Improvement of the search engine functionality on the NFDI4BioImage training website

SeverusYixin commented 1 month ago

Our current search engine on our NFDI4BioImage training website performs ineffectually and is not user friendly enough. https://nfdi4bioimage.github.io/training

There are currently the following issues:

Search results and interface not accurate and clear enough
Search interface not user-friendly
Missing features like advanced search

Possible improvement suggestions:

Enhance the search algorithm:

Possible selections:

I. Elasticsearch + Angular
II. Algolia + Angular
https://www.algolia.com/competitors/compare-algolia-vs-elasticsearch

Perhaps add relevant natural language processing queries

Add advanced search options:

For search engines:

I. Enable Boolean search operators (AND, OR, NOT) for more precise queries.
II. Perhaps with filters such as categories, content types and Tags so that users can refine their search results.

For the search interface:

I. Redesign the search bar to make it more prominent and user-friendly.

For search results:

I. Perhaps add options such as "content type" to search results
II. It would be great to be able to show "Abstract" in the search results.
III. In the future, it might also be nice to add a release date to the data in the database

Literature search engine display might be a good idea:
    https://ieeexplore.ieee.org/Xplore/home.jsp
    https://dl.acm.org/

In the future, when the search engine is built, it would be good to collect user feedback and update it on a regular basis.

For example, the collected user feedback can be analysed regularly through machine learning and other methods to select the most needed updates.

haesleinhuepf commented 1 month ago

Idea: we could search in the YML files, instead of the rendered HTML pages.

Alternatively: We could create 1 html page for each entry in the yml files...

SeverusYixin commented 1 month ago

Idea: we could search in the YML files, instead of the rendered HTML pages.

Alternatively: We could create 1 html page for each entry in the yml files...

It seems to be possible to do this by setting each YML file to the corresponding search tag

SeverusYixin commented 1 month ago

@haesleinhuepf It would be so nice if you could show me which folder or file is for the current search engine.

haesleinhuepf commented 1 month ago

The current search engine can be modified like documented here: https://www.sphinx-doc.org/en/master/usage/advanced/websupport/searchadapters.html

Also consider setting up a new website with a better search, that is not built on Jupyterbook.

SeverusYixin commented 1 month ago

The current search engine can be modified like documented here: https://www.sphinx-doc.org/en/master/usage/advanced/websupport/searchadapters.html

Also consider setting up a new website with a better search, that is not built on Jupyterbook.

I'm partial to setting up a new site, but I'd like to see the code you wrote for our current search engine, and by that I mean the folder or file where that code resides ：）

haesleinhuepf commented 1 month ago

I mean the folder or file where that code resides

I do not know. It's not my code. The search is part of sphinx, which is part of jupyter-book, which we are using.

SeverusYixin commented 1 month ago

Ok, I got it

SeverusYixin commented 1 month ago

Hi @haesleinhuepf ,look at this "III. In the future, it might also be nice to add a release date to the data in the database", maybe they want to discuss the similar things. :)

haesleinhuepf commented 1 month ago

"III. In the future, it might also be nice to add a release date to the data in the database",

Can you give a source (link) for this citation?

SeverusYixin commented 1 month ago

Our current search engine on our NFDI4BioImage training website performs ineffectually and is not user friendly enough. https://nfdi4bioimage.github.io/training

There are currently the following issues:

Search results and interface not accurate and clear enough

Search interface not user-friendly

Missing features like advanced search

Possible improvement suggestions:

Enhance the search algorithm:
Possible selections:
I. Elasticsearch + Angular
II. Algolia + Angular
https://www.algolia.com/competitors/compare-algolia-vs-elasticsearch
Perhaps add relevant natural language processing queries
Add advanced search options:
For search engines:
I. Enable Boolean search operators (AND, OR, NOT) for more precise queries.
II. Perhaps with filters such as categories, content types and Tags so that users can refine their search results.
For the search interface:
I. Redesign the search bar to make it more prominent and user-friendly.
For search results:
I. Perhaps add options such as "content type" to search results
II. It would be great to be able to show "Abstract" in the search results.
III. In the future, it might also be nice to add a release date to the data in the database

Literature search engine display might be a good idea:
  https://ieeexplore.ieee.org/Xplore/home.jsp
  https://dl.acm.org/
In the future, when the search engine is built, it would be good to collect user feedback and update it on a regular basis.

For example, the collected user feedback can be analysed regularly through machine learning and other methods to select the most needed updates.

In my description: "For search results:

I. Perhaps add options such as "content type" to search results II. It would be great to be able to show "Abstract" in the search results. III. In the future, it might also be nice to add a release date to the data in the database

Literature search engine display might be a good idea: https://ieeexplore.ieee.org/Xplore/home.jsp https://dl.acm.org/"

NFDI4BIOIMAGE / search_engine

Improvement of the search engine functionality on the NFDI4BioImage training website #1