Aislo/B03_FileInput/B03_FileInput_Repository.py

"""B03 input_files 테이블의 aiomysql Raw SQL 접근."""

import json
from pathlib import PurePosixPath
from typing import Any
from uuid import UUID

import aiomysql


async def create_input_file(
    connection: aiomysql.Connection,
    *,
    project_id: UUID,
    file_type: str,
    original_filename: str,
    relative_path: str,
    file_size_bytes: int,
    upload_by: int | None,
    crs_epsg: int | None,
    metadata: dict[str, Any],
) -> int:
    """업로드 원본 파일 메타데이터를 저장하고 생성된 ID를 반환한다."""
    normalized_path = PurePosixPath(relative_path)
    if normalized_path.is_absolute() or ".." in normalized_path.parts:
        raise ValueError("DB에는 프로젝트 루트 기준 상대 경로만 저장할 수 있습니다.")
    if normalized_path.parts[:2] != ("B03_FileInput", "input"):
        raise ValueError("입력 파일 경로는 B03_FileInput/input 아래여야 합니다.")

    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            INSERT INTO input_files (
                project_id,
                file_type,
                original_filename,
                raw_file_path,
                file_size_mb,
                upload_by,
                crs_epsg,
                metadata,
                status
            )
            VALUES (%s, %s, %s, %s, %s, %s, %s, %s, 'UPLOADED')
            """,
            (
                str(project_id),
                file_type,
                original_filename,
                normalized_path.as_posix(),
                file_size_bytes / (1024 * 1024),
                upload_by,
                crs_epsg,
                json.dumps(metadata, ensure_ascii=False),
            ),
        )
        input_file_id = cursor.lastrowid

    if not input_file_id:
        raise RuntimeError("input_files 레코드 생성 결과에 ID가 없습니다.")
    return int(input_file_id)


async def get_project_input_readiness(
    connection: aiomysql.Connection,
    project_id: UUID,
) -> tuple[set[str], int | None, int | None]:
    """업로드 파일 유형, 최신 포인트클라우드 입력 ID, 최신 계획노선 CSV 입력 ID를 반환한다."""
    async with connection.cursor(aiomysql.DictCursor) as cursor:
        await cursor.execute(
            """
            SELECT id, LOWER(file_type) AS file_type
            FROM input_files
            WHERE project_id = %s AND status IN ('UPLOADED', 'PROCESSED')
            ORDER BY id DESC
            """,
            (str(project_id),),
        )
        rows = await cursor.fetchall()

    file_types = {str(row["file_type"]) for row in rows if row.get("file_type")}
    point_cloud_id = next(
        (int(row["id"]) for row in rows if str(row.get("file_type") or "") in {"las", "laz"}),
        None,
    )
    # LAS 없는 설계(2026-08-30)의 WF1 입력 — 계획노선 CSV가 분석 원천이 된다.
    route_csv_id = next(
        (int(row["id"]) for row in rows if str(row.get("file_type") or "") == "csv"),
        None,
    )
    return file_types, point_cloud_id, route_csv_id


async def get_project_storage_relative_path(
    connection: aiomysql.Connection,
    project_id: UUID,
) -> str:
    """프로젝트의 검증된 저장소 상대 경로를 조회한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            SELECT storage_path
            FROM projects
            WHERE id = %s AND deleted_at IS NULL
            """,
            (str(project_id),),
        )
        row = await cursor.fetchone()

    if not row or not row[0]:
        raise LookupError("프로젝트 또는 프로젝트 저장 경로를 찾을 수 없습니다.")

    normalized_path = PurePosixPath(str(row[0]).replace("\\", "/"))
    if normalized_path.is_absolute() or ".." in normalized_path.parts:
        raise ValueError("프로젝트 저장 경로는 안전한 상대 경로여야 합니다.")
    return normalized_path.as_posix()


async def create_upload_session(
    connection: aiomysql.Connection,
    *,
    session_id: str,
    project_id: UUID,
    original_filename: str,
    file_size_bytes: int,
    chunk_size_bytes: int,
    total_chunks: int,
) -> None:
    """청크 업로드 세션을 생성하거나 동일 세션 ID를 갱신한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            INSERT INTO upload_sessions (
                id,
                project_id,
                original_filename,
                file_size_bytes,
                chunk_size_bytes,
                total_chunks,
                completed_chunks,
                status,
                created_at,
                updated_at
            )
            VALUES (%s, %s, %s, %s, %s, %s, 0, 'in_progress', NOW(), NOW())
            ON DUPLICATE KEY UPDATE
                original_filename = VALUES(original_filename),
                file_size_bytes = VALUES(file_size_bytes),
                chunk_size_bytes = VALUES(chunk_size_bytes),
                total_chunks = VALUES(total_chunks),
                status = 'in_progress',
                updated_at = NOW()
            """,
            (
                session_id,
                str(project_id),
                original_filename,
                file_size_bytes,
                chunk_size_bytes,
                total_chunks,
            ),
        )


async def get_upload_session(
    connection: aiomysql.Connection,
    *,
    project_id: UUID,
    session_id: str,
) -> dict[str, Any]:
    """청크 업로드 세션을 조회한다."""
    async with connection.cursor(aiomysql.DictCursor) as cursor:
        await cursor.execute(
            """
            SELECT
                id,
                project_id,
                original_filename,
                file_size_bytes,
                chunk_size_bytes,
                total_chunks,
                completed_chunks,
                status
            FROM upload_sessions
            WHERE id = %s AND project_id = %s
            """,
            (session_id, str(project_id)),
        )
        row = await cursor.fetchone()
    if not row:
        raise LookupError("업로드 세션을 찾을 수 없습니다.")
    return dict(row)


async def upsert_upload_chunk(
    connection: aiomysql.Connection,
    *,
    session_id: str,
    chunk_index: int,
    chunk_hash: str,
    size_bytes: int,
    stored_at: str,
) -> int:
    """청크 저장 정보를 기록하고 완료 청크 수를 반환한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            INSERT INTO upload_chunks (
                session_id,
                chunk_index,
                chunk_hash,
                size_bytes,
                stored_at,
                completed_at
            )
            VALUES (%s, %s, %s, %s, %s, NOW())
            ON DUPLICATE KEY UPDATE
                chunk_hash = VALUES(chunk_hash),
                size_bytes = VALUES(size_bytes),
                stored_at = VALUES(stored_at),
                completed_at = NOW()
            """,
            (session_id, chunk_index, chunk_hash, size_bytes, stored_at),
        )
        await cursor.execute(
            """
            SELECT COUNT(*)
            FROM upload_chunks
            WHERE session_id = %s
            """,
            (session_id,),
        )
        row = await cursor.fetchone()
        completed_chunks = int(row[0]) if row else 0
        await cursor.execute(
            """
            UPDATE upload_sessions
            SET completed_chunks = %s, updated_at = NOW()
            WHERE id = %s
            """,
            (completed_chunks, session_id),
        )
    return completed_chunks


async def list_completed_chunk_indexes(
    connection: aiomysql.Connection,
    *,
    session_id: str,
) -> list[int]:
    """완료된 청크 인덱스 목록을 반환한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            SELECT chunk_index
            FROM upload_chunks
            WHERE session_id = %s
            ORDER BY chunk_index ASC
            """,
            (session_id,),
        )
        rows = await cursor.fetchall()
    return [int(row[0]) for row in rows]


async def mark_upload_session_completed(
    connection: aiomysql.Connection,
    *,
    session_id: str,
) -> None:
    """업로드 세션을 완료 상태로 표시한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            UPDATE upload_sessions
            SET status = 'completed', updated_at = NOW()
            WHERE id = %s
            """,
            (session_id,),
        )


async def mark_upload_session_failed(
    connection: aiomysql.Connection,
    *,
    session_id: str,
) -> None:
    """업로드 세션을 실패 상태로 표시한다."""
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            UPDATE upload_sessions
            SET status = 'failed', updated_at = NOW()
            WHERE id = %s
            """,
            (session_id,),
        )


async def list_project_input_files(
    connection: aiomysql.Connection,
    project_id: UUID,
) -> list[dict[str, Any]]:
    """업로드 완료된 입력 파일 목록(재접속 현황 표시용) — 서버가 정본이다.

    같은 파일명을 다시 올리면 새 레코드가 쌓이므로 파일명별 최신 것만 남긴다.
    """
    async with connection.cursor(aiomysql.DictCursor) as cursor:
        await cursor.execute(
            """
            SELECT f.id, f.file_type, f.original_filename, f.file_size_mb, f.status,
                   f.upload_at
            FROM input_files f
            INNER JOIN (
                SELECT MAX(id) AS id
                FROM input_files
                WHERE project_id = %s AND status IN ('UPLOADED', 'PROCESSED')
                GROUP BY original_filename
            ) latest ON latest.id = f.id
            ORDER BY f.id ASC
            """,
            (str(project_id),),
        )
        rows = await cursor.fetchall()
    return [dict(row) for row in rows]


async def list_incomplete_upload_sessions(
    connection: aiomysql.Connection,
    project_id: UUID,
) -> list[dict[str, Any]]:
    """중단된(미완료) 청크 업로드 세션 목록 — 재접속 시 이어올리기 안내용."""
    async with connection.cursor(aiomysql.DictCursor) as cursor:
        await cursor.execute(
            """
            SELECT id, original_filename, file_size_bytes, chunk_size_bytes,
                   total_chunks, completed_chunks, updated_at
            FROM upload_sessions
            WHERE project_id = %s AND status = 'in_progress'
            ORDER BY updated_at DESC
            """,
            (str(project_id),),
        )
        rows = await cursor.fetchall()
    return [dict(row) for row in rows]


async def find_input_file_by_name(
    connection: aiomysql.Connection,
    project_id: UUID,
    original_filename: str,
) -> dict[str, Any] | None:
    """같은 이름으로 등록된 최신 입력 파일 1건. 없으면 None.

    같은 파일을 다시 올렸는지 가리는 데 쓴다 — 중복 판정 기준은 파일명이고, 내용이 같은지는
    이 행의 메타데이터에 적힌 지문으로 본다(2026-08-08 사용자 결정).
    """
    async with connection.cursor(aiomysql.DictCursor) as cursor:
        await cursor.execute(
            """
            SELECT id, original_filename, file_size_mb, metadata
            FROM input_files
            WHERE project_id = %s AND original_filename = %s
              AND status IN ('UPLOADED', 'PROCESSED')
            ORDER BY id DESC
            LIMIT 1
            """,
            (str(project_id), original_filename),
        )
        row = await cursor.fetchone()
    return dict(row) if row else None


async def supersede_previous_input_files(
    connection: aiomysql.Connection,
    project_id: UUID,
    original_filename: str,
    keep_input_file_id: int,
) -> int:
    """같은 이름의 옛 행을 `SUPERSEDED`로 내린다. 내린 건수를 돌려준다.

    조회 쿼리들이 `UPLOADED`/`PROCESSED`만 보므로, 이렇게만 해도 목록·분석에서 빠진다.
    행을 지우지 않는 이유는 언제 무엇이 교체됐는지 추적할 근거를 남기기 위해서다.
    """
    async with connection.cursor() as cursor:
        await cursor.execute(
            """
            UPDATE input_files
            SET status = 'SUPERSEDED'
            WHERE project_id = %s AND original_filename = %s AND id <> %s
              AND status IN ('UPLOADED', 'PROCESSED')
            """,
            (str(project_id), original_filename, keep_input_file_id),
        )
        return cursor.rowcount