wangpeng/yidaima_tools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Optimize cloud drive upload flow

Browse Source
This commit is contained in:
王鹏
2026-05-12 17:49:51 +08:00
parent a6d79d9a14
commit b95e34da7e
3 changed files with 1185 additions and 203 deletions
Download Patch File Download Diff File Expand all files Collapse all files

498
quark_uploader.py
View File

@@ -2,19 +2,30 @@ from __future__ import annotations
import os
import time
from typing import Optional, Callable
from typing import Callable, Optional
from playwright.sync_api import TimeoutError as PlaywrightTimeoutError
from playwright.sync_api import sync_playwright
class QuarkUploader:
"""
夸克网盘上传工具 - 终极稳健版 (双重状态锁)
夸克网盘上传工具。
通过真实 Chrome + 持久用户目录上传，避免依赖非公开接口。
"""
def __init__(self, chrome_path: str, cookies_dir: str, log_callback: Optional[Callable[[str], None]] = None):
def __init__(
self,
chrome_path: str,
cookies_dir: str,
log_callback: Optional[Callable[[str], None]] = None,
):
self.chrome_path = chrome_path
self.cookies_dir = cookies_dir
self.log_callback = log_callback
self.url = "https://pan.quark.cn/"
if not os.path.exists(self.cookies_dir):
os.makedirs(self.cookies_dir, exist_ok=True)
@@ -23,144 +34,403 @@ class QuarkUploader:
if self.log_callback:
self.log_callback(message)
def upload_file(self, file_path: str, target_folder_name: str, root_path: str = "精品项目整理") -> bool:
def upload_file(
self,
file_path: str,
target_folder_name: str,
root_path: str = "精品项目整理",
) -> bool:
"""
上传文件到夸克网盘
上传文件到夸克网盘。
"""
if not os.path.exists(file_path):
self._log(f"错误: 本地文件不存在 {file_path}")
return False
filename = os.path.basename(file_path)
file_size_mb = os.path.getsize(file_path) / 1024 / 1024
context = None
try:
with sync_playwright() as p:
self._log(f"正在启动浏览器...")
self._log("正在启动浏览器...")
launch_args = {
"user_data_dir": self.cookies_dir,
"headless": False,
"viewport": {"width": 1280, "height": 800}
"viewport": {"width": 1280, "height": 800},
}
if self.chrome_path and os.path.exists(self.chrome_path):
launch_args["executable_path"] = self.chrome_path
context = p.chromium.launch_persistent_context(**launch_args)
page = context.new_page()
self._log("正在打开夸克网盘...")
page.goto(self.url, wait_until="networkidle")
# 1. 登录检测
try:
page.wait_for_selector("text=全部文件", timeout=10000)
except:
self._log("请在浏览器中完成登录...")
page.wait_for_selector("text=全部文件", timeout=180000)
# 2. 导航到目录
self._log(f"定位目录: {root_path} > {target_folder_name}")
page.get_by_text("全部文件").first.click()
page.wait_for_timeout(2000)
# 进入根目录
if not page.get_by_text(root_path, exact=True).first.is_visible():
self._create_folder(page, root_path)
page.get_by_text(root_path, exact=True).first.dblclick()
page.wait_for_timeout(1000)
# 进入项目目录
if not page.get_by_text(target_folder_name, exact=True).first.is_visible():
self._create_folder(page, target_folder_name)
page.get_by_text(target_folder_name, exact=True).first.dblclick()
page.wait_for_timeout(2000)
# 3. 查重
if page.get_by_text(filename, exact=True).count() > 0:
self._log(f"云端已存在 '{filename}'，跳过上传。")
context.close()
return True
# 4. 执行上传
self._log(f"开始上传: {filename}")
# 清除可能存在的旧“上传成功”提示
page.evaluate("() => { const els = document.querySelectorAll('.ant-notification-notice'); els.forEach(e => e.remove()); }")
page.set_input_files("input[type=file]", file_path)
# 5. 确认上传已启动 (关键一步)
self._log("等待上传任务初始化...")
try:
# 等待出现任何进度指示元素
page.wait_for_selector("text=% , .ant-progress-inner, text=正在上传", timeout=20000)
self._log("检测到上传流已建立。")
except:
self._log("警告：未检测到明显的进度条，可能由于文件较小或 UI 延迟，继续监测列表状态。")
# 6. 深度监测循环
self._log("进入深度监测模式，请保持浏览器前台运行...")
start_time = time.time()
timeout = 1800 # 30分钟
# 连续稳定次数计数
stable_count = 0
while time.time() - start_time < timeout:
# A. 检查全局进度标志
is_uploading = page.get_by_text("%").is_visible() or \
page.get_by_text("正在上传").is_visible() or \
page.locator(".ant-progress-inner").is_visible()
# B. 检查列表中该行的具体状态 (精准匹配)
# 找到包含文件名的那一行 tr，看里面是否有“正在上传”字样
row_status_text = ""
try:
row = page.locator(f"tr:has-text('{filename}')").first
if row.is_visible():
row_status_text = row.inner_text()
except:
pass
is_item_active = "正在上传" in row_status_text or "%" in row_status_text or "等待上传" in row_status_text
# C. 确认列表里确实有这个文件
in_list = page.get_by_text(filename, exact=True).count() > 0
if not is_uploading and not is_item_active and in_list:
stable_count += 1
if stable_count >= 5: # 连续 5 次检测(约 15 秒)都处于稳定态
self._log("检测到上传任务已从任务列表中消失，且列表文件状态正常。")
self._log("为了绝对安全，最后等待 15 秒进行数据落盘同步...")
page.wait_for_timeout(15000)
self._log("上传确认圆满成功！")
context.close()
return True
else:
self._log(f"上传疑似完成，正在进行稳定性校验 ({stable_count}/5)...")
else:
if is_uploading or is_item_active:
self._log("监测到活跃传输流...")
stable_count = 0 # 只要发现还在传，重置计数
time.sleep(3)
context = p.chromium.launch_persistent_context(**launch_args)
context.add_init_script(
"Object.defineProperty(navigator, 'webdriver', {get: () => undefined})"
)
page = context.new_page()
page.set_default_timeout(60000)
self._log("正在打开夸克网盘...")
page.goto(self.url, wait_until="domcontentloaded", timeout=60000)
if not self._wait_for_login(page, timeout=10):
self._log("请在浏览器中完成夸克登录，登录后会自动继续...")
if not self._wait_for_login(page, timeout=180):
raise RuntimeError("等待登录超时")
self._log(f"定位目录: {root_path} > {target_folder_name}")
self._open_all_files(page)
self._wait_for_file_list_loaded(page)
if not self._ensure_folder(page, root_path):
raise RuntimeError(f"无法进入或创建根目录: {root_path}")
if not self._ensure_folder(page, target_folder_name):
raise RuntimeError(f"无法进入或创建项目目录: {target_folder_name}")
if self._check_file_exists(page, filename):
self._log(f"云端已存在 '{filename}'，跳过上传。")
return True
self._clear_old_notifications(page)
self._log(f"开始上传: {filename} ({file_size_mb:.2f} MB)")
self._select_upload_file(page, file_path)
if not self._wait_for_upload_started(page, filename):
self._log("警告：未检测到明显的上传任务，继续监测文件列表。")
else:
self._log("检测到上传任务已建立。")
success = self._wait_for_upload_finished(page, filename)
if success:
self._log("夸克网盘上传确认成功。")
return True
self._log("传输任务监测超时，请确认网速或手动核实。")
context.close()
return False
except Exception as e:
self._log(f"代码执行异常: {str(e)}")
except Exception as exc:
self._log(f"代码执行异常: {exc}")
return False
finally:
if context:
try:
context.close()
except Exception:
pass
def _wait_for_login(self, page, timeout: float) -> bool:
start_time = time.time()
while time.time() - start_time < timeout:
try:
body_text = self._body_text(page)
if "全部文件" in body_text or "我的文件" in body_text:
return True
except Exception:
pass
time.sleep(1)
return False
def _open_all_files(self, page):
if self._click_text(page, ["全部文件", "我的文件"]):
page.wait_for_timeout(1000)
return
self._log("未找到“全部文件”入口，尝试直接在当前页面继续。")
def _wait_for_file_list_loaded(
self,
page,
timeout: float = 12,
poll_interval: float = 0.4,
) -> bool:
start_time = time.time()
while time.time() - start_time < timeout:
try:
state = page.evaluate(
"""
() => {
const text = document.body ? document.body.innerText : '';
const loadingWords = ['加载中', '正在加载', '请稍候'];
const hasLoadingText = loadingWords.some(word => text.includes(word));
const itemCount = document.querySelectorAll(
'[title], [aria-label], [class*="file"], [class*="File"]'
).length;
const hasAction = ['新建', '上传', '全部文件', '我的文件']
.some(word => text.includes(word));
const hasEmptyState = ['暂无文件', '空文件夹', '拖拽文件']
.some(word => text.includes(word));
return {hasLoadingText, itemCount, hasAction, hasEmptyState};
}
"""
)
if (
state["hasAction"]
and (state["itemCount"] > 0 or state["hasEmptyState"] or not state["hasLoadingText"])
):
return True
except Exception:
pass
time.sleep(poll_interval)
return False
def _ensure_folder(self, page, folder_name: str) -> bool:
self._wait_for_file_list_loaded(page)
if self._open_folder(page, folder_name):
return True
self._log(f"目录不存在，准备创建: {folder_name}")
if not self._create_folder(page, folder_name):
return False
def _create_folder(self, page, folder_name: str):
"""创建目录"""
end_time = time.time() + 20
while time.time() < end_time:
if self._open_folder(page, folder_name):
return True
time.sleep(1)
return False
def _open_folder(self, page, folder_name: str) -> bool:
try:
page.get_by_text("新建").first.click()
locator = page.get_by_text(folder_name, exact=True)
count = locator.count()
for index in range(min(count, 5)):
item = locator.nth(index)
if item.is_visible():
item.dblclick(force=True, timeout=5000)
page.wait_for_timeout(1200)
self._wait_for_file_list_loaded(page)
self._log(f"已进入目录: {folder_name}")
return True
except Exception:
pass
return False
def _create_folder(self, page, folder_name: str) -> bool:
try:
if not self._click_text(page, ["新建"]):
self._log("未找到“新建”按钮。")
return False
page.wait_for_timeout(500)
page.get_by_text("新建文件夹").first.click()
if not self._click_text(page, ["新建文件夹", "文件夹"]):
self._log("未找到“新建文件夹”入口。")
return False
page.wait_for_timeout(500)
page.keyboard.type(folder_name)
page.keyboard.press("Enter")
self._log(f"已创建目录: {folder_name}")
page.wait_for_timeout(1500)
except:
return True
except Exception as exc:
self._log(f"创建目录失败: {exc}")
return False
def _select_upload_file(self, page, file_path: str):
input_locator = page.locator("input[type=file]")
if input_locator.count() == 0:
self._click_text(page, ["上传", "上传文件"])
page.wait_for_selector("input[type=file]", timeout=10000)
page.locator("input[type=file]").first.set_input_files(file_path)
def _wait_for_upload_started(
self,
page,
filename: str,
timeout: float = 20,
) -> bool:
start_time = time.time()
while time.time() - start_time < timeout:
state = self._upload_state(page, filename)
if state["uploading"] or state["item_active"] or state["file_exists"]:
return True
time.sleep(0.5)
return False
def _wait_for_upload_finished(
self,
page,
filename: str,
timeout: float = 1800,
) -> bool:
start_time = time.time()
stable_count = 0
last_active_log = 0.0
self._log("进入上传监测模式，请保持浏览器前台运行...")
while time.time() - start_time < timeout:
state = self._upload_state(page, filename)
if state["has_error"]:
self._log("检测到页面提示上传失败。")
return False
if state["file_exists"] and not state["uploading"] and not state["item_active"]:
stable_count += 1
if stable_count >= 5:
self._log("文件已出现在列表中，且上传状态连续稳定。")
self._log("最后等待 15 秒进行网盘同步确认...")
page.wait_for_timeout(15000)
return self._check_file_exists(page, filename)
self._log(f"上传疑似完成，稳定性校验 ({stable_count}/5)...")
else:
stable_count = 0
now = time.time()
if (state["uploading"] or state["item_active"]) and now - last_active_log >= 15:
self._log("监测到活跃传输流，继续等待...")
last_active_log = now
time.sleep(3)
return False
def _upload_state(self, page, filename: str) -> dict:
try:
state = page.evaluate(
"""
(filename) => {
const normalize = value => (value || '').replace(/\\s+/g, ' ').trim();
const visible = element => {
if (!element) return false;
const style = window.getComputedStyle(element);
const rect = element.getBoundingClientRect();
return style.display !== 'none'
&& style.visibility !== 'hidden'
&& rect.width > 0
&& rect.height > 0;
};
const bodyText = normalize(document.body ? document.body.innerText : '');
const nodes = [
...document.querySelectorAll(
'[title], [aria-label], [class*="file"], [class*="File"], [class*="name"], [class*="Name"]'
)
].filter(visible);
const matchingNode = nodes.find(node => {
const title = normalize(node.getAttribute('title'));
const label = normalize(node.getAttribute('aria-label'));
const text = normalize(node.textContent);
return title === filename || label === filename || text === filename;
});
const container = matchingNode
? matchingNode.closest('tr, li, [class*="item"], [class*="Item"], [class*="row"], [class*="Row"], [class*="file"], [class*="File"]')
: null;
const containerText = normalize(container ? container.innerText : '');
const activeWords = ['正在上传', '上传中', '等待上传', '传输中', '解析中', '处理中'];
const errorWords = ['上传失败', '传输失败', '网络异常', '上传出错'];
const progressPattern = /(?:^|\\D)(?:100|\\d{1,2})(?:\\.\\d+)?%/;
const itemHasProgress = container
? [...container.querySelectorAll(
'.ant-progress-inner, [class*="progress"], [class*="Progress"]'
)].some(visible)
: false;
const itemActive = activeWords.some(word => containerText.includes(word))
|| progressPattern.test(containerText)
|| itemHasProgress;
const uploading = activeWords.some(word => bodyText.includes(word))
|| itemActive;
const hasError = errorWords.some(word => bodyText.includes(word));
return {
file_exists: Boolean(matchingNode),
item_active: itemActive,
uploading,
has_error: hasError,
};
}
""",
filename,
)
return {
"file_exists": bool(state.get("file_exists")),
"item_active": bool(state.get("item_active")),
"uploading": bool(state.get("uploading")),
"has_error": bool(state.get("has_error")),
}
except Exception:
return {
"file_exists": self._check_file_exists(page, filename),
"item_active": False,
"uploading": False,
"has_error": False,
}
def _check_file_exists(self, page, filename: str) -> bool:
try:
locator = page.get_by_text(filename, exact=True)
count = locator.count()
for index in range(min(count, 5)):
if locator.nth(index).is_visible():
return True
except Exception:
pass
try:
names = page.evaluate(
"""
() => {
const normalize = value => (value || '').replace(/\\s+/g, ' ').trim();
const visible = element => {
const style = window.getComputedStyle(element);
const rect = element.getBoundingClientRect();
return style.display !== 'none'
&& style.visibility !== 'hidden'
&& rect.width > 0
&& rect.height > 0;
};
const result = [];
for (const node of document.querySelectorAll('[title], [aria-label]')) {
if (!visible(node)) continue;
const title = normalize(node.getAttribute('title'));
const label = normalize(node.getAttribute('aria-label'));
if (title) result.push(title);
if (label) result.push(label);
}
return result;
}
"""
)
return filename in names
except Exception:
return False
def _click_text(self, page, texts: list[str]) -> bool:
for text in texts:
for exact in (True, False):
try:
locator = page.get_by_text(text, exact=exact)
count = locator.count()
for index in range(min(count, 5)):
item = locator.nth(index)
if item.is_visible():
item.click(force=True, timeout=5000)
return True
except PlaywrightTimeoutError:
continue
except Exception:
continue
return False
def _clear_old_notifications(self, page):
try:
page.evaluate(
"""
() => {
document
.querySelectorAll('.ant-notification-notice, .ant-message-notice')
.forEach(element => element.remove());
}
"""
)
except Exception:
pass
def _body_text(self, page) -> str:
return page.evaluate("() => document.body ? document.body.innerText : ''")