Code-Completions

Code Completions

Note

For Codeium code completion, you need to set up dependencies:

dependencies = { "nvim-lua/plenary.nvim", "MunifTanjim/nui.nvim", "Exafunction/codeium.nvim" },

• Settings
  • virtual text
  • blink.cmp
  • nvim-cmp

Note

Your model needs to support FIM (Fill-in-Middle).

In fact, you should also make the most of FIM as it has an advantage in completion, being able to fill in based on context rather than just continuing from the preceding text.

The url used by the code completion tool and the url used by the chat task are usually two different ones.

1. You can use autocmd to enable the completion feature of llm.nvim.

vim.api.nvim_create_autocmd({ "VimEnter" }, {
  callback = function()
    vim.api.nvim_command("LLMAppHandler Completion")
  end,
})

2. You can make the completion of llm.nvim take effect immediately by disabling lazy loading. (Completion AI tool requires setting auto_trigger = true)

{
  "Kurama622/llm.nvim",
  dependencies = { "nvim-lua/plenary.nvim", "MunifTanjim/nui.nvim" },
  lazy = false,
  config = function()
    ...
  end,
}

Settings

virtual text

Completion AI tool requires setting style = "virtual_text"

Set key mapping for virtual_text

Completion = {
  opts = {
    keymap = {
      virtual_text = {
        accept = {
          mode = "i",
          keys = "<A-a>",
        },
        next = {
          mode = "i",
          keys = "<A-n>",
        },
        prev = {
          mode = "i",
          keys = "<A-p>",
        },
        toggle = {
          mode = "n",
          keys = "<leader>cp",
        },
      },
    },
  },
}

blink.cmp

Completion AI tool requires setting style = "blink.cmp"

• blink.cmp config

{
    "saghen/blink.cmp",
    opts = {
      completion = {
        trigger = {
          prefetch_on_insert = false
          -- allow triggering by white space
          show_on_blocked_trigger_characters = {},
        },
      },

      keymap = {
        ["<C-y>"] = {
          function(cmp)
            cmp.show({ providers = { "llm" } })
          end,
        },
      },

      sources = {
        -- if you want to use auto-complete
        default = { "llm" },
        providers = {
          llm = {
            name = "llm",
            module = "llm.common.completion.frontends.blink",
            timeout_ms = 10000,
            score_offset = 100,
            async = true,
          },
        },
      },
    },
  }

nvim-cmp

Completion AI tool requires setting style = "nvim-cmp"

{
  "hrsh7th/nvim-cmp",
  optional = true,
  opts = function(_, opts)
    -- if you wish to use autocomplete
    table.insert(opts.sources, 1, {
      name = "llm",
      group_index = 1,
      priority = 100,
    })

    opts.performance = {
      -- It is recommended to increase the timeout duration due to
      -- the typically slower response speed of LLMs compared to
      -- other completion sources. This is not needed when you only
      -- need manual completion.
      fetching_timeout = 5000,
    }
  end,
},