Closed yekalkan closed 1 year ago
I've made a progress in the branch https://github.com/abpframework/abp/tree/alper/elastic-search-with-scriban but this is a problematic situation. The Scriban render is made on the web layer because , user preferences are grabbed from the URL. But in the startup, while all docs are being cached to ElasticSearch, we don't know the variables. One way is caching all possible cartesian of variables with different cache keys. See this URL as an example https://docs.abp.io/en/commercial/latest/getting-started?UI=MVC&DB=EF&Tiered=No
this project needs a radical refactor! elastic search section should move to appservice layer. also, rendering MD to HTML should be done in appservice. now, elastic search indexing is done in domain and markdown converters work in Web layer.
Search is performed before documents is converted to plain text from markdown format and processed by scriban.
related https://github.com/abpframework/abp/pull/2901 related https://github.com/abpframework/abp/issues/1621