meeting-minutes: [BUG] There was an error when taking minutes, I am sure the model I run with ollama supports function calls.
There was an error when taking minutes, I am sure the model I run with ollama supports function calls.The following is my log output:
INFO: 127.0.0.1:64409 - "GET /get-summary/meeting-1746788343164 HTTP/1.1" 202 Accepted 2025-05-09 21:12:18,213 - DEBUG - [_trace.py:87] - receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Content-Type', b'application/json'), (b'Date', b'Fri, 09 May 2025 13:12:18 GMT'), (b'Transfer-Encoding', b'chunked')]) 2025-05-09 21:12:18,214 - INFO - [_client.py:1740] - HTTP Request: POST http://localhost:11434/v1/chat/completions "HTTP/1.1 200 OK" 2025-05-09 21:12:18,214 - DEBUG - [_trace.py:87] - receive_response_body.started request=<Request [b'POST']> 2025-05-09 21:12:18,214 - DEBUG - [_trace.py:87] - receive_response_body.complete 2025-05-09 21:12:18,215 - DEBUG - [_trace.py:87] - response_closed.started 2025-05-09 21:12:18,215 - DEBUG - [_trace.py:87] - response_closed.complete 2025-05-09 21:12:18,215 - DEBUG - [_base_client.py:1518] - HTTP Response: POST http://localhost:11434/v1/chat/completions "200 OK" Headers({'content-type': 'application/json', 'date': 'Fri, 09 May 2025 13:12:18 GMT', 'transfer-encoding': 'chunked'}) 2025-05-09 21:12:18,215 - DEBUG - [_base_client.py:1526] - request_id: None 2025-05-09 21:12:18,220 - DEBUG - [_base_client.py:482] - Request options: {'method': 'post', 'url': '/chat/completions', 'files': None, 'idempotency_key': 'stainless-python-retry-8254cbc5-089c-4786-a28e-357741366236', 'json_data': {'messages': [{'role': 'user', 'content': "Given the following meeting transcript chunk, extract the relevant information according to the required JSON structure. If a specific section (like Critical Deadlines) has no relevant information in this chunk, return an empty list for its 'blocks'. Ensure the output is only the JSON data.\n\n Transcript Chunk:\n ---\n 這是 這是 龍 disks 這是 這是 龍 荒诞 我起了他的狼蛋 还说呢\n救活大人\n我好奇化 因我本所 徒龙了\n我说 跟上了就和他 得得死\n小龍 小龍 小龍\n还是 从他们的屠龙开始?\n(知名的海蜜) (如果我化作苦雨) (黑夜去獨當)\n如果能再理解也夠蹉跎\n今天在七阳和彩虹上的日子里,你会想起哪一天的日子。 我不喜欢夕阳和彩虹 黑夜不好吗 只有黑夜才是永恒的\n大人的腰力和腰力都不足 大人的腰力和腰力都不足 大人的腰力和腰力都不足 大人的腰力和腰力都不足 大人的腰力和腰力都不足\n你恢复了 龙丹的灵力修补了我的药丹 我们的下了\n辛苦了 大人 那咱们能不能要一个角力啊 可\n(字幕君:我看你不太會寫字幕)\n给嘉兰收好了\n我 为何擅自去屠龙 加蓝喜欢强者\n任何挡在九霍大人面前的 都得死 那若有一天 我不强了\n你是不是也会见我当做蝴蝶一样 让我死在最美的那一刻\n(字幕製作:貝爾)\n家兰的酒货大人怎么会死呢\n ---\n "}, {'role': 'assistant', 'content': '', 'tool_calls': [{'id': 'call_m1aj8gg3', 'type': 'function', 'function': {'name': 'final_result', 'arguments': '{"ClosingRemarks":"","CriticalDeadlines":[],"ImmediateActionItems":[],"KeyItemsDecisions":[],"MeetingName":"Unknown Meeting Transcript","NextSteps":[],"OtherImportantPoints":"","SectionSummary":""}'}}]}, {'role': 'tool', 'tool_call_id': 'call_m1aj8gg3', 'content': '7 validation errors: [\n {\n "type": "model_type",\n "loc": [\n "SectionSummary"\n ],\n "msg": "Input should be an object",\n "input": ""\n },\n {\n "type": "model_type",\n "loc": [\n "CriticalDeadlines"\n ],\n "msg": "Input should be an object",\n "input": []\n },\n {\n "type": "model_type",\n "loc": [\n "KeyItemsDecisions"\n ],\n "msg": "Input should be an object",\n "input": []\n },\n {\n "type": "model_type",\n "loc": [\n "ImmediateActionItems"\n ],\n "msg": "Input should be an object",\n "input": []\n },\n {\n "type": "model_type",\n "loc": [\n "NextSteps"\n ],\n "msg": "Input should be an object",\n "input": []\n },\n {\n "type": "model_type",\n "loc": [\n "OtherImportantPoints"\n ],\n "msg": "Input should be an object",\n "input": ""\n },\n {\n "type": "model_type",\n "loc": [\n "ClosingRemarks"\n ],\n "msg": "Input should be an object",\n "input": ""\n }\n]\n\nFix the errors and try again.'}, {'role': 'assistant', 'content': 'It seems like there are some issues with the input formats. Please provide the meeting transcript data in a proper JSON structure as per the meeting summary schema so that I could extract the relevant information efficiently. Here is an example of how you should format the data:\n\n```json\n{\n "MeetingName": "Dragon Hunt Discussion",\n "SectionSummary": "Participants speculate on dragon hunting strategies and outcomes.",\n "CriticalDeadlines": {\n "blocks": []\n },\n "KeyItemsDecisions": {\n "blocks": []\n },\n "ImmediateActionItems": {\n "blocks": []\n },\n "NextSteps": {\n "blocks": []\n },\n "OtherImportantPoints": {\n "blocks": []\n },\n "ClosingRemarks": ""\n}\n```\n\nPlease adjust the content inside the JSON object according to the discussion in the transcript.'}, {'role': 'user', 'content': 'Plain text responses are not permitted, please call one of the functions instead.\n\nFix the errors and try again.'}, {'role': 'assistant', 'content': '', 'tool_calls': [{'id': 'call_zte9spd8', 'type': 'function', 'function': {'name': 'final_result', 'arguments': '{"ClosingRemarks":"","CriticalDeadlines":{"blocks":[]},"ImmediateActionItems":{"blocks":[]},"KeyItemsDecisions":{"blocks":[]},"MeetingName":"Dragon Hunt Discussion","NextSteps":{"blocks":[]},"OtherImportantPoints":{"blocks":[]},"SectionSummary":"Participants debated strategies for hunting dragons and the philosophical implications of such actions."}'}}]}, {'role': 'tool', 'tool_call_id': 'call_zte9spd8', 'content': '7 validation errors: [\n {\n "type": "model_type",\n "loc": [\n "SectionSummary"\n ],\n "msg": "Input should be an object",\n "input": "Participants debated strategies for hunting dragons and the philosophical implications of such actions."\n },\n {\n "type": "missing",\n "loc": [\n "CriticalDeadlines",\n "title"\n ],\n "msg": "Field required",\n "input": {\n "blocks": []\n }\n },\n {\n "type": "missing",\n "loc": [\n "KeyItemsDecisions",\n "title"\n ],\n "msg": "Field required",\n "input": {\n "blocks": []\n }\n },\n {\n "type": "missing",\n "loc": [\n "ImmediateActionItems",\n "title"\n ],\n "msg": "Field required",\n "input": {\n "blocks": []\n }\n },\n {\n "type": "missing",\n "loc": [\n "NextSteps",\n "title"\n ],\n "msg": "Field required",\n "input": {\n "blocks": []\n }\n },\n {\n "type": "missing",\n "loc": [\n "OtherImportantPoints",\n "title"\n ],\n "msg": "Field required",\n "input": {\n "blocks": []\n }\n },\n {\n "type": "model_type",\n "loc": [\n "ClosingRemarks"\n ],\n "msg": "Input should be an object",\n "input": ""\n }\n]\n\nFix the errors and try again.'}, {'role': 'assistant', 'content': '', 'tool_calls': [{'id': 'call_mskm7jua', 'type': 'function', 'function': {'name': 'final_result', 'arguments': '{"ClosingRemarks":"","CriticalDeadlines":{"blocks":[],"title":"Critical Deadlines"},"ImmediateActionItems":{"blocks":[],"title":"Immediate Action Items"},"KeyItemsDecisions":{"blocks":[],"title":"Key Items \\u0026 Decisions"},"MeetingName":"Dragon Hunting Debate Meeting","NextSteps":{"blocks":[],"title":"Next Steps"},"OtherImportantPoints":{"blocks":[],"title":"Other Important Points"},"SectionSummary":"Participants debated strategies for hunting dragons and the philosophical implications of such actions."}'}}]}, {'role': 'tool', 'tool_call_id': 'call_mskm7jua', 'content': '2 validation errors: [\n {\n "type": "model_type",\n "loc": [\n "SectionSummary"\n ],\n "msg": "Input should be an object",\n "input": "Participants debated strategies for hunting dragons and the philosophical implications of such actions."\n },\n {\n "type": "model_type",\n "loc": [\n "ClosingRemarks"\n ],\n "msg": "Input should be an object",\n "input": ""\n }\n]\n\nFix the errors and try again.'}, {'role': 'assistant', 'content': 'To provide you with a meeting summary, we will need more specific information under each section like SectionSummary, CriticalDeadlines, KeyItemsDecisions, ImmediateActionItems, NextSteps, OtherImportantPoints, and ClosingRemarks. Since some of these sections were not provided in the initial input, let\'s assume some hypothetical content for a clear summary response.\n\nHere\'s an example on how to structure your input data:\n\n```json\n{\n "MeetingName": "Dragon Hunt Discussion",\n "SectionSummary": {\n "title": "Summarized Discussion",\n "blocks": [\n {\n "color": "green",\n "content": "Participants debated strategies for hunting dragons and the philosophical implications of such actions.",\n "id": "01"\n }\n ]\n },\n "CriticalDeadlines": {\n "title": "Due Dates",\n "blocks": [\n {\n "color": "blue",\n "content": "Submit dragon hunting plan by next meeting.",\n "id": "02"\n }\n ]\n },\n "KeyItemsDecisions": {\n "title": "Milestones Reached",\n "blocks": [\n {\n "color": "purple",\n "content": "Agreed on the importance of considering animal rights before hunting.",\n "id": "03"\n }\n ]\n },\n "ImmediateActionItems": {\n "title": "Pending Actions",\n "blocks": [\n {\n "color": "orange",\n "content": "Research potential impacts of dragon hunting on the ecosystem.",\n "id": "04"\n }\n ]\n },\n "NextSteps": {\n "title": "Next Steps Plan",\n "blocks": [\n {\n "color": "green",\n "content": "Plan a mock hunting situation to test our strategies.",\n "id": "05"\n }\n ]\n },\n "OtherImportantPoints": {\n "title": "Miscellaneous Notes",\n "blocks": [\n {\n "color": "gray",\n "content": "Discussion pointed out the need for more experienced hunters on the team.",\n "id": "06"\n }\n ]\n },\n "ClosingRemarks": {\n "title": "Meeting Final Thoughts",\n "blocks": [\n {\n "color": "red",\n "content": "Remember, ethics should not be compromised while achieving hunting goals.",\n "id": "07"\n }\n ]\n }\n}\n```\n\nPlease replace these placeholders with actual data from your meeting transcript and provide it as input.'}, {'role': 'user', 'content': 'Plain text responses are not permitted, please call one of the functions instead.\n\nFix the errors and try again.'}], 'model': 'qwen2-fun:7b-q5', 'n': 1, 'parallel_tool_calls': True, 'stream': False, 'tool_choice': 'required', 'tools': [{'type': 'function', 'function': {'name': 'final_result', 'description': 'Represents the meeting summary response based on a section of the transcript', 'parameters': {'$defs': {'Block': {'description': 'Represents a block of content in a section', 'properties': {'id': {'title': 'Id', 'type': 'string'}, 'type': {'title': 'Type', 'type': 'string'}, 'content': {'title': 'Content', 'type': 'string'}, 'color': {'title': 'Color', 'type': 'string'}}, 'required': ['id', 'type', 'content', 'color'], 'title': 'Block', 'type': 'object'}, 'Section': {'description': 'Represents a section in the meeting summary', 'properties': {'title': {'title': 'Title', 'type': 'string'}, 'blocks': {'items': {'$ref': '#/$defs/Block'}, 'title': 'Blocks', 'type': 'array'}}, 'required': ['title', 'blocks'], 'title': 'Section', 'type': 'object'}}, 'properties': {'MeetingName': {'title': 'Meetingname', 'type': 'string'}, 'SectionSummary': {'$ref': '#/$defs/Section'}, 'CriticalDeadlines': {'$ref': '#/$defs/Section'}, 'KeyItemsDecisions': {'$ref': '#/$defs/Section'}, 'ImmediateActionItems': {'$ref': '#/$defs/Section'}, 'NextSteps': {'$ref': '#/$defs/Section'}, 'OtherImportantPoints': {'$ref': '#/$defs/Section'}, 'ClosingRemarks': {'$ref': '#/$defs/Section'}}, 'required': ['MeetingName', 'SectionSummary', 'CriticalDeadlines', 'KeyItemsDecisions', 'ImmediateActionItems', 'NextSteps', 'OtherImportantPoints', 'ClosingRemarks'], 'title': 'SummaryResponse', 'type': 'object'}}}]}} 2025-05-09 21:12:18,220 - DEBUG - [_base_client.py:1480] - Sending HTTP Request: POST http://localhost:11434/v1/chat/completions 2025-05-09 21:12:18,220 - DEBUG - [_trace.py:87] - send_request_headers.started request=<Request [b'POST']> 2025-05-09 21:12:18,221 - DEBUG - [_trace.py:87] - send_request_headers.complete 2025-05-09 21:12:18,221 - DEBUG - [_trace.py:87] - send_request_body.started request=<Request [b'POST']> 2025-05-09 21:12:18,221 - DEBUG - [_trace.py:87] - send_request_body.complete 2025-05-09 21:12:18,221 - DEBUG - [_trace.py:87] - receive_response_headers.started request=<Request [b'POST']> 2025-05-09 21:12:18,783 - DEBUG - [core.py:104] - executing <function connect.<locals>.connector at 0x107e510d0> 2025-05-09 21:12:18,784 - DEBUG - [core.py:106] - operation <function connect.<locals>.connector at 0x107e510d0> completed 2025-05-09 21:12:18,784 - DEBUG - [core.py:104] - executing functools.partial(<built-in method execute of sqlite3.Connection object at 0x107a66e40>, '\n SELECT t.*, p.status, p.result \n FROM transcript_chunks t \n JOIN summary_processes p ON t.meeting_id = p.meeting_id \n WHERE t.meeting_id = ?\n ', ('meeting-1746788343164',)) 2025-05-09 21:12:18,784 - DEBUG - [core.py:106] - operation functools.partial(<built-in method execute of sqlite3.Connection object at 0x107a66e40>, '\n SELECT t.*, p.status, p.result \n FROM transcript_chunks t \n JOIN summary_processes p ON t.meeting_id = p.meeting_id \n WHERE t.meeting_id = ?\n ', ('meeting-1746788343164',)) completed 2025-05-09 21:12:18,784 - DEBUG - [core.py:104] - executing functools.partial(<built-in method fetchone of sqlite3.Cursor object at 0x107e597a0>) 2025-05-09 21:12:18,784 - DEBUG - [core.py:106] - operation functools.partial(<built-in method fetchone of sqlite3.Cursor object at 0x107e597a0>) completed 2025-05-09 21:12:18,785 - DEBUG - [core.py:104] - executing functools.partial(<built-in method close of sqlite3.Cursor object at 0x107e597a0>) 2025-05-09 21:12:18,785 - DEBUG - [core.py:106] - operation functools.partial(<built-in method close of sqlite3.Cursor object at 0x107e597a0>) completed 2025-05-09 21:12:18,785 - DEBUG - [core.py:104] - executing functools.partial(<built-in method close of sqlite3.Connection object at 0x107a66e40>) 2025-05-09 21:12:18,785 - DEBUG - [core.py:106] - operation functools.partial(<built-in method close of sqlite3.Connection object at 0x107a66e40>) completed INFO: 127.0.0.1:64444 - "GET /get-summary/meeting-1746788343164 HTTP/1.1" 202 Accepted 2025-05-09 21:12:24,783 - DEBUG - [core.py:104] - executing <function connect.<locals>.connector at 0x107e510d0> 2025-05-09 21:12:24,784 - DEBUG - [core.py:106] - operation <function connect.<locals>.connector at 0x107e510d0> completed 2025-05-09 21:12:24,785 - DEBUG - [core.py:104] - executing functools.partial(<built-in method execute of sqlite3.Connection object at 0x107a66d50>, '\n SELECT t.*, p.status, p.result \n FROM transcript_chunks t \n JOIN summary_processes p ON t.meeting_id = p.meeting_id \n WHERE t.meeting_id = ?\n ', ('meeting-1746788343164',)) 2025-05-09 21:12:24,786 - DEBUG - [core.py:106] - operation functools.partial(<built-in method execute of sqlite3.Connection object at 0x107a66d50>, '\n SELECT t.*, p.status, p.result \n FROM transcript_chunks t \n JOIN summary_processes p ON t.meeting_id = p.meeting_id \n WHERE t.meeting_id = ?\n ', ('meeting-1746788343164',)) completed 2025-05-09 21:12:24,787 - DEBUG - [core.py:104] - executing functools.partial(<built-in method fetchone of sqlite3.Cursor object at 0x107e59810>) 2025-05-09 21:12:24,787 - DEBUG - [core.py:106] - operation functools.partial(<built-in method fetchone of sqlite3.Cursor object at 0x107e59810>) completed 2025-05-09 21:12:24,787 - DEBUG - [core.py:104] - executing functools.partial(<built-in method close of sqlite3.Cursor object at 0x107e59810>) 2025-05-09 21:12:24,787 - DEBUG - [core.py:106] - operation functools.partial(<built-in method close of sqlite3.Cursor object at 0x107e59810>) completed 2025-05-09 21:12:24,787 - DEBUG - [core.py:104] - executing functools.partial(<built-in method close of sqlite3.Connection object at 0x107a66d50>) 2025-05-09 21:12:24,788 - DEBUG - [core.py:106] - operation functools.partial(<built-in method close of sqlite3.Connection object at 0x107a66d50>) completed INFO: 127.0.0.1:64464 - "GET /get-summary/meeting-1746788343164 HTTP/1.1" 202 Accepted 2025-05-09 21:12:28,215 - DEBUG - [_trace.py:87] - receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Content-Type', b'application/json'), (b'Date', b'Fri, 09 May 2025 13:12:28 GMT'), (b'Content-Length', b'683')]) 2025-05-09 21:12:28,215 - INFO - [_client.py:1740] - HTTP Request: POST http://localhost:11434/v1/chat/completions "HTTP/1.1 200 OK" 2025-05-09 21:12:28,215 - DEBUG - [_trace.py:87] - receive_response_body.started request=<Request [b'POST']> 2025-05-09 21:12:28,216 - DEBUG - [_trace.py:87] - receive_response_body.complete 2025-05-09 21:12:28,216 - DEBUG - [_trace.py:87] - response_closed.started 2025-05-09 21:12:28,216 - DEBUG - [_trace.py:87] - response_closed.complete 2025-05-09 21:12:28,216 - DEBUG - [_base_client.py:1518] - HTTP Response: POST http://localhost:11434/v1/chat/completions "200 OK" Headers({'content-type': 'application/json', 'date': 'Fri, 09 May 2025 13:12:28 GMT', 'content-length': '683'}) 2025-05-09 21:12:28,216 - DEBUG - [_base_client.py:1526] - request_id: None 2025-05-09 21:12:28,217 - ERROR - [transcript_processor.py:157] - Error processing chunk 1: Exceeded maximum retries (5) for result validation Traceback (most recent call last): File "/opt/homebrew/Cellar/meetily-backend/0.0.4/backend/app/transcript_processor.py", line 133, in process_transcript summary_result = await agent.run( File "/opt/homebrew/Cellar/meetily-backend/0.0.4/backend/venv/lib/python3.9/site-packages/pydantic_ai/agent.py", line 309, in run final_result, tool_responses = await self._handle_model_response( File "/opt/homebrew/Cellar/meetily-backend/0.0.4/backend/venv/lib/python3.9/site-packages/pydantic_ai/agent.py", line 1092, in _handle_model_response return await self._handle_text_response(text, run_context, result_schema) File "/opt/homebrew/Cellar/meetily-backend/0.0.4/backend/venv/lib/python3.9/site-packages/pydantic_ai/agent.py", line 1110, in _handle_text_response self._incr_result_retry(run_context) File "/opt/homebrew/Cellar/meetily-backend/0.0.4/backend/venv/lib/python3.9/site-packages/pydantic_ai/agent.py", line 1278, in _incr_result_retry raise exceptions.UnexpectedModelBehavior( pydantic_ai.exceptions.UnexpectedModelBehavior: Exceeded maximum retries (5) for result validation 2025-05-09 21:12:28,220 - INFO - [transcript_processor.py:159] - Finished processing all 1 chunks.
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Comments: 15
前后端分离的,你最好安装在本地。他的模型对中文识别不是很好。 加参数 -l zh实现中文输出
This is an open source speech recognition model I found, called PaddleSpeech, I hope it can be used. @sujithatzackriya
Can deepseek be added to the way of generating summaries? @sujithatzackriya