bug/Local API Error:Input should be 'by_title'

[tangkunyin]

Describe the bug

Basic chunking strategy not recognized with a markdown file, and it seems only valid with 'by_title'

How did the error occurred

I have checked the official documentation, and found by_page and by_similarity are only available in Unstructured API and Platform.

I used the https://github.com/Unstructured-IO/unstructured-api @0.0.82 source code and built a local api service, everything was ok except the chunkingStrategy with basic

Loader Options

{
  apiKey: '',
  apiUrl: 'http://127.0.0.1:8000/general/v0/general',
  strategy: 'auto',
  chunkingStrategy: 'basic'
}

Response by API

Failed to partition file /tmp/uploads/ec20069fa25bd07537eae8559fde792dff9d944288bc3c2ecbc8785ea98bc5c8.md with error 422 and message {"detail":[{"type":"literal_error","loc":["body","chunking_strategy"],"msg":"Input should be 'by_title'","input":"basic","ctx":{"expected":"'by_title'"}}]}

Expected behavior
No errors and file parsed correctly.

Any feedback is greatful. 🙏 😊 🙌

[tangkunyin]

Unfortunately, I found a issue, it said only by_title supported with Self-hosting version.

It would be better upgrading the official document and making a table list that make clarify the difference between the self-hosted version and the public cloud api version.

If not, we won't know where the real problem it is.