- Issue created by @webbywe
- πΊπΈUnited States majorrobot
Thanks for this feature, @webbywe!
I reviewed the MR and tested locally. Everything worked locally.
I left a few comments and questions on the MR -- nothing big. Thank you again!
During development it was noticed that the tokens were large being logged which would increase cost potential during development along with potential inaccurate results.
Run the current slice1 migration and the complete HTML from the external page is sent to AI.
Allow for setting of the configuration from within a html_processor configuration for the AI plugin.
ai:
html_processor:
....
A few example of the configuration options would be added to the ai_migration_example migration yml.
The container would be specified as a string or array. This will allow specific regions of the html to be targeted and sent to the AI after processing. For arrays, each container filter would be processed and append results to the content string sent AI.
Utilize the Symfony HTML Sanitizer Component for the heavy lifting of sanitizing the html content to allow full flexibility to the developer to the options given from the component. It is based on the HTML Sanitizer W3C Standard Proposal.
https://symfony.com/packages/HTML%20Sanitizer
https://github.com/symfony/html-sanitizer
This will allow the developer further flexibility in targeting specific content to strip that would not be possible with the sanitizer.
Active
1.0
Code
Thanks for this feature, @webbywe!
I reviewed the MR and tested locally. Everything worked locally.
I left a few comments and questions on the MR -- nothing big. Thank you again!