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/chatllm.h

@@ -1,16 +1,29 @@
 #ifndef CHATLLM_H
 #define CHATLLM_H
 
-#include <QObject>
-#include <QThread>
-#include <QFileInfo>
-
-#include <memory>
-
-#include "database.h"
 #include "modellist.h"
+
 #include "../gpt4all-backend/llmodel.h"
 
+#include <QByteArray>
+#include <QElapsedTimer>
+#include <QFileInfo>
+#include <QObject>
+#include <QPair>
+#include <QString>
+#include <QThread>
+#include <QVector>
+#include <QtGlobal>
+
+#include <atomic>
+#include <cstdint>
+#include <memory>
+#include <string>
+
+class QDataStream;
+struct ResultInfo;
+template <typename T> class QList;
+
 enum LLModelType {
     GPTJ_,
     LLAMA_,