feat: Add github workspace command

Supports resolving github code index workspace data and searching in it

TODO: Currently this api do not accepts ghu_ github copilot token, and I need to use `gh cli` instead
that creates hosts.yml

https://github.blog/engineering/the-technology-behind-githubs-new-code-search/

Signed-off-by: Tomas Slusny <[email protected]>

Author: deathbeam

lua/CopilotChat/client.lua

@@ -824,6 +824,24 @@ function Client:embed(inputs, model)
   return results
 end
 
+--- Search for the given query
+---@param query string: The query to search for
+---@param repository string: The repository to search in
+---@param model string: The model to use for search
+---@return table<CopilotChat.context.embed>
+function Client:search(query, repository, model)
+  local models = self:fetch_models()
+
+  local provider_name, search = resolve_provider_function('search', model, models, self.providers)
+  local headers = self:authenticate(provider_name)
+  local ok, response = pcall(search, query, repository, headers)
+  if not ok then
+    log.warn('Failed to search: ', response)
+    return {}
+  end
+  return response
+end
+
 --- Stop the running job
 ---@return boolean
 function Client:stop()

lua/CopilotChat/config/contexts.lua

@@ -5,7 +5,7 @@ local utils = require('CopilotChat.utils')
 ---@class CopilotChat.config.context
 ---@field description string?
 ---@field input fun(callback: fun(input: string?), source: CopilotChat.source)?
----@field resolve fun(input: string?, source: CopilotChat.source, prompt: string):table<CopilotChat.context.embed>
+---@field resolve fun(input: string?, source: CopilotChat.source, prompt: string, model: string):table<CopilotChat.context.embed>
 
 ---@type table<string, CopilotChat.config.context>
 return {
@@ -160,4 +160,10 @@ return {
       return context.quickfix()
     end,
   },
+  workspace = {
+    description = 'Includes all non-hidden files in the current workspace in chat context.',
+    resolve = function(_, _, prompt, model)
+      return context.workspace(prompt, model)
+    end,
+  },
 }

lua/CopilotChat/config/mappings.lua

@@ -409,7 +409,8 @@ return {
       async.run(function()
         local embeddings = {}
         if section and not section.answer then
-          embeddings = copilot.resolve_embeddings(section.content, chat.config)
+          local _, selected_model = pcall(copilot.resolve_model, section.content, chat.config)
+          embeddings = copilot.resolve_embeddings(section.content, selected_model, chat.config)
         end
 
         for _, embedding in ipairs(embeddings) do

lua/CopilotChat/config/providers.lua

@@ -44,6 +44,7 @@ local utils = require('CopilotChat.utils')
 ---@field get_agents nil|fun(headers:table):table<CopilotChat.Provider.agent>
 ---@field get_models nil|fun(headers:table):table<CopilotChat.Provider.model>
 ---@field embed nil|string|fun(inputs:table<string>, headers:table):table<CopilotChat.Provider.embed>
+---@field search nil|string|fun(query:string, repository:string, headers:table):table<CopilotChat.Provider.output>
 ---@field prepare_input nil|fun(inputs:table<CopilotChat.Provider.input>, opts:CopilotChat.Provider.options):table
 ---@field prepare_output nil|fun(output:table, opts:CopilotChat.Provider.options):CopilotChat.Provider.output
 ---@field get_url nil|fun(opts:CopilotChat.Provider.options):string
@@ -101,11 +102,41 @@ local function get_github_token()
   error('Failed to find GitHub token')
 end
 
+local cached_gh_apps_token = nil
+
+--- Get the github apps token (gho_ token)
+---@return string
+local function get_gh_apps_token()
+  if cached_gh_apps_token then
+    return cached_gh_apps_token
+  end
+
+  async.util.scheduler()
+
+  local config_path = utils.config_path()
+  if not config_path then
+    error('Failed to find config path for GitHub token')
+  end
+
+  local file_path = config_path .. '/gh/hosts.yml'
+  if vim.fn.filereadable(file_path) == 1 then
+    local content = table.concat(vim.fn.readfile(file_path), '\n')
+    local token = content:match('oauth_token:%s*([%w_]+)')
+    if token then
+      cached_gh_apps_token = token
+      return token
+    end
+  end
+
+  error('Failed to find GitHub token')
+end
+
 ---@type table<string, CopilotChat.Provider>
 local M = {}
 
 M.copilot = {
   embed = 'copilot_embeddings',
+  search = 'copilot_search',
 
   get_headers = function(token)
     return {
@@ -279,6 +310,7 @@ M.copilot = {
 
 M.github_models = {
   embed = 'copilot_embeddings',
+  search = 'copilot_search',
 
   get_headers = function(token)
     return {
@@ -360,4 +392,80 @@ M.copilot_embeddings = {
   end,
 }
 
+M.copilot_search = {
+  get_headers = M.copilot.get_headers,
+
+  get_token = function()
+    return get_gh_apps_token(), nil
+  end,
+
+  search = function(query, repository, headers)
+    utils.curl_post(
+      'https://api.github.com/repos/' .. repository .. '/copilot_internal/embeddings_index',
+      {
+        headers = headers,
+      }
+    )
+
+    local response, err = utils.curl_get(
+      'https://api.github.com/repos/' .. repository .. '/copilot_internal/embeddings_index',
+      {
+        headers = headers,
+      }
+    )
+
+    if err then
+      error(err)
+    end
+
+    if response.status ~= 200 then
+      error('Failed to check search: ' .. tostring(response.status))
+    end
+
+    local body = vim.json.decode(response.body)
+
+    if
+      body.can_index ~= 'ok'
+      or not body.bm25_search_ok
+      or not body.lexical_search_ok
+      or not body.semantic_code_search_ok
+      or not body.semantic_doc_search_ok
+      or not body.semantic_indexing_enabled
+    then
+      error('Failed to search: ' .. vim.inspect(body))
+    end
+
+    local body = vim.json.encode({
+      query = query,
+      scopingQuery = '(repo:' .. repository .. ')',
+      similarity = 0.766,
+      limit = 100,
+    })
+
+    local response, err = utils.curl_post('https://api.individual.githubcopilot.com/search/code', {
+      headers = headers,
+      body = utils.temp_file(body),
+    })
+
+    if err then
+      error(err)
+    end
+
+    if response.status ~= 200 then
+      error('Failed to search: ' .. tostring(response.body))
+    end
+
+    local out = {}
+    for _, result in ipairs(vim.json.decode(response.body)) do
+      table.insert(out, {
+        filename = result.path,
+        filetype = result.languageName:lower(),
+        score = result.score,
+        content = result.contents,
+      })
+    end
+    return out
+  end,
+}
+
 return M

lua/CopilotChat/context.lua

@@ -639,6 +639,20 @@ function M.quickfix()
   return out
 end
 
+--- Get the content of the current workspace
+---@param prompt string
+---@param model string
+function M.workspace(prompt, model)
+  local git_remote =
+    vim.trim(utils.system({ 'git', 'config', '--get', 'remote.origin.url' }).stdout)
+  local repo_path = git_remote:match('github.com[:/](.+).git$')
+  if not repo_path then
+    error('Could not determine GitHub repository from git remote: ' .. git_remote)
+  end
+
+  return client:search(prompt, repo_path, model)
+end
+
 --- Filter embeddings based on the query
 ---@param prompt string
 ---@param model string

lua/CopilotChat/init.lua

@@ -220,9 +220,10 @@ end
 
 --- Resolve the embeddings from the prompt.
 ---@param prompt string
+---@param model string
 ---@param config CopilotChat.config.shared
 ---@return table<CopilotChat.context.embed>, string
-function M.resolve_embeddings(prompt, config)
+function M.resolve_embeddings(prompt, model, config)
   local contexts = {}
   local function parse_context(prompt_context)
     local split = vim.split(prompt_context, ':')
@@ -262,7 +263,7 @@ function M.resolve_embeddings(prompt, config)
   for _, context_data in ipairs(contexts) do
     local context_value = M.config.contexts[context_data.name]
     for _, embedding in
-      ipairs(context_value.resolve(context_data.input, state.source or {}, prompt))
+      ipairs(context_value.resolve(context_data.input, state.source or {}, prompt, model))
     do
       if embedding then
         embeddings:set(embedding.filename, embedding)
@@ -648,7 +649,7 @@ function M.ask(prompt, config)
   local ok, err = pcall(async.run, function()
     local selected_agent, prompt = M.resolve_agent(prompt, config)
     local selected_model, prompt = M.resolve_model(prompt, config)
-    local embeddings, prompt = M.resolve_embeddings(prompt, config)
+    local embeddings, prompt = M.resolve_embeddings(prompt, selected_model, config)
 
     local has_output = false
     local query_ok, filtered_embeddings =