Unable to search issues for part of a word.

[sgtwilko]

• Gitea version (or commit ref): 1.2.3
• Git version: 2.15.0
• Operating system: Windows
• Database (use [x]):
  • PostgreSQL
  • MySQL
  • MSSQL
  • SQLite
• Can you reproduce the bug at https://try.gitea.io:
  • Yes (provide example URL)
  • No
  • Not relevant
• Log gist:

Description

The Issue search does not find an issue when searching for part of a word within the issue name/description.

We discovered this as we had created an issue for each file we needed to deal with.
Our filenames have a specification that indicates where and what the purpose is of the file.
For example:
Customer_SupportTicketRaise

If we search the issues for "SupportTicket" or "Customer" no issues are returned as this would only match a partial string.
I can understand that searching for part of a string could return too many results for most people, but either including those results below the full string match, or having the option to search for a partial string would be better as, at first, we thought the entire search was broken.

Screenshots

Created serveral issues:

Searched for "dave" but it failed to find "test_dave":

[lunny]

I think this is a designed feature. And currently it also don't support CJK languages.

[ncantelmo]

Hey all,

I just upgraded from Gogs yesterday specifically because I saw that issue search had been added to Gitea. I'd like to second this enhancement request, but also request that fuzzy matching be considered as part of it.

When you're adding to a large backlog and you don't remember the exact terms you used (e.g. migrate vs. migration vs. migrations), it can be easy to miss an existing issue and create a duplicate.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs during the next 2 weeks. Thank you for your contributions.

[JVimes]

Search for click didn't find clicks for me. Very surprising. This is a moderately large problem for me.

[lunny]

Since Gitea has 3 engines for issue indexing, I assume we are talking about bleve. #22829 has resolved camelCase problem.

[JVimes]

@lunny Does #22829 make query "click" find result "clicks" (with the "s")?

[lunny]

@lunny Does #22829 make query "click" find result "clicks" (with the "s")?

Not yet. Looks like bleve doesn't support that, we may need to write a customized token filter

[wxiaoguang]

Bleve has some stemers:

• Plural Stemmer for English blevesearch/bleve#1750

That issue mentions that "snowball and porter stemmers" can stem words to their origins.

(Just come across, FYI)

[JVimes]

If stemmers can implement partial-word matches that would solve the general case. Otherwise I think they would help a bit.

[jgschis]

You can use the ngram analyser to do partial string matching.

So customer_support_ticket

could be ngrammed as

cus
cust
custo
custom
custome
customer
customer_
etc

[Leon-001]

Or at least a wildcard symbol such as "*" or "%" would be great!

[coolbombom]

any news regarding this issue. when will it be integrated/fixed?

[hakito]

At least a prefix based search now works (when using an external indexer like meilisearch)

But only for the global issues search and not on the repo level :-(