Fuzzy matching filters for ddc.vim.

The de facto standard filters, match_head and sorter_rank, behave like the other completion such as VSCode. They suggest terms containing an input as a substring. However, they omit words with a typo. It means that we cannot obtain a suggestion word from wrd. Moreover, it is hard to select a long word with a short input.

Our filters fix this problem with the fuzzy filter that behaves like CtrlP.

To install our filters,

Plug 'tani/ddc-fuzzy'

To use our filters,

call ddc#custom#patch_global('ui', 'pum')
call ddc#custom#patch_global('sourceOptions', {
  \   '_': {
  \     'matchers': ['matcher_fuzzy'],
  \     'sorters': ['sorter_fuzzy'],
  \     'converters': ['converter_fuzzy']
  \   }
  \ })

matcher_fuzzy filter provides an option splitMode. For the default mode, { 'splitMode': 'character' }, each input character matches any characters of a candidate. For example, abc matches axbxc. On the other hand, for the option {'splitMode': 'word'}, each input character matches the beginning of a word of a candidate. For example, abc does not match axbxc but abc matches a_bc and a_b_c.

call ddc#custom#patch_global('filterParams', {
  \   'matcher_fuzzy': {
  \     'splitMode': 'word'
  \   }
  \ })

converter_fuzzy filter provides an option hlGroup to change the highlight group in vim. This filter applies the highlight group to characters matched by the input pattern.

call ddc#custom#patch_global('filterParams', {
  \   'converter_fuzzy': {
  \     'hlGroup': 'SpellBad'
  \   }
  \ })

• tani/ddc-git, git commit/file/branch completion
• tani/ddc-oldfiles, oldfiles completion
• tani/ddc-path, path namescompletion in current working directory
• tani/ddc-onp, yet another fuzzy matcher with O(NP) algorithm

This work is licensed under the MIT License. Copyright © 2021 TANIGUCHI Masaya. All rights reserved.