Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Searching for Substrings of the room name / number („Raumkennung“) #1733

Open
BaumBanane opened this issue Dec 27, 2024 · 1 comment
Open
Labels
feature New feature or request

Comments

@BaumBanane
Copy link
Contributor

BaumBanane commented Dec 27, 2024

Is your feature request related to a problem? Please describe.
When searching for the room 1.221 in the Hauptgebäude, searching for 1.221 will return no useful results. You either need to search for 1221 (Architektenname) or 01.221.

Describe the solution you'd like
When searching for a room, consider all superstrings of that room in the Architektenname and the Raumkennung

Describe alternatives you've considered
I have not considered any alternatives.

Additional context
Current behavior:

image image
image image
@BaumBanane BaumBanane added the feature New feature or request label Dec 27, 2024
@CommanderStorm
Copy link
Member

The current way the costing is done, these are our ranking rules:

.with_ranking_rules([
"words",
"typo",
"rank:desc",
"proximity",
"attribute",
"sort",
"exactness",
])

The rule that would be able to help here is typo.
The reason the rooms are currently lower ranked is that

  • maybe rank:desc (i.e. the rule that adds a bias for larger/more relevant "things" in searching) should have a lower priority

  • maybe the type bias (i.e. a lecture hall is more relevant than ) should be spit out of rank

  • I currently don't have a second pass to enable better typo searching
    => stripping leading zeros and adding this as another index to search on
    => keeping the unmodifled version around (maybe with typos disabled to aid in relevancy => only have exactness on this one?)

  • The typo tolerance is currently not configured.
    Thing is about relevancy is that ther alwaysis a tradeoff. When I experimented with higher typo tolerance values, the results were worse in the "benchmark" (might be biased or not have).

    Typo on the first character Meilisearch considers a typo on a query's first character as two typos.

    I don't quite know from the top of my head what the minimum word size for one/two typos are.

In general, searching is really hard to get relevant results

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature New feature or request
Projects
None yet
Development

No branches or pull requests

2 participants