Add the ability to complete from spell dictionaries about asyncomplete-look HOT 19 CLOSED

Yes, it works, thank you!
Is it possible to add information about completion source? For example, if I use completion from buffer, it shows [buffer] to the right of the proposed word:

Something like [spell] or [dictionary]?

Maybe better experience with fuzzy filter such as machakann/asyncomplete-ezfilter.vim.

Thank you, I like it :)

from asyncomplete-look.

Thank you a lot!

from asyncomplete-look.

Your plugin is very useful! I suggest to add this information to the repo readme and submit this plugin to the asyncomplete sources too.

from asyncomplete-look.

Maybe is it possible to add small completion delay to reduce input lag?

Sorry, I don't know how to implement it.

Now I decided to support only grep way because I think it is suitable for this plugin. I have already merged into master.
Please be careful that necessary setting changed from the above comment (#1 (comment)).

from asyncomplete-look.

Hello! Please try this branch and setting.
https://github.com/htlsne/asyncomplete-look/tree/spell.

au User asyncomplete_setup call asyncomplete#register_source({
    \ 'name': 'spell',
    \ 'whitelist': ['text', 'markdown'],
    \ 'completor': function('asyncomplete#sources#spell#completor'),
    \ })

" if you not set it other way
set spell

I think it is the same completion mechanism as MUcomplete (although I have never used that plugin).

from asyncomplete-look.

Maybe better experience with fuzzy filter such as machakann/asyncomplete-ezfilter.vim.

let g:asyncomplete_preprocessor =
  \ [function('asyncomplete#preprocessor#ezfilter#filter')]

let g:asyncomplete#preprocessor#ezfilter#config = {}
let g:asyncomplete#preprocessor#ezfilter#config.spell =
  \ {ctx, items -> ctx.osa_filter(items, 1)}

from asyncomplete-look.

Merged into master.

Added in latest version:

Is it possible to add information about completion source? For example, if I use completion from buffer, it shows [buffer] to the right of the proposed word:
Something like [spell] or [dictionary]?

from asyncomplete-look.

After a long use, I began to notice that when using machakann/asyncomplete-ezfilter.vim, Vim starts to work much slower, especially when moving the cursor. Now I use your plugin without it.

from asyncomplete-look.

I was wrong. The problem is in spell completion itself. Input response becomes much slower with it :(
This is especially noticeable when quickly entering a large number of characters. Try entering something likeajsdashdkjahdkjahdkjahskdhakjhsdkjahsdhlkljkl, you will see how your cursor hangs.

I understand that this, most likely, cannot be done faster :(

from asyncomplete-look.

I hear that spellsuggest() is slow.
For provide complete from spell dictionaries, another approach may be preferable.

Try this branch. It grep your dictionary file directly.
https://github.com/htlsne/asyncomplete-look/tree/grep

au User asyncomplete_setup call asyncomplete#register_source({
    \ 'name': 'look',
    \ 'whitelist': ['text', 'markdown'],
    \ 'completor': function('asyncomplete#sources#look#completor'),
    \ })

" Find your dictionary path such as /home/<username>/.vim/spell/en.utf-8.add
let g:asc_look_spell_file = /full/path/to/your/en.utf-8.add

au User asyncomplete_setup call asyncomplete#register_source({
    \ 'name': 'spell',
    \ 'whitelist': ['text', 'markdown'],
    \ 'completor': function('asyncomplete#sources#spell#completor'),
    \ })

from asyncomplete-look.

Sorry for newbie questions, but where can I find en.utf-8.add file? I using Neovim and have only preinstalled /usr/share/nvim/runtime/spell/en.utf-8.spl and ~/.local/share/nvim/site/spell/ru.utf-8.spl. I tried both, but completion not works.

from asyncomplete-look.

Does ~/.config/nvim/spell/en.utf-8.add exists? .add file is plain text file that contains your spell dictionary word.

from asyncomplete-look.

No, the spell directory even not exists.

from asyncomplete-look.

Hmm... When you use zg to add a good word, a message like Word 'xxx' added to ~/.config/nvim/spell/en.utf-8.add should be shown. This is .add file.

from asyncomplete-look.

Oh, sorry, I just didn’t add the words :)
I tested it with the following configuration:

au User asyncomplete_setup call asyncomplete#register_source({
    \ 'name': 'look',
    \ 'whitelist': ['text', 'markdown'],
    \ 'completor': function('asyncomplete#sources#look#completor'),
    \ })

let g:asc_look_spell_file = '/home/gena/.config/nvim/spell/en.utf-8.add'

au User asyncomplete_setup call asyncomplete#register_source({
    \ 'name': 'spell',
    \ 'whitelist': ['text', 'markdown'],
    \ 'completor': function('asyncomplete#sources#spell#completor'),
    \ })

But it not works, added word plugins not suggested. The file /home/gena/.config/nvim/spell/en.utf-8.add exists and contains plugins.

from asyncomplete-look.

Oh, sorry, I just didn’t add the words :)

I thought "spell dictionaries" means these added words, didn't you mean that...?

But it not works, added word plugins not suggested. The file /home/gena/.config/nvim/spell/en.utf-8.add exists and contains plugins.

Using grep branch, and editing filetype=text or markdown file?

from asyncomplete-look.

I thought "spell dictionaries" means these added words, didn't you mean that...?

I meant auto-completion from the words that I added + words from the standard spell dictionary. The current master branch works in this way.

Using grep branch, and editing filetype=text or markdown file?

Yes, I even checked on both filetypes.

from asyncomplete-look.

I meant auto-completion from the words that I added + words from the standard spell dictionary. The current master branch works in this way.

If you mean "standard spell dictionary" is same dictionary as z=, all I can do now is current master branch.

grep branch uses your added words + look command dictionary (/usr/share/dict/words). I expect this is faster than master branch (not certainly because speed depends on environments).
It works like this.

If you don't like this, I don't have any other idea, sorry. If you like it, I can support to work this.

from asyncomplete-look.

If you mean "standard spell dictionary" is same dictionary as z=, all I can do now is current master branch.

Yes, I understand :(
Maybe is it possible to add small completion delay to reduce input lag?

grep branch uses your added words + look command dictionary (/usr/share/dict/words). I expect this is faster than master branch (not certainly because speed depends on environments).
It works like this.

Yes, this is nice idea. The spellsuggest() is better, but it's slow when I typing a lot of letters :(

from asyncomplete-look.