Closed Mrkwtkr closed 8 months ago
@Mrkwtkr was this a non-English file, or did you set it to translate to a non-English language?
Can you post the full logs from your Logs tab?
Both. It's a non-English file. The audio is talked in Chinese. I set the Summary Language (Advanced)
to Chinese.
Here is the full logs:
@Mrkwtkr I've just shipped a much more robust makeParagraphs()
method that should handle Chinese text more gracefully. It also guarantees each paragraph can be no more than 1,800 characters, so it should prevent this error.
You can update your workflow by following these instructions: https://thomasjfrank.com/how-to-transcribe-audio-to-text-with-chatgpt-and-notion/#update
Thank you, I just confirmed that it has been fixed.
I was uploading a 30-minute audio file to my Google Drive. It seems like there is an issue with the workflow on rich text chunking, which should be less than 2000 characters.
I've tried twice with the same audio file. Each time, it did incur API costs (about $0.35 totally), but nothing was added to my Notion database.
Here is the details:
APIResponseError body failed validation: body.children[1].paragraph.rich_text[0].text.content.length should be ≤
2000
, instead was2039
. at null.buildRequestError (/pipedream/dist/code/4cf355a52ab0f9c275ba953eea42492276c4b796f961fdffefa87942b1ced4df/node_modules/.pnpm/@notionhq+client@2.2.13/node_modules/@notionhq/client/build/src/errors.js:162:16) at Client.request (/pipedream/dist/code/4cf355a52ab0f9c275ba953eea42492276c4b796f961fdffefa87942b1ced4df/node_modules/.pnpm/@notionhq+client@2.2.13/node_modules/@notionhq/client/build/src/Client.js:378:54) at process.processTicksAndRejections (node:internal/process/task_queues:95:5) at null.retry.retries (file:///pipedream/dist/code/4cf355a52ab0f9c275ba953eea42492276c4b796f961fdffefa87942b1ced4df/code/Notion-Voice-Notes.mjs:1747:23)