Skip to content

Match Settings

MituuZ edited this page Jun 6, 2026 · 7 revisions

Match settings

Is part of Global Settings.

Test bench

You can live test the match options using the test bench (does not update in place). It shows each individual score for each file and you can tweak it as you like.

Test bench showing the preview

Tolerance

Used to make the file search a bit more flexible. Tells how many wrong characters can be present before deciding that a file path cannot match the search string.

Match characters multiple times

Counts occurrences of each unique character present in the search string.

e.g. search string "file" has 5 occurrences in "/dir/file"

This option can affect performance more than the others, as it requires full passes over the file paths.

Single char

Set the multiplier for each individual character that was matched with "Match characters multiple times".

The final result is divided by 10 to avoid using floats and still get more precise control.

Partial path match

Tells how much static score does a partial path match add. This can be used to prioritize filename matches or pointing to the correct package/directory with a space.

Examples

  • "file" matches "/dir/file"
  • "dir" matches "/dir/file"
  • "ext" matches "/dir/file.ext"

Streak modifier

Note! The actual streak might not be the longest possible streak.

When checking that each character from the search string is present and ordered in the file path the longest streak is used to score the entry.

This is used to set the multiplier for the longest streak.

Filename modifier

Search string is checked against each filename and the longest streak is used to score the entry.

This is used to set the multiplier for the filename streak. Thus enabling pushing only filename matches upper in the list.

Recently used files

Introduced in version 2.4.0, Fuzzier stores files that are opened using the file search to a persistent settings file (.idea/FuzzierSettings.xml).

This works as an LRU cache with a maximum of 100 entries. The list is ordered and contains the file path and access count that are used by the scoring algorithm.

Frequency boost

Frequency weight scales up in steps to a default max of 50 accesses in history, which equals 10 extra points. Any accesses beyond that are capped at the same maximum.

The frequency weight can be adjusted to increase or decrease that maximum value, e.g. boost of 20 would mean that 50 accesses equals 20 extra points.

Recency boost

Recency boost starts at a default maximum of 10 extra points for the most recently used file. It decreases in steps by 1 point every 2 positions in the recent-files history, reaching 0 points at index 20. Any files older than that receive no recency boost.

The recency weight can be adjusted to increase or decrease the maximum contribution; for example, a recency weight of 20 would mean the most recent file gets 20 extra points.

Clone this wiki locally