宜博 / pixelphone

Commit 6cd9f893747a00be87a43500821d7d31b4402122

Authored by yibo
1 parent 06d4382c

feat(phoneuse): send_wechat_voice_tts 路由到 cgi 全合成路 

send_voice_tts_via_hook 的最后一步从 send_voice_via_hook(SEND_VOICE_INTERNAL)
改为 send_voice_cgi_via_hook(SEND_VOICE_CGI)——让 Hermes 的 send_wechat_voice_tts
MCP 工具（wechat-voice-sender skill 调的）走已验证的 uploadvoice cgi 全合成路：
音频上传+投递接收方+发送方本地已发送气泡。docstring 同步更新。

已 adb push 同步到设备 ~/.hermes/event_bus/（android_mcp.py + wechat_hook_send.py），
设备 wechat-image-sender skill 也更新为调 send_wechat_image MCP 工具（弃旧 ACTION_SEND
UI 模拟）。设备文件权限修为 termux uid 可读（原 root:root 导致 MCP 子进程 import 失败）。

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Inline Side-by-side
Showing 2 changed files with 9 additions and 6 deletions
phoneuse/android_mcp.py
  View file @6cd9f89
@@ -928,7 +928,8 @@ def _build_server(): @@ -928,7 +928,8 @@ def _build_server():
928 928
929 Full pipeline (runs on-device): 929 Full pipeline (runs on-device):
930 text → edge-tts (MP3) → ffmpeg (16kHz PCM) → silk-wasm (SILK V3) 930 text → edge-tts (MP3) → ffmpeg (16kHz PCM) → silk-wasm (SILK V3)
931 - → SEND_VOICE_INTERNAL broadcast → WeChat DB insert (status=4, delivered) 931 + → SEND_VOICE_CGI (internal uploadvoice cgi): audio uploaded + delivered to
  932 + recipient + sent bubble on the sender side. Verified end-to-end.
932 933
933 **DO NOT implement SILK encoding yourself.** This tool handles the entire 934 **DO NOT implement SILK encoding yourself.** This tool handles the entire
934 pipeline. Just pass the text you want spoken. 935 pipeline. Just pass the text you want spoken.
phoneuse/wechat_hook_send.py
  View file @6cd9f89
@@ -472,7 +472,8 @@ def send_voice_tts_via_hook( @@ -472,7 +472,8 @@ def send_voice_tts_via_hook(
472 """Generate TTS from *text* and send as WeChat voice message to *wxid*. 472 """Generate TTS from *text* and send as WeChat voice message to *wxid*.
473 473
474 End-to-end: edge-tts → ffmpeg (16kHz PCM) → silk-wasm (SILK V3) → 474 End-to-end: edge-tts → ffmpeg (16kHz PCM) → silk-wasm (SILK V3) →
475 - wechathook SEND_VOICE_INTERNAL broadcast → WeChat DB insert, status=4. 475 + wechathook SEND_VOICE_CGI (internal uploadvoice cgi): uploads the audio + delivers
  476 + to the recipient AND inserts a sent bubble on the sender side. Verified end-to-end.
476 477
477 Args: 478 Args:
478 wxid: talker id (e.g. 'filehelper'). IMPORTANT: test only with 'filehelper'. 479 wxid: talker id (e.g. 'filehelper'). IMPORTANT: test only with 'filehelper'.
@@ -498,10 +499,11 @@ def send_voice_tts_via_hook( @@ -498,10 +499,11 @@ def send_voice_tts_via_hook(
498 except Exception as e: 499 except Exception as e:
499 return {"status": "error", "error": f"TTS/SILK generation failed: {e}"} 500 return {"status": "error", "error": f"TTS/SILK generation failed: {e}"}
500 501
501 - # Delegate staging + broadcast to send_voice_via_hook (it handles both  
502 - # on-device and off-device paths). Must call while tmp dir is still alive.  
503 - result = send_voice_via_hook(wxid=wxid, silk_path=silk_path,  
504 - duration_ms=duration_ms) 502 + # Delegate staging + broadcast to the fully-synthetic cgi sender (uploads audio
  503 + # inline via the internal uploadvoice cgi + delivers to the recipient + inserts a
  504 + # sent bubble). Must call while tmp dir is still alive.
  505 + result = send_voice_cgi_via_hook(wxid=wxid, silk_path=silk_path,
  506 + duration_ms=duration_ms)
505 result["tts_text"] = text[:100] 507 result["tts_text"] = text[:100]
506 result["tts_voice"] = voice 508 result["tts_voice"] = voice
507 return result 509 return result