"""Diagnostic script: navigate to a Zhihu question, fill + submit an answer, take screenshots at key steps, and upload them to Slack. Usage: python scripts/zhihu_debug_publish.py \ --url "https://www.zhihu.com/question/1992659642908702603" """ import sys import os import time from pathlib import Path sys.path.insert(0, str(Path(__file__).resolve().parents[1])) from health.utils.env_loader import load_env_with_extras load_env_with_extras() import click from slack_sdk import WebClient from playwright.sync_api import sync_playwright from playwright_stealth import Stealth STATE_FILE = Path(__file__).resolve().parents[1].parent / "data" / "zhihu_state.json" SCREENSHOT_DIR = Path("/tmp/zhihu_debug") SCREENSHOT_DIR.mkdir(exist_ok=True) DUMMY_TEXT = "这是一段测试文字，用于诊断知乎发布流程是否正常。请勿实际发布。" def upload(client: WebClient, channel: str, path: Path, comment: str) -> None: with path.open("rb") as f: client.files_upload_v2( channel=channel, file=f, filename=path.name, initial_comment=comment, ) print(f"Uploaded: {path.name}") @click.command() @click.option("--url", required=True, help="Zhihu question URL to test") @click.option("--submit", is_flag=True, default=False, help="Actually click 发布 (default: dry run, no submit)") def main(url: str, submit: bool) -> None: token = os.environ.get("OBSIDIAN_SLACK_BOT_TOKEN") channel = os.environ.get("ZHIHU_NOTIFY_CHANNEL", "") client = WebClient(token=token) import json with sync_playwright() as p: browser = p.chromium.launch(headless=True) context = browser.new_context( viewport={"width": 1280, "height": 800}, user_agent=( "Mozilla/5.0 (Windows NT 10.0; Win64; x64) " "AppleWebKit/537.36 (KHTML, like Gecko) " "Chrome/122.0.0.0 Safari/537.36" ), ) # Load session if STATE_FILE.exists(): data = json.loads(STATE_FILE.read_text()) context.add_cookies(data.get("cookies", [])) print(f"Loaded {len(data.get('cookies', []))} cookies") else: print("WARNING: no state file found, not logged in") page = context.new_page() Stealth().apply_stealth_sync(page) # ── Step 1: navigate ────────────────────────────────────────────────── print(f"Navigating to: {url}") page.goto(url, wait_until="domcontentloaded", timeout=30000) page.wait_for_timeout(3000) s1 = SCREENSHOT_DIR / "1_after_navigate.png" page.screenshot(path=str(s1), full_page=False) upload(client, channel, s1, f"📸 Step 1: after navigation to question page (login={page.url})") # ── Step 2: scroll + find 写回答 button ────────────────────────────── page.evaluate("window.scrollBy(0, 400)") page.wait_for_timeout(800) # Log all buttons texts = page.evaluate(""" () => Array.from(document.querySelectorAll('button, a, [role="button"]')) .map(e => e.textContent.trim()) .filter(t => t.length > 0 && t.length < 30) """) print(f"Buttons/links on page: {texts[:40]}") client.chat_postMessage(channel=channel, text=f"🔍 页面按钮列表：`{texts[:40]}`") # Try click clicked = page.evaluate(""" () => { const walker = document.createTreeWalker(document.body, NodeFilter.SHOW_TEXT); let node; while ((node = walker.nextNode())) { const t = node.textContent.trim(); if (['写回答','回答问题','回答','发起回答'].includes(t)) { const el = node.parentElement; const btn = el.closest('button, a, [role="button"]') || el; btn.click(); return t; } } return null; } """) print(f"Clicked '写回答' text: {clicked!r}") if not clicked: s_err = SCREENSHOT_DIR / "2_no_write_button.png" page.screenshot(path=str(s_err)) upload(client, channel, s_err, "❌ Step 2: 未找到写回答按钮，截图如下") browser.close() return page.wait_for_timeout(2000) s2 = SCREENSHOT_DIR / "2_after_click_write.png" page.screenshot(path=str(s2), full_page=False) upload(client, channel, s2, f"📸 Step 2: after clicking '{clicked}' — editor should be visible") # ── Step 3: fill editor ─────────────────────────────────────────────── editor_sels = [ ".DraftEditor-content[contenteditable='true']", ".RichText [contenteditable='true']", "[contenteditable='true'].public-DraftEditor-content", ] editor = None for sel in editor_sels: try: page.wait_for_selector(sel, timeout=5000) editor = page.locator(sel).first print(f"Found editor via: {sel}") break except Exception: continue if editor is None: s_err2 = SCREENSHOT_DIR / "3_no_editor.png" page.screenshot(path=str(s_err2)) upload(client, channel, s_err2, "❌ Step 3: 未找到编辑器，截图如下") browser.close() return editor.click() page.wait_for_timeout(400) page.evaluate( """(text) => { const ed = document.querySelector( '.DraftEditor-content[contenteditable="true"],' + '.RichText [contenteditable="true"],' + '[contenteditable="true"].public-DraftEditor-content' ); if (ed) { ed.focus(); document.execCommand('insertText', false, text); } }""", DUMMY_TEXT, ) page.wait_for_timeout(600) content_len = page.evaluate(""" () => { const ed = document.querySelector('[contenteditable="true"]'); return ed ? ed.textContent.length : 0; } """) print(f"Editor content length after fill: {content_len}") s3 = SCREENSHOT_DIR / "3_after_fill.png" page.screenshot(path=str(s3), full_page=False) upload(client, channel, s3, f"📸 Step 3: after filling editor ({content_len} chars)") if not submit: print("Dry run — not clicking 发布. Pass --submit to actually submit.") client.chat_postMessage(channel=channel, text="ℹ️ 诊断完成（dry run，未点击发布）") browser.close() return # ── Step 4: click 发布 ──────────────────────────────────────────────── publish_sels = [ "button:has-text('发布回答')", ".AnswerForm button:has-text('发布')", ".AnswerForm button:has-text('提交')", "button:has-text('发布')", "button:has-text('提交')", "button.SubmitAnswer", "[data-za-element-name='SubmitAnswer']", ] published = False matched_sel = None for sel in publish_sels: try: btn = page.locator(sel).first btn.wait_for(state="visible", timeout=3000) btn.click() published = True matched_sel = sel print(f"Clicked 发布 via: {sel}") break except Exception: continue if not published: s_err3 = SCREENSHOT_DIR / "4_no_publish_btn.png" page.screenshot(path=str(s_err3)) upload(client, channel, s_err3, "❌ Step 4: 未找到发布按钮，截图如下") browser.close() return # Wait and screenshot multiple times to catch any dialog for i, delay in enumerate([1000, 2000, 3000], start=1): page.wait_for_timeout(delay) # Log buttons at this moment btns_now = page.evaluate(""" () => Array.from(document.querySelectorAll('button, a[role="button"]')) .map(e => e.textContent.trim()) .filter(t => t.length > 0 && t.length < 30) """) print(f"+{i}s buttons: {btns_now[:20]}") s = SCREENSHOT_DIR / f"4_{i}s_after_publish.png" page.screenshot(path=str(s), full_page=False) upload(client, channel, s, f"📸 Step 4 (+{i}s after 发布 click): url={page.url}") print(f"Final URL: {page.url}") browser.close() if __name__ == "__main__": main()