Package restructure and API changes, several fixes

- More emphasis on `api` package. It now holds database model structs
  from `lmcli/models` (which is now gone) as well as the tool spec,
  call, and result types. `tools.Tool` is now `api.ToolSpec`.
  `api.ChatCompletionClient` was renamed to
  `api.ChatCompletionProvider`.

- Change ChatCompletion interface and implementations to no longer do
  automatic tool call recursion - they simply return a ToolCall message
  which the caller can decide what to do with (e.g. prompt for user
  confirmation before executing)

- `api.ChatCompletionProvider` functions have had their ReplyCallback
  parameter removed, as now they only return a single reply.

- Added a top-level `agent` package, moved the current built-in tools
  implementations under `agent/toolbox`. `tools.ExecuteToolCalls` is now
  `agent.ExecuteToolCalls`.

- Fixed request context handling in openai, google, ollama (use
  `NewRequestWithContext`), cleaned up request cancellation in TUI

- Fix tool call tui persistence bug (we were skipping message with empty
  content)

- Now handle tool calling from TUI layer

TODO:
- Prompt users before executing tool calls
- Automatically send tool results to the model (or make this toggleable)

pkg/tui/views/chat/chat.go

@@ -4,7 +4,6 @@ import (
 	"time"
 
 	"git.mlow.ca/mlow/lmcli/pkg/api"
-	models "git.mlow.ca/mlow/lmcli/pkg/lmcli/model"
 	"git.mlow.ca/mlow/lmcli/pkg/tui/shared"
 	"github.com/charmbracelet/bubbles/cursor"
 	"github.com/charmbracelet/bubbles/spinner"
@@ -16,37 +15,39 @@ import (
 
 // custom tea.Msg types
 type (
-	// sent on each chunk received from LLM
-	msgResponseChunk api.Chunk
-	// sent when response is finished being received
-	msgResponseEnd string
-	// a special case of common.MsgError that stops the response waiting animation
-	msgResponseError error
-	// sent on each completed reply
-	msgResponse models.Message
 	// sent when a conversation is (re)loaded
 	msgConversationLoaded struct {
-		conversation *models.Conversation
-		rootMessages []models.Message
+		conversation *api.Conversation
+		rootMessages []api.Message
 	}
 	// sent when a new conversation title generated
 	msgConversationTitleGenerated string
-	// sent when a conversation's messages are laoded
-	msgMessagesLoaded []models.Message
 	// sent when the conversation has been persisted, triggers a reload of contents
 	msgConversationPersisted struct {
 		isNew        bool
-		conversation *models.Conversation
-		messages     []models.Message
+		conversation *api.Conversation
+		messages     []api.Message
 	}
+	// sent when a conversation's messages are laoded
+	msgMessagesLoaded []api.Message
+	// a special case of common.MsgError that stops the response waiting animation
+	msgChatResponseError error
+	// sent on each chunk received from LLM
+	msgChatResponseChunk api.Chunk
+	// sent on each completed reply
+	msgChatResponse *api.Message
+	// sent when the response is canceled
+	msgChatResponseCanceled struct{}
+	// sent when results from a tool call are returned
+	msgToolResults []api.ToolResult
 	// sent when the given message is made the new selected reply of its parent
-	msgSelectedReplyCycled *models.Message
+	msgSelectedReplyCycled *api.Message
 	// sent when the given message is made the new selected root of the current conversation
-	msgSelectedRootCycled *models.Message
+	msgSelectedRootCycled *api.Message
 	// sent when a message's contents are updated and saved
-	msgMessageUpdated *models.Message
+	msgMessageUpdated *api.Message
 	// sent when a message is cloned, with the cloned message
-	msgMessageCloned *models.Message
+	msgMessageCloned *api.Message
 )
 
 type focusState int
@@ -77,14 +78,14 @@ type Model struct {
 
 	// app state
 	state           state // current overall status of the view
-	conversation    *models.Conversation
-	rootMessages    []models.Message
-	messages        []models.Message
+	conversation    *api.Conversation
+	rootMessages    []api.Message
+	messages        []api.Message
 	selectedMessage int
 	editorTarget    editorTarget
 	stopSignal      chan struct{}
-	replyChan       chan models.Message
-	replyChunkChan  chan api.Chunk
+	replyChan       chan api.Message
+	chatReplyChunks chan api.Chunk
 	persistence     bool // whether we will save new messages in the conversation
 
 	// ui state
@@ -111,12 +112,12 @@ func Chat(shared shared.Shared) Model {
 		Shared: shared,
 
 		state:        idle,
-		conversation: &models.Conversation{},
+		conversation: &api.Conversation{},
 		persistence:  true,
 
-		stopSignal:     make(chan struct{}),
-		replyChan:      make(chan models.Message),
-		replyChunkChan: make(chan api.Chunk),
+		stopSignal:      make(chan struct{}),
+		replyChan:       make(chan api.Message),
+		chatReplyChunks: make(chan api.Chunk),
 
 		wrap:            true,
 		selectedMessage: -1,
@@ -144,8 +145,8 @@ func Chat(shared shared.Shared) Model {
 
 	system := shared.Ctx.GetSystemPrompt()
 	if system != "" {
-		m.messages = []models.Message{{
-			Role:    models.MessageRoleSystem,
+		m.messages = []api.Message{{
+			Role:    api.MessageRoleSystem,
 			Content: system,
 		}}
 	}
@@ -166,6 +167,5 @@ func Chat(shared shared.Shared) Model {
 func (m Model) Init() tea.Cmd {
 	return tea.Batch(
 		m.waitForResponseChunk(),
-		m.waitForResponse(),
 	)
 }