chat: fix #includes with include-what-you-use (#2401)

Also use qGuiApp instead of qApp.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>

gpt4all-chat/embllm.h

@@ -1,13 +1,20 @@
 #ifndef EMBLLM_H
 #define EMBLLM_H
 
+#include <QByteArray>
 #include <QNetworkAccessManager>
-#include <QNetworkReply>
 #include <QObject>
+#include <QString>
 #include <QStringList>
 #include <QThread>
+#include <QVariant>
+#include <QVector>
 
-#include "../gpt4all-backend/llmodel.h"
+#include <atomic>
+#include <vector>
+
+class LLModel;
+class QNetworkAccessManager;
 
 struct EmbeddingChunk {
     int folder_id;