From 4059494132a1827d1ad384a3372943f7525e312b Mon Sep 17 00:00:00 2001
From: zjf <zjf@bowong.ai>
Date: Wed, 28 May 2025 18:23:24 +0800
Subject: [PATCH] rpa m

---
 src/cluster/rpa.py | 35 ++++++++++++++++++++++-------------
 1 file changed, 22 insertions(+), 13 deletions(-)

diff --git a/src/cluster/rpa.py b/src/cluster/rpa.py
index 17ac884..5bd06ce 100644
--- a/src/cluster/rpa.py
+++ b/src/cluster/rpa.py
@@ -24,6 +24,7 @@ rpa_image = (modal.Image.debian_slim(python_version="3.11")
     "mkdir /root/browsers",
     "playwright install-deps",
     "playwright install chromium",
+    "playwright install msedge",
 ])
              .add_local_python_source('cluster')
              .add_local_python_source('BowongModalFunctions')
@@ -456,6 +457,10 @@ with rpa_image.imports():
 
         webcast_id: str = data.live_id
         logger.info(f"Starting RPA run for webcast_id: {webcast_id}")
+
+        live_url = f"https://bowong-gateway.bowong.cc/api/t1/live/liveUrl?name={data.live_stream_name}"
+        result = LiveStreamResult(live_id=webcast_id, live_stream_url=live_url)
+
         # createLivePull2Push
         with httpx.Client() as client:
             res = client.post(url="https://bowong-gateway.bowong.cc/api/t1/live/createLivePull2Push", json={
@@ -469,17 +474,18 @@ with rpa_image.imports():
         data_res = res.json()
         print(f"!!!!!! data_res:{data_res}")
         if status_code == 200 and data_res['code'] == 0:
-            live_url = f"https://bowong-gateway.bowong.cc/api/t1/live/liveUrl?name={data.live_stream_name}"
-            result = LiveStreamResult(live_id=webcast_id, live_stream_url=live_url)
             result.ext_data = data_res['data']
+        else:
+            result.msg="createLivePull2Push 推流失败"
+            return data_res
 
         async with async_playwright() as p:
-            browser: Browser = await p.chromium.launch(headless=True)
-            try:
-                result = await goto_live(browser, data, result, webcast_id)
-            except BaseException as e:
-                logger.error(f"Failed to start RPA run: {str(e)}")
-                await browser.close()
+            browser: Browser = await p.chromium.launch(headless=True, channel='msedge')
+            # browser = await p.chromium.launch(channel="msedge")
+
+
+            result = await goto_live(browser, data, result, webcast_id)
+
 
         # Save result to JSON file with ULID naming,TODO 后续s3上传
         await save_result_to_json(result, data)
@@ -487,7 +493,7 @@ with rpa_image.imports():
         return result.dict()  # Return as dict for JSON compatibility
 
 
-    @retry(stop=stop_after_attempt(3), wait=wait_fixed(5))
+    # @retry(stop=stop_after_attempt(3), wait=wait_fixed(5))
     async def goto_live(browser, data: LiveStreamProductWatchRequest, result: LiveStreamResult, webcast_id):
         context = await browser.new_context()
         await context.add_cookies(cookies)
@@ -495,10 +501,13 @@ with rpa_image.imports():
         page: Page = await context.new_page()
         url = f"https://live.douyin.com/{webcast_id}?open_promotion_list=1"
         logger.info(f"Navigating to {url}")
-        await page.goto(url, timeout=60000)
-        await page.wait_for_load_state("networkidle", timeout=60000)
+        await page.goto(url,wait_until="domcontentloaded", timeout=1800 * 1000)
+        logger.info(f"Navigating to end {url}")
+        await page.wait_for_load_state("networkidle", timeout=1800 * 1000)
+        logger.info(f"Navigating to end {url}")
         result.title = await page.title()
-        logger.info(f"Retrieved title: {result.title}")
+        logger.info(f"------Retrieved title: {result.title}")
+
         live_status = await get_live_status_by_web_page(page, webcast_id)
 
         if live_status == LIVE_END:
@@ -521,7 +530,7 @@ with rpa_image.imports():
         # 312898731614
         # webcast_ids = ["36861064178"]
         # webcast_ids = ["36861064178"]
-        webcast_id = "206240580433"
+        webcast_id = "629972366172"
 
         data: LiveStreamProductWatchRequest = LiveStreamProductWatchRequest(live_id=webcast_id)
         # TODO other info