", summary="Enable maintenance mode (blocks write operations)")
async def enable_maintenance():
    global _MAINTENANCE_MODE
    _MAINTENANCE_MODE = True
    lock_path = _maintenance_lock_path()
    try:
        lock_path.parent.mkdir(parents=True, exist_ok=True)
        lock_path.write_text(str(time.time()), encoding="utf-8")
    except Exception as exc:
        _logger.warning("Could not write maintenance lock file: %s", exc)
    _logger.warning("Maintenance mode ENABLED — all write operations blocked")
    return {"success": True, "message": "Maintenance mode enabled", "maintenance": True}


@router.post("/maintenance/disable", summary="Disable maintenance mode")
async def disable_maintenance():
    global _MAINTENANCE_MODE
    _MAINTENANCE_MODE = False
    lock_path = _maintenance_lock_path()
    try:
        lock_path.unlink(missing_ok=True)
    except Exception as exc:
        _logger.warning("Could not remove maintenance lock file: %s", exc)
    _logger.warning("Maintenance mode DISABLED — write operations resumed")
    return {"success": True, "message": "Maintenance mode disabled", "maintenance": False}


@router.get("/maintenance", summary="Check maintenance mode status")
async def maintenance_status():
    lock_path = _maintenance_lock_path()
    source = "off"
    if _MAINTENANCE_MODE:
        source = "api_flag"
    elif os.environ.get("APP_UNDER_MAINTENANCE"):
        source = "env_var"
    elif lock_path.is_file():
        source = "lock_file"
    return {
        "success": True,
        "maintenance": is_maintenance(),
        "source": source,
    }