v2.2.3 - add server settings

Added the ability to change the settings of the LM Studio server.

port - changes the port to look for the server at
max_tokens - changes the maximum response length from LM Studio.  The maximum response from LM Studio + your question and context provided to LM Studio, must not exceed the "Context Length" setting within LM Studio.

config.yaml

@@ -29,15 +29,17 @@ AVAILABLE_MODELS:
 - sentence-transformers/gtr-t5-large
 - sentence-transformers/gtr-t5-base
 COMPUTE_DEVICE: cuda
-EMBEDDING_MODEL_NAME: C:/PATH/Scripts/LM Search Vector Database/v2_0_1 - working/Embedding_Models/hkunlp--instructor-xl
+EMBEDDING_MODEL_NAME: C:/PATH/Scripts/ChromaDB-Plugin-for-LM-Studio/v2_2_2 - working/Embedding_Models/BAAI--bge-base-en-v1.5
+server:
+  api_key: ''
+  connection_str: http://localhost:1234/v1
+  model_max_tokens: 500
+  model_temperature: 0.1
+  prefix: '[INST]'
+  suffix: '[/INST]'
 styles:
   button: 'background-color: #323842; color: light gray; font: 10pt "Segoe UI Historic";
     width: 29;'
   frame: 'background-color: #161b22;'
   input: 'background-color: #2e333b; color: light gray; font: 13pt "Segoe UI Historic";'
   text: 'background-color: #092327; color: light gray; font: 12pt "Segoe UI Historic";'
-tabs:
-- name: Settings
-  placeholder: Placeholder text for Settings tab.
-- name: Models
-  placeholder: Placeholder text for Models tab.

gui.py

@@ -49,7 +49,7 @@ def init_ui(self):
         self.left_frame = QFrame()
         left_vbox = QVBoxLayout()
 
-        tab_widget = create_tabs(tabs_config)
+        tab_widget = create_tabs()
         left_vbox.addWidget(tab_widget)
 
         button_data = [

gui_tabs.py

@@ -1,30 +1,56 @@
 from PySide6.QtWebEngineWidgets import QWebEngineView
-from PySide6.QtWidgets import QTextEdit, QTabWidget
+from PySide6.QtWidgets import QTextEdit, QTabWidget, QVBoxLayout, QWidget, QGroupBox
 from PySide6.QtCore import QUrl
 import os
+from gui_tabs_settings_server import ServerSettingsTab
+from gui_tabs_settings_models import ModelsSettingsTab
 
-def create_tabs(tabs_config):
+def create_tabs():
     tab_widget = QTabWidget()
     tab_widget.setTabPosition(QTabWidget.South)
 
-    tab_widgets = [QTextEdit(tab.get('placeholder', '')) for tab in tabs_config]
-    for i, tab in enumerate(tabs_config):
-        tab_widget.addTab(tab_widgets[i], tab.get('name', ''))
+    user_manual_folder = os.path.join(os.path.dirname(__file__), 'User_Manual')
+
+    # Create Settings tab
+    settings_tab = QWidget()
+    settings_layout = QVBoxLayout()
+
+    # Add Server and EmbeddingModel settings to Settings tab
+    server_settings = ServerSettingsTab()
+    server_group = QGroupBox("Server/LLM Settings")
+    server_layout = QVBoxLayout()
+    server_layout.addWidget(server_settings)
+    server_group.setLayout(server_layout)
+
+    models_settings = ModelsSettingsTab()
+    models_group = QGroupBox("Embedding Models Settings")
+    models_layout = QVBoxLayout()
+    models_layout.addWidget(models_settings)
+    models_group.setLayout(models_layout)
+
+    settings_layout.addWidget(server_group)
+    settings_layout.addWidget(models_group)
+    settings_tab.setLayout(settings_layout)
 
-    # Adding the Tutorial tab
+    tab_widget.addTab(settings_tab, 'Settings')
+
+    # Create Models tab
+    models_tab = QTextEdit("Placeholder text for Models tab.")
+    tab_widget.addTab(models_tab, 'Models')
+
+    # Create Floating Point Formats tab
     tutorial_tab = QWebEngineView()
-    tab_widget.addTab(tutorial_tab, 'Floating Point Formats')
-    user_manual_folder = os.path.join(os.path.dirname(__file__), 'User_Manual')
     tutorial_html_path = os.path.join(user_manual_folder, 'number_format.html')
     tutorial_tab.setUrl(QUrl.fromLocalFile(tutorial_html_path))
+    tab_widget.addTab(tutorial_tab, 'Floating Point Formats')
 
-    # Adding the Whisper tab
+    # Create Whisper tab
     whisper_tab = QWebEngineView()
     whisper_html_path = os.path.join(user_manual_folder, 'whisper_quants.html')
     whisper_tab.setUrl(QUrl.fromLocalFile(whisper_html_path))
     tab_widget.addTab(whisper_tab, 'Whisper')
 
-    # Adding the Tips tab
+    # Create Tips tab
     tips_tab = QWebEngineView()
     tips_html_path = os.path.join(user_manual_folder, 'tips.html')
     tips_tab.setUrl(QUrl.fromLocalFile(tips_html_path))

gui_tabs_settings_models.py

@@ -0,0 +1,11 @@
+from PySide6.QtWidgets import QWidget, QLabel, QVBoxLayout
+
+class ModelsSettingsTab(QWidget):
+    def __init__(self):
+        super(ModelsSettingsTab, self).__init__()
+
+        self.label = QLabel("Placeholder text for embedding model settings, which are coming soon.")
+
+        layout = QVBoxLayout()
+        layout.addWidget(self.label)
+        self.setLayout(layout)

gui_tabs_settings_server.py

@@ -0,0 +1,61 @@
+from PySide6.QtWidgets import QWidget, QLabel, QLineEdit, QPushButton, QGridLayout, QMessageBox
+from PySide6.QtGui import QIntValidator
+import yaml
+
+class ServerSettingsTab(QWidget):
+    def __init__(self):
+        super(ServerSettingsTab, self).__init__()
+
+        with open('config.yaml', 'r') as file:
+            config_data = yaml.safe_load(file)
+            self.connection_str = config_data.get('server', {}).get('connection_str', '')
+            self.current_port = self.connection_str.split(":")[-1].split("/")[0]
+            self.current_max_tokens = config_data.get('server', {}).get('model_max_tokens', '')
+
+        # Server port setting widgets
+        self.port_label = QLabel(f"Current Port: {self.current_port}")
+        self.new_port_edit = QLineEdit()
+        self.new_port_edit.setPlaceholderText("Enter new port...")
+        self.new_port_edit.setValidator(QIntValidator())
+
+        # Max token setting widgets
+        self.max_tokens_label = QLabel(f"Max Tokens: {self.current_max_tokens}")
+        self.new_max_tokens_edit = QLineEdit()
+        self.new_max_tokens_edit.setPlaceholderText("Enter new max tokens...")
+        self.new_max_tokens_edit.setValidator(QIntValidator())
+
+        # Update config button
+        self.update_button = QPushButton("Update")
+        self.update_button.clicked.connect(self.update_config)
+
+        layout = QGridLayout()
+        layout.addWidget(self.port_label, 0, 0)
+        layout.addWidget(self.new_port_edit, 0, 1)
+        layout.addWidget(self.max_tokens_label, 1, 0)
+        layout.addWidget(self.new_max_tokens_edit, 1, 1)
+        layout.addWidget(self.update_button, 2, 1)
+
+        self.setLayout(layout)
+
+    def update_config(self):
+        new_port = self.new_port_edit.text()
+        new_max_tokens = self.new_max_tokens_edit.text()
+
+        # Update config.yaml if at least one setting was changed - otherwise display message
+        if not new_port and not new_max_tokens:
+            QMessageBox.warning(self, 'No Updates', 'No new values were entered.')
+            return
+
+        with open('config.yaml', 'r') as file:
+            config_data = yaml.safe_load(file)
+
+        if new_port:
+            config_data['server']['connection_str'] = self.connection_str.replace(self.current_port, new_port)
+            self.port_label.setText(f"Current Port: {new_port}")
+
+        if new_max_tokens:
+            config_data['server']['model_max_tokens'] = int(new_max_tokens)
+            self.max_tokens_label.setText(f"Max Tokens: {new_max_tokens}")
+
+        with open('config.yaml', 'w') as file:
+            yaml.safe_dump(config_data, file)

server_connector.py

@@ -15,17 +15,25 @@
     chroma_db_impl="duckdb+parquet", persist_directory=PERSIST_DIRECTORY, anonymized_telemetry=False
 )
 
-openai.api_base = 'http://localhost:1234/v1'
-openai.api_key = ''
-
-prefix = "[INST]"
-suffix = "[/INST]"
-
 def connect_to_local_chatgpt(prompt):
+    with open('config.yaml', 'r') as config_file:
+        config = yaml.safe_load(config_file)
+        server_config = config.get('server', {})
+        openai_api_base = server_config.get('connection_str')
+        openai_api_key = server_config.get('api_key')
+        prefix = server_config.get('prefix')
+        suffix = server_config.get('suffix')
+        model_temperature = server_config.get('model_temperature')
+        model_max_tokens = server_config.get('model_max_tokens')
+
+    openai.api_base = openai_api_base
+    openai.api_key = openai_api_key
+
     formatted_prompt = f"{prefix}{prompt}{suffix}"
     response = openai.ChatCompletion.create(
         model="local model",
-        temperature=0.1,
+        temperature=model_temperature,
+        max_tokens=model_max_tokens,
         messages=[{"role": "user", "content": formatted_prompt}]
     )
     return response.choices[0].message["content"]