feat: detect server crash before Generate/Chat calls

Co-Authored-By: Virgil <virgil@lethean.io>

model.go

@@ -4,6 +4,7 @@ package rocm
 
 import (
 	"context"
+	"fmt"
 	"iter"
 	"sync"
 
@@ -22,6 +23,13 @@ type rocmModel struct {
 
 // Generate streams tokens for the given prompt via llama-server's /v1/completions endpoint.
 func (m *rocmModel) Generate(ctx context.Context, prompt string, opts ...inference.GenerateOption) iter.Seq[inference.Token] {
+	if !m.srv.alive() {
+		m.mu.Lock()
+		m.lastErr = fmt.Errorf("rocm: server has exited: %w", m.srv.exitErr)
+		m.mu.Unlock()
+		return func(yield func(inference.Token) bool) {}
+	}
+
 	cfg := inference.ApplyGenerateOpts(opts)
 
 	req := llamacpp.CompletionRequest{
@@ -51,6 +59,13 @@ func (m *rocmModel) Generate(ctx context.Context, prompt string, opts ...inferen
 
 // Chat streams tokens from a multi-turn conversation via llama-server's /v1/chat/completions endpoint.
 func (m *rocmModel) Chat(ctx context.Context, messages []inference.Message, opts ...inference.GenerateOption) iter.Seq[inference.Token] {
+	if !m.srv.alive() {
+		m.mu.Lock()
+		m.lastErr = fmt.Errorf("rocm: server has exited: %w", m.srv.exitErr)
+		m.mu.Unlock()
+		return func(yield func(inference.Token) bool) {}
+	}
+
 	cfg := inference.ApplyGenerateOpts(opts)
 
 	chatMsgs := make([]llamacpp.ChatMessage, len(messages))

server.go

@@ -25,6 +25,16 @@ type server struct {
 	exitErr error
 }
 
+// alive reports whether the llama-server process is still running.
+func (s *server) alive() bool {
+	select {
+	case <-s.exited:
+		return false
+	default:
+		return true
+	}
+}
+
 // findLlamaServer locates the llama-server binary.
 // Checks ROCM_LLAMA_SERVER_PATH first, then PATH.
 func findLlamaServer() (string, error) {

server_test.go

@@ -3,10 +3,13 @@
 package rocm
 
 import (
+	"context"
+	"fmt"
 	"os"
 	"strings"
 	"testing"
 
+	"forge.lthn.ai/core/go-inference"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -95,3 +98,50 @@ func TestGuessModelType(t *testing.T) {
 		})
 	}
 }
+
+func TestServerAlive_Running(t *testing.T) {
+	s := &server{exited: make(chan struct{})}
+	assert.True(t, s.alive())
+}
+
+func TestServerAlive_Exited(t *testing.T) {
+	exited := make(chan struct{})
+	close(exited)
+	s := &server{exited: exited, exitErr: fmt.Errorf("process killed")}
+	assert.False(t, s.alive())
+}
+
+func TestGenerate_ServerDead(t *testing.T) {
+	exited := make(chan struct{})
+	close(exited)
+	s := &server{
+		exited:  exited,
+		exitErr: fmt.Errorf("process killed"),
+	}
+	m := &rocmModel{srv: s}
+
+	var count int
+	for range m.Generate(context.Background(), "hello") {
+		count++
+	}
+	assert.Equal(t, 0, count)
+	assert.ErrorContains(t, m.Err(), "server has exited")
+}
+
+func TestChat_ServerDead(t *testing.T) {
+	exited := make(chan struct{})
+	close(exited)
+	s := &server{
+		exited:  exited,
+		exitErr: fmt.Errorf("process killed"),
+	}
+	m := &rocmModel{srv: s}
+
+	msgs := []inference.Message{{Role: "user", Content: "hello"}}
+	var count int
+	for range m.Chat(context.Background(), msgs) {
+		count++
+	}
+	assert.Equal(t, 0, count)
+	assert.ErrorContains(t, m.Err(), "server has exited")
+}