Demo of Manticore Search

[sanikolaev]

We have some great tools for learning Manticore Search like the manual which is powered by Manticore Search and online courses. Plus, there are neat little demos on the play platform (like this one: https://play.manticoresearch.com/didyoumean/).

I'm thinking it's time we build a bigger demo. How about an app that uses Manticore to search for things like songs, movies, or news articles? This way, people can see a bigger project working with Manticore. We can share the app's code and also write a simple guide on how to create it.

Here's the plan:

• Pick a topic: we should start with something fun that shows what Manticore can do.
• Make the demo: develop a simple yet cool app. It should be easy for beginners to get and also show some advanced tricks for the others.

[AbstractiveNord]

• Text articles is great, especially if they contain source code opened.
• In case of videos, if this type of content is planned, would be wonderful if ManticoreSearch will have series like this. Videos usually more attractive for newcomers. May be effective even only highlighting interactive demos with videos.

[sanikolaev]

Further ideas/plans:

• Check if we can integrate with https://github.com/ankane/onnxruntime-php
  • and some default model, e.g. E5 or BERT
  • create a task to expose the KNN functionality in the PHP client
• Review the readme
• Make a blogpost

[sanikolaev]

Empty result in this case:

[sanikolaev]

The Open/Closed states should be hidden/blocked in case of "Comments".

[sanikolaev]

• [x] Let's add ... around the highlighting passages:

[sanikolaev]

• [x] Bug: not all comments get fetched from Github (e.g. search by "fixed in").

[sanikolaev]

• [x] TODO: Reset all advanced filters when clicking on "Comments".

[sanikolaev]

More bugs:

• [x] Labels are not displayed
• [x] "Everywhere + Closed + Sort by: most commented" displays 0 results

[sanikolaev]

• [x] add a space:

[sanikolaev]

• [x] This is displayed in black in Chrome and Safari:

[sanikolaev]

• [x] order facets by count
• [x] let's include comments when filtering by author:

[sanikolaev]

Check if we can integrate with https://github.com/ankane/onnxruntime-php

Blocked by https://github.com/manticoresoftware/manticoresearch-php/issues/184

[sanikolaev]

• [x] Filter by author works wrong:

[sanikolaev]

• [x] Filters are not reflected in the UI:

[sanikolaev]

• [x] Let's hide comments when the selected filters can't be applied to comments:

[sanikolaev]

• [x] In case of the infiinite scrolling, wrong results get loaded.

[sanikolaev]

• [x] Let's rtrim the slash:

  Same with quickwit-oss/quickwit/

• [x] Exception, potentially a breach:
• [x] "issues & PRs"
• [x] <span> shouldn't be there:
• [x] Can't reset the query if you do:
  • search for something
  • Advanced -> author -> sanikolaev ON
  • Advanced -> author -> sanikolaev OFF
  • Cmd+R to refresh the page
  • "Reset" doesn't reset the query
• [x] Filter by comments counts and "Search in comments" shouldn't be available at the same time, it doesn't make sense:

[sanikolaev]

• [x] Remove the quotes mark and think how to deal with this query in general:

  

  e.g. in case of a syntax error - make another attempt with a query cleaned from special characters (" etc.).

• [x] FT query is not reset anymore which is good, but:
  • search for something
  • Advanced -> author -> sanikolaev ON
  • Advanced -> author -> sanikolaev OFF
  • Cmd+R to refresh the page
  • Now "Reset" resets the query
• [x] Filter by comments counts and "Search in comments" can be still used at the same time:
  • Click "Comments"
  • Select "Comments count < 3"
  • Click "Comments" again
  • Select "Comments count < 3" again

[sanikolaev]

• [x] ... issue: try trimming ...... in the end/beginning replacing that with ...

[sanikolaev]

• [ ] experiment more with the facets UI to make it less confusing (grey out filters that would result in zero results etc.)

[sanikolaev]

What's left before we announce the demo:

• [x] update the installation instruction, so the user can download and restore some data (e.g. our repo https://github.com/manticoresoftware/manticoresearch)
• [x] add morphology='lemmatizer_en' min_infix_len='2' expand_keywords='1'
• [x] rebuild data on https://github.manticoresearch.com/

[sanikolaev]

Check if we can integrate with https://github.com/ankane/onnxruntime-php

Blocked by https://github.com/manticoresoftware/manticoresearch-php/issues/184

Unblocked.

[sanikolaev]

• [ ] improve exceptions handling, especially the exception which is thrown when the token has exceeded its limit

[donhardman]

I'm currently waiting for a response because I can't use ext-php-rs with the latest PHP 8.3. I need this to move forward with building a tokenizer as an extension in PHP. For more details, you can check out this issue: https://github.com/davidcole1340/ext-php-rs/issues/314

[donhardman]

I believe there's no reason to keep this task open since the main project is complete, and it would be better to apply future iterations in separate tasks and steps. I've divided what we have here into two tasks and will be closing this one:

Vector Search: https://github.com/manticoresoftware/manticore-github-issue-search/issues/10 Features and Extras: https://github.com/manticoresoftware/manticore-github-issue-search/issues/11