Allow disable frontend auto metadata extraction feature via config

[t83714]

Allow disable frontend auto metadata extraction feature via config

We have frontend (in-browser) metadata extraction pipelines here: https://github.com/magda-io/magda/tree/main/magda-web-client/src/Components/Dataset/MetadataExtraction

Those pipelines (running in web workers) were designed to auto extract metadata (e.g. auto generate keywords & title, spatial coverage etc) from various file formats (e.g. PDF, words, excel, CSV etc)

However, when a large files are supplied (e.g. 200MB CSV with 1M rows), the processing could take a long time.

Considering we might have better approach, e.g.:

• leveraging LLM to generate metatdata (e.g. write summary description)
• allow users to trigger the process via UI interaction around the specific field rather part of file uploading process

We want:

• by default, disable the feature since v5 release
• allow users to enable the feature via Helm chart config