add script to update faiss, and update faiss

Author: samcraftt

server_api/chatbot/faiss_index/index.faiss

@@ -0,0 +1,32 @@
+# client/src/App.js
+
+Root component of the PyTC Client application. Wraps the app in context providers and performs cache clearing on boot.
+
+## Structure
+
+```
+App
+  ├── ContextWrapper (GlobalContext)
+  │   └── YamlContextWrapper (YamlContext)
+  │       └── CacheBootstrapper
+  │           └── MainContent (Views)
+```
+
+## Components
+
+### `CacheBootstrapper`
+- Runs `resetFileState()` from `AppContext` on mount
+- Clears local cache (files, fileList, etc.) before rendering children
+- Renders nothing until cache is cleared; then renders main content
+
+### `MainContent`
+- Renders `Views` component (main application layout with tabs)
+
+## Contexts
+
+- **AppContext** — Global state (files, configs, paths, etc.)
+- **YamlContext** — YAML-specific state (GPUs, batch size, etc.)
+
+## Usage
+
+Used as the root in `index.js` via `ReactDOM.createRoot`.

server_api/chatbot/faiss_index/index.pkl

@@ -0,0 +1,47 @@
+# client/src/api.js
+
+Central API client for the PyTC Client frontend. Exposes functions for all backend communication.
+
+## API Base URL
+
+`${REACT_APP_SERVER_PROTOCOL || "http"}://${REACT_APP_SERVER_URL || "localhost:4242"}`
+
+## Exports
+
+### `apiClient`
+Axios instance with base URL, `withCredentials: true`. Used for general requests (e.g. `/files`, `/files/upload`).
+
+### Visualization
+- **`getNeuroglancerViewer(image, label, scales)`** — Launches Neuroglancer viewer. Accepts file objects or path strings. Uses FormData for browser uploads.
+
+### File Checks
+- **`checkFile(file)`** — POST to `/check_files` to detect if file is likely a label (heuristic).
+
+### Generic
+- **`makeApiRequest(url, method, data)`** — HTTP request helper with JSON Content-Type.
+
+### Model Training
+- **`startModelTraining(trainingConfig, logPath, outputPath)`** — Injects OUTPUT_PATH into YAML config and POSTs to `/start_model_training`
+- **`stopModelTraining()`**
+- **`getTrainingStatus()`**
+- **`getTensorboardURL()`**
+
+### Model Inference
+- **`startModelInference(inferenceConfig, outputPath, checkpointPath)`** — Injects OUTPUT_PATH into YAML, sets NUM_GPUS=1, POSTs to `/start_model_inference`
+- **`stopModelInference()`**
+- **`getInferenceStatus()`**
+
+### Chatbot
+- **`queryChatBot(query)`** — POST to `/chat/query`
+- **`clearChat()`** — POST to `/chat/clear`
+
+### Config Presets
+- **`getConfigPresets()`** — GET `/pytc/configs`
+- **`getConfigPresetContent(path)`** — GET `/pytc/config`
+- **`getModelArchitectures()`** — GET `/pytc/architectures`
+
+## Helper Functions
+
+- `buildFilePath(file)` — Extracts path from various file object shapes ( Ant Design Upload, folderPath+name, etc.)
+- `hasBrowserFile(file)` — Checks if file is a browser File object
+- `handleError(error)` — Throws errors with response detail when available

server_api/chatbot/file_summaries/App.md

@@ -1,65 +1,25 @@
-# Chatbot Component
+# client/src/components/Chatbot.js
 
-## Overview
-
-The ChatBot component provides an AI-powered conversational interface for user support and assistance. It integrates with a RAG (Retrieval-Augmented Generation) backend to provide intelligent responses about the PyTorch Connectomics platform.
-
-## Functionality
-
-- **Real-time Chat Interface**: Provides a full-screen chat interface with message history
-- **Session Management**: Creates and maintains chat sessions with unique session IDs
-- **Message Persistence**: Saves chat history and session data in localStorage
-- **AI Integration**: Connects to backend RAG system for intelligent responses
-- **User Experience**: Clean, modern UI with message bubbles and typing indicators
-
-## Key Features
-
-- **Session-based Conversations**: Each chat session maintains context and memory
-- **Message History**: Persistent storage of conversation history
-- **Error Handling**: Graceful error handling for network issues
-- **Responsive Design**: Full-height layout with proper scrolling
-- **Input Validation**: Prevents empty message sending
+AI assistant panel for helping users navigate PyTC Client. Renders a chat UI with message history, Markdown rendering, and server-backed responses.
 
 ## Props
 
-- `onClose`: Function to close the chat interface
-
-## State Management
-
-- `messages`: Array of chat messages with user/bot distinction
-- `inputValue`: Current text input value
-- `sessionId`: Unique identifier for the chat session
-- `isSending`: Boolean indicating if a message is being processed
-
-## API Integration
-
-- **createChatSession()**: Creates a new chat session with the backend
-- **queryChatBot(sessionId, query)**: Sends user queries to the AI assistant
-
-## User Interactions
-
-1. **Opening Chat**: Click the message icon to open the chat interface
-2. **Sending Messages**: Type in the text area and press Enter or click Send
-3. **Session Persistence**: Chat history is automatically saved and restored
-4. **Closing Chat**: Click the X button to close the interface
+- **`onClose`** — Callback when user closes the chat (e.g. drawer)
 
-## Technical Implementation
+## Features
 
-- Uses Ant Design components for UI consistency
-- Implements proper cleanup with useEffect hooks
-- Handles async operations with proper error boundaries
-- Manages local storage for data persistence
+- **Message persistence** — Saves messages to `localStorage` under `chatMessages`
+- **Markdown rendering** — Uses `react-markdown` with `remarkGfm` for lists, tables, code blocks
+- **Keyboard shortcut** — Enter (without Shift) sends message
+- **Clear chat** — Popconfirm to clear; calls `clearChat()` API and resets to initial greeting
 
-## Use Cases
+## API Calls
 
-- **User Support**: Help users navigate the platform and understand features
-- **Troubleshooting**: Assist with common issues and error resolution
-- **Feature Guidance**: Provide step-by-step instructions for complex workflows
-- **Documentation Access**: Answer questions about platform capabilities
+- `queryChatBot(query)` — Sends user message, displays response
+- `clearChat()` — Clears server-side history and local state
 
-## Integration Points
+## UI Layout
 
-- **Global Context**: Accesses shared application state
-- **API Layer**: Communicates with backend RAG system
-- **Local Storage**: Persists user data across sessions
-- **UI Framework**: Integrates with Ant Design component library
+- Header: "AI Assistant" title, Clear button, Close button
+- Scrollable message list (user messages right-aligned, blue; bot messages left-aligned, gray)
+- TextArea + Send button at bottom