xhs-note-crawling/workplace/repository_config_snapshot.py

#!/usr/bin/env python3
"""仓库根 ``config.ini`` 与 ``save/project-config`` 的一次性快照。"""

from __future__ import annotations

import configparser
import json
import sys
from dataclasses import dataclass
from pathlib import Path
from typing import Any

_SINGLETON_MODULE_SOURCE_FILE_PATH = Path(__file__).resolve()
REPOSITORY_ROOT_DIRECTORY: Path = _SINGLETON_MODULE_SOURCE_FILE_PATH.parent.parent
_repository_root_directory_string = str(REPOSITORY_ROOT_DIRECTORY)
if _repository_root_directory_string not in sys.path:
    sys.path.insert(0, _repository_root_directory_string)

_CONFIG_INI_FILE_PATH = REPOSITORY_ROOT_DIRECTORY / "config.ini"
_PROJECT_CONFIG_PRIMARY_FILE_PATH = REPOSITORY_ROOT_DIRECTORY / "save" / "project-config"
_PROJECT_CONFIG_LEGACY_FILE_PATH = REPOSITORY_ROOT_DIRECTORY / "save" / "input-keyword.config"


@dataclass(frozen=True)
class PyAutoGUIScreenCalibration:
    viewport_to_screen_scale_x: float = 1.0
    viewport_to_screen_scale_y: float = 1.0
    screen_offset_x: int = 0
    screen_offset_y: int = 0


@dataclass(frozen=True)
class RepositoryConfigIniSnapshot:
    home_page_url: str
    python_executable_relative_path: str
    pyautogui_screen_calibration: PyAutoGUIScreenCalibration
    project_config: dict[str, Any]


_REPOSITORY_CONFIG_INI_SNAPSHOT_SINGLETON: RepositoryConfigIniSnapshot | None = None


def _read_project_config_mapping_from_repository_save_directory() -> dict[str, Any]:
    if _PROJECT_CONFIG_PRIMARY_FILE_PATH.is_file():
        raw_text = _PROJECT_CONFIG_PRIMARY_FILE_PATH.read_text(encoding="utf-8")
        parsed_object = json.loads(raw_text)
        return parsed_object if isinstance(parsed_object, dict) else {}
    if _PROJECT_CONFIG_LEGACY_FILE_PATH.is_file():
        raw_text = _PROJECT_CONFIG_LEGACY_FILE_PATH.read_text(encoding="utf-8")
        parsed_object = json.loads(raw_text)
        return parsed_object if isinstance(parsed_object, dict) else {}
    return {}


def _read_repository_config_ini_snapshot_from_disk() -> RepositoryConfigIniSnapshot:
    parser = configparser.ConfigParser()
    parser.read(_CONFIG_INI_FILE_PATH, encoding="utf-8")
    python_executable_relative_path = parser.get(
        "paths",
        "pythonRelative",
        fallback="python/py/python.exe",
    )
    home_page_url = parser.get(
        "urls",
        "homeUrl",
        fallback="https://www.xiaohongshu.com/",
    )
    pyautogui_screen_calibration = PyAutoGUIScreenCalibration()
    project_config_mapping = _read_project_config_mapping_from_repository_save_directory()
    return RepositoryConfigIniSnapshot(
        home_page_url=home_page_url.strip(),
        python_executable_relative_path=python_executable_relative_path.strip(),
        pyautogui_screen_calibration=pyautogui_screen_calibration,
        project_config=project_config_mapping,
    )


def get_repository_config_ini_snapshot() -> RepositoryConfigIniSnapshot:
    global _REPOSITORY_CONFIG_INI_SNAPSHOT_SINGLETON
    if _REPOSITORY_CONFIG_INI_SNAPSHOT_SINGLETON is None:
        _REPOSITORY_CONFIG_INI_SNAPSHOT_SINGLETON = (
            _read_repository_config_ini_snapshot_from_disk()
        )
    return _REPOSITORY_CONFIG_INI_SNAPSHOT_SINGLETON


__all__ = [
    "PyAutoGUIScreenCalibration",
    "RepositoryConfigIniSnapshot",
    "REPOSITORY_ROOT_DIRECTORY",
    "get_repository_config_ini_snapshot",
]