ZEM

Neovim plugin that can quickly navigate to previously indexed locations.

Early development, but useable.

It uses the python3 remote plugin api and sqlite.

You can use :Zem to open a prompt with preview window that shows the currently matching elements, or you can use :ZemEdit QUERY to directly go to the location.

Install

1. put the zem folder inside a rplugin/python3 folder that nvim searches. (for example using vim-plug Plug 'wonkodv/zem')

2. in nvim: :UpdateRemotePlugins and restart

3. (optionaly) define mappings in init.vim

   nnoremap <S-Y> :Zem<CR>
   nnoremap <C-y> :Zem<C-R><C-W><CR>
   nnoremap <C-]> :ZemEdit =Impl <C-R><C-W><CR>
   

Indexing

Before using the :Zem command, the Index must be filled. By executing :ZemUpdateIndex, all configured sources are indexed and the elements stored in an sqlite3 database.

Elements

Each element consists of these fields:

• match: Text that is matched against to find it (e.g. function name)
• type: what this element represents, (e.g. File, Define, ...)
• file: which file the element is in
• location: how to find the element in the file (Line number or /{pattern}/)

Query

Each letter of a word must match in order, but words can match out of order:

    query   | matches     |  doesn't match
------------|-------------|-----------------
 fob        |  FooBar     |  Foo 
 ba foo     |  FooBar     |  Foo 
 foofoo     |  FooBarFoo  |  Foo 
 foo foo    |  Foo        |  Bar 

Tokens that start with = match against types. If none are given, all types are matched, otherwise only those types that start with the given ones are matched

    query  | matches
-----------|-------------
=I         | ['Implementation', 'Import'] 
=Impl      | ['Implementation'] 
=Impl =De  | ['Implementation', 'Define'] 

Matching multiple types might be useful to define mappings that open zem and narrow down the number of matched types (e.g. nmap <S-T> :Zem =Define =Type <CR>)

Sources

The two implemented sources are:

• files: index files
• tags: scan a file generated by ctags and index each tag.

Config

the following configuration variables can be set in init.vim and/or, in a .nvimrc file inside each project. (.nvimrc files are sourced in the CWD if the option exrc is set.)

• g:zem_db ('.zem.sqlite') Name of the index database.
• g:zem_height (20) Number of rows in the preview window
• g:zem_prompt ('ZEM> ') Prompt
• g:zem_markup (...) python str.format()-string to display match, typ, file, location
• g:zem_sources (['files', 'tags']) names of scanners to use. must be implemented in zem/scanner.py or vim functions that return elements as (match, typ, file, location)-tuples.
• g:zem_source_FUNCNAME Parameter that is passed to the scanner. Can be omitted to use defaults.
  • g:zem_source_files dictionary with these fields:
    • root (".") where to search for files
    • pattern (["**/*.*"]) search pattern (python pathlib.Path.glob() syntax)
    • exclude (["*~",".git.","*.pyc",...]) files to exclude (unusual syntax)
    • type ("File") type of the matches
  • g:zem_source_tags dictionary with these fields:
    • files (['.tags', 'tags']) files to scan
    • type_map ({'d':'Define','f':'Implementation', ...}) Map the tag-kind to Element-type. The default maps most tag kinds onto one of: Define, Prototyp, Typedef, Implementation