feat: fetch ollama models to display in the model selector (#2287)

lua/avante/model_selector.lua

@@ -75,6 +75,9 @@ function M.open()
     ::continue::
   end
 
+  -- Sort models by name for stable display
+  table.sort(models, function(a, b) return (a.name or "") < (b.name or "") end)
+
   if #models == 0 then
     Utils.warn("No models available in config")
     return

lua/avante/providers/ollama.lua

@@ -233,4 +233,59 @@ M.on_error = function(result)
   Utils.error(error_msg, { title = "Ollama" })
 end
 
+-- List available models using Ollama's tags API
+function M:models_list()
+  -- Return cached models if available
+  if self._model_list_cache then return self._model_list_cache end
+
+  -- Parse provider config and construct tags endpoint URL
+  local provider_conf = Providers.parse_config(self)
+  if not provider_conf.endpoint then error("Ollama requires endpoint configuration") end
+
+  local curl = require("plenary.curl")
+  local tags_url = Utils.url_join(provider_conf.endpoint, "/api/tags")
+  local base_headers = {
+    ["Content-Type"] = "application/json",
+    ["Accept"] = "application/json",
+  }
+  local headers = Utils.tbl_override(base_headers, self.extra_headers)
+
+  -- Request the model tags from Ollama
+  local response = curl.get(tags_url, { headers = headers })
+  if response.status ~= 200 then
+    Utils.error("Failed to fetch Ollama models: " .. (response.body or response.status))
+    return {}
+  end
+
+  -- Parse the response body
+  local ok, res_body = pcall(vim.json.decode, response.body)
+  if not ok or not res_body.models then return {} end
+
+  -- Helper to format model display string from its details
+  local function format_display_name(details)
+    local parts = {}
+    for _, key in ipairs({ "family", "parameter_size", "quantization_level" }) do
+      if details[key] then table.insert(parts, details[key]) end
+    end
+    return table.concat(parts, ", ")
+  end
+
+  -- Format the models list
+  local models = {}
+  for _, model in ipairs(res_body.models) do
+    local details = model.details or {}
+    local display = format_display_name(details)
+    table.insert(models, {
+      id = model.name,
+      name = string.format("ollama/%s (%s)", model.name, display),
+      display_name = model.name,
+      provider_name = "ollama",
+      version = model.digest,
+    })
+  end
+
+  self._model_list_cache = models
+  return models
+end
+
 return M

lua/avante/types.lua

@@ -291,6 +291,19 @@ vim.g.avante_login = vim.g.avante_login
 ---@alias AvanteLLMStopCallback fun(opts: AvanteLLMStopCallbackOptions): nil
 ---@alias AvanteLLMConfigHandler fun(opts: AvanteSupportedProvider): AvanteDefaultBaseProvider, table<string, any>
 ---
+---@class AvanteProviderModel
+---@field id string
+---@field name string
+---@field display_name string
+---@field provider_name string
+---@field version string
+---@field tokenizer? string
+---@field max_input_tokens? integer
+---@field max_output_tokens? integer
+---@field policy? boolean
+---
+---@alias AvanteProviderModelList AvanteProviderModel[]
+---
 ---@class AvanteProvider: AvanteSupportedProvider
 ---@field parse_curl_args? AvanteCurlArgsParser
 ---@field parse_stream_data? AvanteStreamParser
@@ -315,6 +328,7 @@ vim.g.avante_login = vim.g.avante_login
 ---@field on_error? fun(result: table<string, any>): nil
 ---@field transform_tool? fun(self: AvanteProviderFunctor, tool: AvanteLLMTool): AvanteOpenAITool | AvanteClaudeTool
 ---@field get_rate_limit_sleep_time? fun(self: AvanteProviderFunctor, headers: table<string, string>): integer | nil
+---@field models_list? fun(self): AvanteProviderModelList | nil
 ---
 ---@alias AvanteBedrockPayloadBuilder fun(self: AvanteBedrockModelHandler | AvanteBedrockProviderFunctor, prompt_opts: AvantePromptOptions, request_body: table<string, any>): table<string, any>
 ---