myagent-ai 1.23.2 → 1.23.4

package/core/stt.py

@@ -37,12 +37,21 @@ def _convert_to_wav(audio_data: bytes, audio_format: Optional[str] = None) -> by
         from pydub import AudioSegment
         audio_buf = io.BytesIO(audio_data)
         seg = AudioSegment.from_file(audio_buf, format=audio_format or "webm")
+        # [v1.23.2] 检查音频时长，过短直接返回原始数据
+        if seg.duration_seconds < 0.1:
+            logger.debug(f"音频过短 ({seg.duration_seconds:.2f}s)，跳过转换")
+            return audio_data
         seg = seg.set_channels(1).set_frame_rate(16000).set_sample_width(2)
         wav_buf = io.BytesIO()
         seg.export(wav_buf, format="wav")
         wav_buf.seek(0)
         return wav_buf.read()
-    except Exception:
+    except Exception as e:
+        import shutil
+        if not shutil.which("ffmpeg"):
+            logger.warning(f"pydub 转换失败且缺少 ffmpeg: {e}")
+        else:
+            logger.warning(f"pydub 音频转换失败: {e}")
         return audio_data
 
 
@@ -68,10 +77,15 @@ async def _stt_sensevoice(audio_data: bytes, audio_format: Optional[str] = None)
             )
             logger.info("SenseVoice 模型已加载 (iic/SenseVoiceSmall, CPU)")
 
-        # 转换为 16kHz WAV
+        # [v1.23.2] 增强: pydub 转换失败记录警告、WAV 头验证、音频长度检查
         wav_data = _convert_to_wav(audio_data, audio_format)
         wav_path = f"/tmp/myagent_stt_{id(audio_data) % 100000}.wav"
         try:
+            # 验证 WAV 文件头 (RIFF....WAVE)
+            if len(wav_data) < 44 or wav_data[:4] != b'RIFF' or wav_data[8:12] != b'WAVE':
+                logger.warning(f"SenseVoice 跳过: 无效 WAV 数据 (size={len(wav_data)}, header={wav_data[:12].hex()})")
+                return None
+
             with open(wav_path, 'wb') as f:
                 f.write(wav_data)
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "myagent-ai",
-  "version": "1.23.2",
+  "version": "1.23.4",
   "description": "本地桌面端执行型AI助手 - Open Interpreter 风格 | Local Desktop Execution-Oriented AI Assistant",
   "main": "main.py",
   "bin": {

package/scripts/cli.py

@@ -402,8 +402,8 @@ async def cmd_docx_create(args):
     a = p.parse_args(args)
 
     content = _parse_json_arg(a.content)
-    from skills.docx_skill import DocxCreateSkill
-    skill = DocxCreateSkill()
+    from skills.docx_skill import DOCXCreateSkill
+    skill = DOCXCreateSkill()
     result = await skill.execute(content=content, title=a.title, output_path=a.output)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -418,8 +418,8 @@ async def cmd_docx_read(args):
     p.add_argument("path", help="Word 文件路径 (.docx)")
     a = p.parse_args(args)
 
-    from skills.docx_skill import DocxReadSkill
-    skill = DocxReadSkill()
+    from skills.docx_skill import DOCXReadSkill
+    skill = DOCXReadSkill()
     result = await skill.execute(path=a.path)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -435,8 +435,8 @@ async def cmd_xlsx_create(args):
     a = p.parse_args(args)
 
     sheets = _parse_json_arg(a.sheets)
-    from skills.xlsx_skill import XlsxCreateSkill
-    skill = XlsxCreateSkill()
+    from skills.xlsx_skill import XLSXCreateSkill
+    skill = XLSXCreateSkill()
     result = await skill.execute(sheets=sheets, title=a.title, output_path=a.output)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -452,8 +452,8 @@ async def cmd_xlsx_read(args):
     p.add_argument("--sheet", default="", help="工作表名称 (默认全部)")
     a = p.parse_args(args)
 
-    from skills.xlsx_skill import XlsxReadSkill
-    skill = XlsxReadSkill()
+    from skills.xlsx_skill import XLSXReadSkill
+    skill = XLSXReadSkill()
     result = await skill.execute(path=a.path, sheet_name=a.sheet)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -471,8 +471,8 @@ async def cmd_xlsx_edit(args):
     a = p.parse_args(args)
 
     data = _parse_json_arg(a.data)
-    from skills.xlsx_skill import XlsxEditSkill
-    skill = XlsxEditSkill()
+    from skills.xlsx_skill import XLSXEditSkill
+    skill = XLSXEditSkill()
     result = await skill.execute(path=a.path, action=a.action, data=data, sheet_name=a.sheet)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -488,8 +488,8 @@ async def cmd_ppt_create(args):
     a = p.parse_args(args)
 
     slides = _parse_json_arg(a.slides)
-    from skills.ppt_skill import PptCreateSkill
-    skill = PptCreateSkill()
+    from skills.ppt_skill import PPTCreateSkill
+    skill = PPTCreateSkill()
     result = await skill.execute(slides=slides, theme=a.theme, output_path=a.output)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -504,8 +504,8 @@ async def cmd_ppt_read(args):
     p.add_argument("path", help="PPT 文件路径 (.pptx)")
     a = p.parse_args(args)
 
-    from skills.ppt_skill import PptReadSkill
-    skill = PptReadSkill()
+    from skills.ppt_skill import PPTReadSkill
+    skill = PPTReadSkill()
     result = await skill.execute(path=a.path)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -521,8 +521,8 @@ async def cmd_pdf_create(args):
     a = p.parse_args(args)
 
     content = _parse_json_arg(a.content)
-    from skills.pdf_skill import PdfCreateSkill
-    skill = PdfCreateSkill()
+    from skills.pdf_skill import PDFCreateSkill
+    skill = PDFCreateSkill()
     result = await skill.execute(content=content, palette=a.palette, output_path=a.output)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})
@@ -539,8 +539,8 @@ async def cmd_pdf_read(args):
     p.add_argument("--end", type=int, default=0, help="结束页 (默认全部)")
     a = p.parse_args(args)
 
-    from skills.pdf_skill import PdfReadSkill
-    skill = PdfReadSkill()
+    from skills.pdf_skill import PDFReadSkill
+    skill = PDFReadSkill()
     result = await skill.execute(path=a.path, start_page=a.start, end_page=a.end)
     _print_result({"success": result.success, "message": result.message,
                    "data": result.data, "error": result.error})

package/web/api_server.py

@@ -2497,36 +2497,57 @@ window.addEventListener('beforeunload', function() {{
                     logger.info("SenseVoice 模型已加载 (iic/SenseVoiceSmall, CPU)")
 
                 # SenseVoice 接受 16kHz WAV
+                # [v1.23.2] 增强: pydub 转换失败时记录警告、验证 WAV 头、检查音频长度
                 wav_path = f"/tmp/myagent_stt_{id(audio_data) % 100000}.wav"
                 wav_buf = io.BytesIO()
+                pydub_ok = False
                 try:
                     from pydub import AudioSegment
                     audio_buf = io.BytesIO(audio_data)
                     seg = AudioSegment.from_file(audio_buf, format=audio_format or "webm")
-                    seg = seg.set_channels(1).set_frame_rate(16000).set_sample_width(2)
-                    seg.export(wav_buf, format="wav")
-                except Exception:
+                    # 检查音频时长，过短直接跳过
+                    if seg.duration_seconds < 0.1:
+                        logger.debug(f"SenseVoice 跳过: 音频过短 ({seg.duration_seconds:.2f}s)")
+                    else:
+                        seg = seg.set_channels(1).set_frame_rate(16000).set_sample_width(2)
+                        seg.export(wav_buf, format="wav")
+                        pydub_ok = True
+                except Exception as conv_err:
+                    import shutil
+                    if not shutil.which("ffmpeg"):
+                        logger.warning(f"pydub 转换失败且缺少 ffmpeg: {conv_err}. 安装: sudo apt install ffmpeg")
+                    else:
+                        logger.warning(f"pydub 音频转换失败: {conv_err}")
+
+                if not pydub_ok:
                     wav_buf = io.BytesIO(audio_data)
+
                 wav_buf.seek(0)
-                with open(wav_path, 'wb') as f:
-                    f.write(wav_buf.read())
-
-                # SenseVoice 推理
-                res = sv_model.generate(input=wav_path, cache={},
-                                        language="auto",  # 自动检测语言
-                                        use_itn=True,     # 逆文本标准化（数字/日期等）
-                                        batch_size_s=300)
-                if res and len(res) > 0 and len(res[0]) > 0:
-                    text = res[0][0]["text"] if isinstance(res[0][0], dict) else str(res[0][0])
-                    # SenseVoice 可能输出带 <|zh|><|en|><|EMO|> 等特殊 token，清理掉
-                    import re
-                    text = re.sub(r'<\|[^|]+\|>', '', text).strip()
-                    if text:
-                        try:
-                            os.remove(wav_path)
-                        except Exception:
-                            pass
-                        return web.json_response({"text": text, "engine": "sensevoice"})
+                wav_bytes = wav_buf.read()
+
+                # 验证 WAV 文件头 (RIFF....WAVE)
+                if len(wav_bytes) < 44 or wav_bytes[:4] != b'RIFF' or wav_bytes[8:12] != b'WAVE':
+                    logger.warning(f"SenseVoice 跳过: 无效 WAV 数据 (size={len(wav_bytes)}, header={wav_bytes[:12].hex()})")
+                else:
+                    with open(wav_path, 'wb') as f:
+                        f.write(wav_bytes)
+
+                    # SenseVoice 推理
+                    res = sv_model.generate(input=wav_path, cache={},
+                                            language="auto",  # 自动检测语言
+                                            use_itn=True,     # 逆文本标准化（数字/日期等）
+                                            batch_size_s=300)
+                    if res and len(res) > 0 and len(res[0]) > 0:
+                        text = res[0][0]["text"] if isinstance(res[0][0], dict) else str(res[0][0])
+                        # SenseVoice 可能输出带 <|zh|><|en|><|EMO|> 等特殊 token，清理掉
+                        import re
+                        text = re.sub(r'<\|[^|]+\|>', '', text).strip()
+                        if text:
+                            try:
+                                os.remove(wav_path)
+                            except Exception:
+                                pass
+                            return web.json_response({"text": text, "engine": "sensevoice"})
                 try:
                     os.remove(wav_path)
                 except Exception: