xhs-note-crawling/workplace/input-keyword/find_search_input.py

#!/usr/bin/env python3
"""
用 ``workplace/ocr-pos.py``（RapidOCR）在**当前屏幕截图**中查找固定文案，
返回文字框中心点的**屏幕像素坐标** ``(x, y)``；入口为 ``start``。

保留 ``selector_search_input`` / ``SEARCH_INPUT_SELECTORS`` 供 ``input-word-by-word.py`` 等旧代码导入。
"""

from __future__ import annotations

import importlib.util
import sys
from pathlib import Path
from typing import TYPE_CHECKING

if TYPE_CHECKING:
    from playwright.sync_api import Page

target_text = "搜索小红书"

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

from workplace.screenshot import PillowImageGrabFullScreenScreenshotCaptureSaver  # noqa: E402

SEARCH_INPUT_SELECTORS: tuple[str, ...] = (
    "#search-input",
    "input#search-input",
    "textarea#search-input",
)
selector_search_input = SEARCH_INPUT_SELECTORS[0]


class SearchInputFlowResult:
    """不用 ``@dataclass``：经 ``exec_module`` 动态加载时，与 ``from __future__ import annotations`` 组合可能触发 CPython dataclasses 内部错误。"""

    __slots__ = ("ok", "message", "center_xy")

    def __init__(
        self,
        ok: bool,
        message: str = "",
        center_xy: tuple[int, int] | None = None,
    ) -> None:
        self.ok = ok
        self.message = message
        self.center_xy = center_xy


def _load_ocr_pos():
    path = Path(__file__).resolve().parent.parent / "ocr-pos.py"
    spec = importlib.util.spec_from_file_location("workplace_ocr_pos_script", path)
    if spec is None or spec.loader is None:
        raise ImportError(f"Cannot load {path}")
    mod = importlib.util.module_from_spec(spec)
    spec.loader.exec_module(mod)
    return mod


def start(
    page: Page,
    keyword: str = "",
    *,
    preset_center_xy: tuple[int, int] | None = None,
) -> SearchInputFlowResult:
    """
    截取整屏（``capture_full_screen_and_store_in_memory`` 返回 BGR 图，并同步写入 ``workplace.screenshot`` 模块变量），
    OCR 匹配 ``target_text``（与 ``keyword`` 无关），返回中心点屏幕坐标。

    若传入 ``preset_center_xy`` 则跳过截图与 OCR，直接返回该坐标。

    ``page`` 仅保留签名兼容。
    """
    _ = page, keyword
    if preset_center_xy is not None:
        x, y = int(preset_center_xy[0]), int(preset_center_xy[1])
        return SearchInputFlowResult(ok=True, message="", center_xy=(x, y))

    ocr_mod = _load_ocr_pos()
    saver = PillowImageGrabFullScreenScreenshotCaptureSaver()
    full_screen_screenshot_bgr_numpy = saver.capture_full_screen_and_store_in_memory()
    pt = ocr_mod.ocr_find_text_center(
        full_screen_screenshot_bgr_numpy,
        target_text,
    )
    if pt is None:
        msg = f"OCR 未在全屏截图中找到「{target_text}」。"
        print(msg, file=sys.stderr)
        return SearchInputFlowResult(ok=False, message=msg, center_xy=None)
    return SearchInputFlowResult(ok=True, message="", center_xy=pt)