From 4204f1dfb42b85e17df84f336895f67a050589e6 Mon Sep 17 00:00:00 2001
From: Meisterlala <6453306+Meisterlala@users.noreply.github.com>
Date: Sun, 7 Jun 2026 19:19:13 +0200
Subject: [PATCH] feat(ai-commit): align streaming status with Ollama API and
 refine error reporting for timeout detection

---
 local/ai-commit/lua/ai-commit/config.lua      |  3 +-
 local/ai-commit/lua/ai-commit/init.lua        |  9 +++
 local/ai-commit/lua/ai-commit/providers.lua   | 61 +++++++++++++++++--
 local/ai-provider/lua/ai-provider/core.lua    | 56 ++++++++++-------
 local/ai-provider/lua/ai-provider/init.lua    |  2 +
 .../lua/ai-provider/providers/ollama.lua      | 19 +++---
 .../tests/ai_provider/core_spec.lua           | 41 +++++++++++++
 .../tests/ai_provider/ollama_options_spec.lua | 28 ++++++++-
 lua/plugins/ai-provider.lua                   | 12 +++-
 9 files changed, 194 insertions(+), 37 deletions(-)

diff --git a/local/ai-commit/lua/ai-commit/config.lua b/local/ai-commit/lua/ai-commit/config.lua
index f4ee05b..f4dbc52 100644
--- a/local/ai-commit/lua/ai-commit/config.lua
+++ b/local/ai-commit/lua/ai-commit/config.lua
@@ -2,6 +2,7 @@ local M = {}
 
 M.summary_source_id = 'ai-commit-summarize'
 M.message_source_id = 'ai-commit-message'
+M.refine_source_id = 'ai-commit-refine'
 
 M.values = {
   context = {
@@ -9,7 +10,7 @@ M.values = {
     recent_commits = true,
     staged_changes = true,
   },
-  max_tokens = 10000,
+  max_tokens = 32768,
   spinner_interval = 80,
   preview_lines = 5,
   max_diff_chars = 100000,
diff --git a/local/ai-commit/lua/ai-commit/init.lua b/local/ai-commit/lua/ai-commit/init.lua
index 94cf296..7aa2b47 100644
--- a/local/ai-commit/lua/ai-commit/init.lua
+++ b/local/ai-commit/lua/ai-commit/init.lua
@@ -16,6 +16,7 @@ function M.setup(opts)
     local sources = {
       { id = config.summary_source_id, name = 'AI Commit: OpenCode Summary' },
       { id = config.message_source_id, name = 'AI Commit: Commit Message' },
+      { id = config.refine_source_id, name = 'AI Commit: Refinement' },
     }
     for _, source in ipairs(sources) do
       ai_provider.register_source(source.id, { name = source.name })
@@ -53,6 +54,10 @@ function M.setup(opts)
   vim.api.nvim_create_user_command('AICommitSummaryModel', providers.select_summary_model, {
     desc = 'Select AI commit summary model',
   })
+
+  vim.api.nvim_create_user_command('AICommitRefinementModel', providers.select_refinement_model, {
+    desc = 'Select AI commit refinement model',
+  })
 end
 
 function M.insert()
@@ -63,4 +68,8 @@ function M.select_model()
   providers.select_model()
 end
 
+function M.select_refinement_model()
+  providers.select_refinement_model()
+end
+
 return M
diff --git a/local/ai-commit/lua/ai-commit/providers.lua b/local/ai-commit/lua/ai-commit/providers.lua
index 5f148de..99c1897 100644
--- a/local/ai-commit/lua/ai-commit/providers.lua
+++ b/local/ai-commit/lua/ai-commit/providers.lua
@@ -77,6 +77,51 @@ local function provider_label(provider)
   return provider:sub(1, 1):upper() .. provider:sub(2)
 end
 
+local function format_error_detail(meta)
+  if type(meta) ~= 'table' then
+    return 'unknown error'
+  end
+
+  local parts = { tostring(meta.error or 'unknown error') }
+  if meta.done_reason then
+    table.insert(parts, 'done_reason=' .. tostring(meta.done_reason))
+  end
+  if meta.requested_model then
+    table.insert(parts, 'requested_model=' .. tostring(meta.requested_model))
+  end
+  if meta.used_model then
+    table.insert(parts, 'used_model=' .. tostring(meta.used_model))
+  end
+  if meta.elapsed_ms then
+    table.insert(parts, string.format('elapsed=%.1fs', meta.elapsed_ms / 1000))
+  end
+  if meta.load_duration then
+    table.insert(parts, string.format('load=%.1fs', meta.load_duration / 1e9))
+  end
+  if meta.prompt_eval_count then
+    table.insert(parts, 'prompt_tokens=' .. tostring(meta.prompt_eval_count))
+  end
+  if meta.eval_count then
+    table.insert(parts, 'eval_tokens=' .. tostring(meta.eval_count))
+  end
+
+  return table.concat(parts, ' | ')
+end
+
+local function format_error_summary(provider, meta)
+  local error = type(meta) == 'table' and tostring(meta.error or '') or ''
+  if error:match 'timed out' or error:match 'Operation timed out' then
+    return provider .. ' request timed out while waiting for the model.'
+  end
+  if type(meta) == 'table' and meta.done_reason == 'length' then
+    return provider .. ' stopped because the output or context limit was reached.'
+  end
+  if error ~= '' and error ~= 'unknown error' then
+    return provider .. ' request failed: ' .. error
+  end
+  return provider .. ' request failed. See ai-commit logs for details.'
+end
+
 function M.select_model()
   local logger = log()
   local ai_provider = require 'ai-provider'
@@ -91,6 +136,13 @@ function M.select_summary_model()
   ai_provider.select_source_model(config.summary_source_id)
 end
 
+function M.select_refinement_model()
+  local logger = log()
+  local ai_provider = require 'ai-provider'
+  logger.info('Opening AI provider model picker for source=' .. config.refine_source_id)
+  ai_provider.select_source_model(config.refine_source_id)
+end
+
 ---@param full_prompt string
 ---@param callback function(string|nil, table|nil)
 ---@param status_callback function(string)|nil
@@ -161,9 +213,10 @@ local function complete_ai_provider(source_id, full_prompt, callback, status_cal
         return
       end
       if not message then
-        logger.error(provider .. ' chat request failed: ' .. tostring(meta and meta.error or 'unknown error'))
-        vim.notify(provider .. ' request failed. See ai-commit logs.', vim.log.levels.ERROR)
-        callback(nil, nil)
+        local detail = format_error_detail(meta)
+        logger.error(provider .. ' chat request failed: ' .. detail)
+        vim.notify(format_error_summary(provider, meta), vim.log.levels.ERROR)
+        callback(nil, meta)
         return
       end
       callback(util.clean_message(message), { requested_model = model, used_model = meta and meta.used_model or model })
@@ -336,7 +389,7 @@ local function maybe_refine_message(context, message, iteration, callback, statu
   dump_prompt(prompt, 'refinement-' .. tostring(next_iteration), context.branch)
 
   local refinement_context = child_request_context(request_context, {
-    source_id = config.message_source_id,
+    source_id = config.refine_source_id,
     status_action = tostring(next_iteration) .. '. Refinement',
   })
   M.complete_prompt(prompt, function(refined_message)
diff --git a/local/ai-provider/lua/ai-provider/core.lua b/local/ai-provider/lua/ai-provider/core.lua
index 10fa7e9..bb70e43 100644
--- a/local/ai-provider/lua/ai-provider/core.lua
+++ b/local/ai-provider/lua/ai-provider/core.lua
@@ -34,12 +34,13 @@ local function is_configured_provider(provider)
 end
 
 local function source_name(source_id)
-  local source = state.sources[source_id]
+  local prefs = M.load_preferences()
+  local source = prefs.sources and prefs.sources[source_id]
   if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
     return source.name
   end
-  local prefs = M.load_preferences()
-  source = prefs.sources and prefs.sources[source_id]
+
+  source = state.sources[source_id]
   if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
     return source.name
   end
@@ -47,18 +48,35 @@ local function source_name(source_id)
 end
 
 local function source_registered_name(source_id)
-  local source = state.sources[source_id]
+  local prefs = M.load_preferences()
+  local source = prefs.sources and prefs.sources[source_id]
   if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
     return source.name
   end
-  local prefs = M.load_preferences()
-  source = prefs.sources and prefs.sources[source_id]
+
+  source = state.sources[source_id]
   if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
     return source.name
   end
   return nil
 end
 
+local function registered_source_metadata(source_id)
+  local metadata = {}
+  local prefs = M.load_preferences()
+  local source = prefs.sources and prefs.sources[source_id]
+  if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
+    metadata.name = source.name
+  end
+
+  source = state.sources[source_id]
+  if type(source) == 'table' and type(source.name) == 'string' and source.name ~= '' then
+    metadata.name = metadata.name or source.name
+  end
+
+  return metadata
+end
+
 local function source_display(source_id)
   return source_name(source_id)
 end
@@ -66,10 +84,7 @@ end
 local function has_source_model_preference(source_id)
   local prefs = M.load_preferences()
   local source = prefs.sources and prefs.sources[source_id]
-  return type(source) == 'table'
-    and is_configured_provider(source.provider)
-    and type(source.model) == 'string'
-    and source.model ~= ''
+  return type(source) == 'table' and is_configured_provider(source.provider) and type(source.model) == 'string' and source.model ~= ''
 end
 
 local function source_selection_display(source_id)
@@ -199,11 +214,7 @@ function M.get_selected_model(provider, source_id)
   local prefs = M.load_preferences()
   if valid_source_id(source_id) then
     local source = prefs.sources and prefs.sources[source_id]
-    if type(source) == 'table'
-      and source.provider == provider
-      and type(source.model) == 'string'
-      and source.model ~= ''
-    then
+    if type(source) == 'table' and source.provider == provider and type(source.model) == 'string' and source.model ~= '' then
       return source.model
     end
   end
@@ -260,10 +271,11 @@ function M.set_source_selection(source_id, provider, model)
   local prefs = M.load_preferences()
   prefs.sources = prefs.sources or {}
   local source = type(prefs.sources[source_id]) == 'table' and prefs.sources[source_id] or {}
+  local metadata = registered_source_metadata(source_id)
   source.provider = provider
   source.model = model
   source.label = nil
-  source.name = nil
+  source.name = source.name or metadata.name
   prefs.sources[source_id] = source
   return M.save_preferences(prefs)
 end
@@ -275,7 +287,9 @@ function M.register_source(source_id, opts)
 
   opts = opts or {}
   state.sources[source_id] = state.sources[source_id] or {}
-  state.sources[source_id].name = opts.name
+  if type(opts.name) == 'string' and opts.name ~= '' then
+    state.sources[source_id].name = opts.name
+  end
 
   local prefs = M.load_preferences()
   prefs.sources = prefs.sources or {}
@@ -293,7 +307,7 @@ function M.register_source(source_id, opts)
   local provider = opts.provider
   local model = opts.model
   if provider and model and is_configured_provider(provider) then
-    prefs.sources[source_id] = { provider = provider, model = model }
+    prefs.sources[source_id] = { provider = provider, model = model, name = source.name }
     if type(opts.name) == 'string' and opts.name ~= '' then
       prefs.sources[source_id].name = opts.name
     end
@@ -408,9 +422,7 @@ function M.chat(first, second)
   request.model = request.model or M.get_selected_model(provider, request.source_id)
   request.provider_config = get_provider_config(provider)
   if not request.model then
-    log.error(
-      string.format('chat requested without selected model source=%s provider=%s', tostring(request.source_id), tostring(provider))
-    )
+    log.error(string.format('chat requested without selected model source=%s provider=%s', tostring(request.source_id), tostring(provider)))
     if request.callback then
       request.callback(nil, nil)
     end
@@ -567,7 +579,7 @@ local function select_helper(opts, callback)
       end
 
       if callback then
-        callback({ provider = choice.provider, model = choice.model, label = choice.label })
+        callback { provider = choice.provider, model = choice.model, label = choice.label }
       end
     end)
   end)
diff --git a/local/ai-provider/lua/ai-provider/init.lua b/local/ai-provider/lua/ai-provider/init.lua
index b9f925d..315da6c 100644
--- a/local/ai-provider/lua/ai-provider/init.lua
+++ b/local/ai-provider/lua/ai-provider/init.lua
@@ -23,6 +23,7 @@
 ---@class AiProviderModelConfig
 ---@field model string Underlying provider model name.
 ---@field context_size? integer Optional model-specific context size override.
+---@field timeout? integer Optional model-specific request timeout in milliseconds.
 ---@field think? boolean Optional Ollama thinking mode override for this logical model profile.
 
 ---@class AiProviderConfig
@@ -60,6 +61,7 @@
 ---@field stream? boolean Whether the provider should stream chunks. Defaults to true when supported.
 ---@field max_tokens? integer Maximum generated tokens/provider equivalent.
 ---@field context_size? integer Per-request context size override.
+---@field timeout? integer Per-request timeout in milliseconds.
 ---@field keep_alive? string|integer Per-request keep-alive/unload timeout override.
 ---@field ps_timeout? integer Per-request loaded-model inspection timeout in milliseconds.
 ---@field think? boolean Per-request Ollama thinking mode override.
diff --git a/local/ai-provider/lua/ai-provider/providers/ollama.lua b/local/ai-provider/lua/ai-provider/providers/ollama.lua
index 09d3aa4..48fd785 100644
--- a/local/ai-provider/lua/ai-provider/providers/ollama.lua
+++ b/local/ai-provider/lua/ai-provider/providers/ollama.lua
@@ -167,6 +167,7 @@ function M.chat(request)
   local raw_model = selected_config.model or selected_model
   local context_size = request.context_size or selected_config.context_size or provider_config.context_size
   local keep_alive = request.keep_alive or provider_config.keep_alive
+  local timeout = request.timeout or selected_config.timeout or provider_config.timeout or 120000
   local think = request.think
   if think == nil then
     think = selected_config.think
@@ -246,7 +247,7 @@ function M.chat(request)
       tostring(context_size),
       tostring(request.max_tokens),
       tostring(keep_alive),
-      tostring(request.timeout or 30000),
+      tostring(timeout),
       tostring(request.stream ~= false),
       tostring(think)
     )
@@ -270,7 +271,7 @@ function M.chat(request)
     return curl.stream_json_lines {
       url = endpoint() .. '/api/chat',
       body = body,
-      timeout = request.timeout or 30000,
+      timeout = timeout,
       is_cancelled = request.is_cancelled,
       on_json_line = function(data)
         if is_cancelled(request) then
@@ -463,17 +464,21 @@ function M.chat(request)
         phase = 'loaded',
         message = 'Model loaded',
       }
+      emit_status {
+        phase = 'context',
+        message = 'Loading prompt context',
+      }
     elseif model_loaded == false then
       emit_status {
         phase = 'loading',
         message = 'Loading model',
       }
+    else
+      emit_status {
+        phase = 'context',
+        message = 'Loading prompt context',
+      }
     end
-
-    emit_status {
-      phase = 'context',
-      message = 'Loading prompt context',
-    }
     job = run_chat()
     if request.register_http_job then
       request.register_http_job(job)
diff --git a/local/ai-provider/tests/ai_provider/core_spec.lua b/local/ai-provider/tests/ai_provider/core_spec.lua
index 140aec3..832e1d8 100644
--- a/local/ai-provider/tests/ai_provider/core_spec.lua
+++ b/local/ai-provider/tests/ai_provider/core_spec.lua
@@ -97,4 +97,45 @@ describe('ai-provider core', function()
     rawset(core, 'load_preferences', original_load_preferences)
     rawset(core, 'save_preferences', original_save_preferences)
   end)
+
+  it('persists registered source names across model selection updates', function()
+    ai_provider.setup {
+      default_provider = 'ollama',
+      providers = {
+        ollama = { default_model = 'gemma4:e2b' },
+      },
+    }
+
+    local core = require 'ai-provider.core'
+    local original_load_preferences = core.load_preferences
+    local original_save_preferences = core.save_preferences
+    local prefs = {}
+
+    rawset(core, 'load_preferences', function()
+      return vim.deepcopy(prefs)
+    end)
+    rawset(core, 'save_preferences', function(next_prefs)
+      prefs = vim.deepcopy(next_prefs)
+      return true
+    end)
+
+    assert.is_true(ai_provider.register_source('ai-commit-refine', { name = 'AI Commit: Refinement' }))
+    assert.are.same('AI Commit: Refinement', ai_provider.get_source_name 'ai-commit-refine')
+    assert.is_true(ai_provider.set_source_selection('ai-commit-refine', 'ollama', 'gemma4:e2b 64k'))
+
+    assert.are.same({
+      name = 'AI Commit: Refinement',
+      provider = 'ollama',
+      model = 'gemma4:e2b 64k',
+    }, prefs.sources['ai-commit-refine'])
+    assert.are.same({
+      provider = 'ollama',
+      model = 'gemma4:e2b 64k',
+      label = 'ollama/gemma4:e2b 64k',
+      name = 'AI Commit: Refinement',
+    }, ai_provider.get_source_selection 'ai-commit-refine')
+
+    rawset(core, 'load_preferences', original_load_preferences)
+    rawset(core, 'save_preferences', original_save_preferences)
+  end)
 end)
diff --git a/local/ai-provider/tests/ai_provider/ollama_options_spec.lua b/local/ai-provider/tests/ai_provider/ollama_options_spec.lua
index e72a4c8..4783e92 100644
--- a/local/ai-provider/tests/ai_provider/ollama_options_spec.lua
+++ b/local/ai-provider/tests/ai_provider/ollama_options_spec.lua
@@ -96,6 +96,30 @@ describe('ollama provider options', function()
     assert.are.same('10m', captured.keep_alive)
   end)
 
+  it('uses configured request timeout for streaming chat', function()
+    local captured_timeout = nil
+    package.loaded['ai-provider.curl'] = {
+      stream_json_lines = function(request)
+        captured_timeout = request.timeout
+        request.callback(0)
+        return { shutdown = function() end }
+      end,
+    }
+
+    local ollama = require 'ai-provider.providers.ollama'
+
+    ollama.chat {
+      model = 'gemma4:e2b',
+      prompt = 'Reply with exactly: ok',
+      provider_config = {
+        timeout = 180000,
+      },
+      callback = function() end,
+    }
+
+    assert.are.same(180000, captured_timeout)
+  end)
+
   it('sends configured thinking mode for reasoning model profiles', function()
     ---@type table|nil
     local captured = nil
@@ -298,7 +322,7 @@ describe('ollama provider options', function()
     end, 10)
 
     assert.are.same('loading', statuses[1])
-    assert.are.same('context', statuses[2])
-    assert.are.same('generating', statuses[3])
+    assert.are.same('generating', statuses[2])
+    assert.is_false(vim.tbl_contains(statuses, 'context'))
   end)
 end)
diff --git a/lua/plugins/ai-provider.lua b/lua/plugins/ai-provider.lua
index 9cbdc54..b97484e 100644
--- a/lua/plugins/ai-provider.lua
+++ b/lua/plugins/ai-provider.lua
@@ -14,17 +14,27 @@ return {
       ollama = {
         default_model = 'gemma4:e4b',
         context_size = 1024 * 8,
-        load_timeout = 120000,
+        timeout = 180000,
         keep_alive = '1h',
         models = {
           ['gemma4:e2b 32k'] = {
             model = 'gemma4:e2b',
             context_size = 1024 * 32,
           },
+          ['gemma4:e2b 32k fast'] = {
+            model = 'gemma4:e2b',
+            context_size = 1024 * 32,
+            think = false,
+          },
           ['gemma4:e2b 64k'] = {
             model = 'gemma4:e2b',
             context_size = 1024 * 64,
           },
+          ['gemma4:e2b 64k fast'] = {
+            model = 'gemma4:e2b',
+            context_size = 1024 * 64,
+            think = false,
+          },
           ['qwen3.5:4b 64k'] = {
             model = 'qwen3.5:4b',
             context_size = 1024 * 64,
-- 
2.51.2