Wip-bring-back-s3-mock-issue-36

.env.example

@@ -78,22 +78,28 @@ FEATURE_AGENT_MODE_AVAILABLE=true
 AGENT_LOOP_STRATEGY=think-act
 # (Adjust above to stage rollouts. For a bare-bones chat set them all to false.)
 
-APP_LOG_DIR=/workspaces/atlas-ui-3-11/logs
+APP_LOG_DIR=/workspaces/atlas-ui-3/logs
 
 CAPABILITY_TOKEN_SECRET=blablah
 
 #############################################
 # S3/MinIO Storage Configuration
 #############################################
-# MinIO endpoint (use localhost for local dev, minio for docker-compose)
-S3_ENDPOINT=http://localhost:9000
-S3_BUCKET_NAME=atlas-files
-S3_ACCESS_KEY=minioadmin
-S3_SECRET_KEY=minioadmin
-S3_REGION=us-east-1
-S3_TIMEOUT=30
-
-S3_USE_SSL=false
+# Choose ONE option below (comment out the other)
+
+# --- Option 1: Mock S3 (Default - No Docker required) ---
+USE_MOCK_S3=true
+
+# --- Option 2: MinIO (Requires Docker) ---
+# Uncomment below and set USE_MOCK_S3=false to use MinIO
+# USE_MOCK_S3=false
+# S3_ENDPOINT=http://localhost:9000
+# S3_BUCKET_NAME=atlas-files  # Must match bucket created in docker-compose.yml
+# S3_ACCESS_KEY=minioadmin
+# S3_SECRET_KEY=minioadmin
+# S3_REGION=us-east-1
+# S3_TIMEOUT=30
+# S3_USE_SSL=false
 
 
 SECURITY_CSP_VALUE="default-src 'self'; img-src 'self' data: blob:; script-src 'self'; style-src 'self' 'unsafe-inline'; connect-src 'self'; frame-src 'self' blob: data:; frame-ancestors 'self'"

.gitignore

@@ -1,6 +1,6 @@
 *.pptx
 *.jsonl
-
+.ruff_cache
 # Environment variables
 .env
 .claude

CLAUDE.md

@@ -61,12 +61,16 @@ mkdir -p logs
 ```bash
 bash agent_start.sh
 ```
-This script handles: killing old processes, clearing logs, building frontend, starting mock S3, and starting backend.
+This script handles: killing old processes, clearing logs, building frontend, starting S3 storage (MinIO or Mock based on `USE_MOCK_S3` in `.env`), and starting backend.
 
 **Options:**
 - `bash agent_start.sh -f` - Only rebuild frontend
 - `bash agent_start.sh -b` - Only restart backend
 
+**Note:** The script automatically reads `USE_MOCK_S3` from `.env`:
+- If `true`: Uses in-process Mock S3 (no Docker)
+- If `false`: Starts MinIO via docker-compose
+
 ### Manual Development Workflow
 
 **Frontend Build (CRITICAL):**
@@ -82,11 +86,24 @@ cd backend
 python main.py  # NEVER use uvicorn --reload (causes problems)
 ```
 
-**Mock S3 (Optional):**
-```bash
-cd mocks/s3-mock
-python main.py  # Runs on http://127.0.0.1:8003
-```
+**S3 Storage (Mock vs MinIO):**
+
+The project supports two S3 storage backends:
+
+1. **Mock S3 (Default, Recommended for Development)**
+   - Set `USE_MOCK_S3=true` in `.env`
+   - Uses in-process FastAPI TestClient (no Docker required)
+   - Files stored in `minio-data/chatui/` on disk
+   - No external server needed - integrated directly into backend
+   - Faster startup, simpler development workflow
+
+2. **MinIO (Production-like)**
+   - Set `USE_MOCK_S3=false` in `.env`
+   - Requires Docker: `docker-compose up -d minio minio-init`
+   - Full S3 compatibility with all features
+   - Use for testing production-like scenarios
+
+The mock automatically activates when the backend starts if `USE_MOCK_S3=true`.
 
 ### Testing
 
@@ -246,9 +263,11 @@ Three agent loop strategies implement different reasoning patterns:
 - **Act** (`backend/application/chat/agent/act_loop.py`): Pure action loop without explicit reasoning steps, fastest with minimal overhead. LLM calls tools directly and signals completion via the "finished" tool
 
 ### File Storage
-S3-compatible storage via `backend/modules/file_storage/s3_client.py`:
-- Production: Real S3 or S3-compatible service
-- Development: Mock S3 (`mocks/s3-mock/`)
+S3-compatible storage via `backend/modules/file_storage/`:
+- Production/MinIO: `s3_client.py` - boto3-based client for real S3/MinIO
+- Development: `mock_s3_client.py` - TestClient-based in-process mock (no Docker)
+- Controlled by `USE_MOCK_S3` env var (default: true)
+- Both implementations share the same interface
 
 ### Security Middleware Stack
 ```

GEMINI.md

@@ -0,0 +1,65 @@
+# GEMINI.md
+
+This file provides guidance to the Gemini AI agent when working with code in this repository.
+
+## Project Overview
+
+Atlas UI 3 is a full-stack LLM chat interface with Model Context Protocol (MCP) integration, supporting multiple LLM providers (OpenAI, Anthropic Claude, Google Gemini), RAG, and agentic capabilities.
+
+**Tech Stack:**
+- Backend: FastAPI + WebSockets, LiteLLM, FastMCP
+- Frontend: React 19 + Vite 7 + Tailwind CSS
+- Python Package Manager: **uv** (NOT pip!)
+- Configuration: Pydantic with YAML/JSON configs
+
+## Building and Running
+
+### Quick Start (Recommended)
+```bash
+bash agent_start.sh
+```
+This script handles: killing old processes, clearing logs, building frontend, and starting the backend.
+
+### Manual Development Workflow
+
+**Frontend Build (CRITICAL):**
+```bash
+cd frontend
+npm install
+npm run build  # Use build, NOT npm run dev (WebSocket issues)
+```
+
+**Backend Start:**
+```bash
+cd backend
+python main.py  # NEVER use uvicorn --reload (causes problems)
+```
+
+### Testing
+
+**Run all tests:**
+```bash
+./test/run_tests.sh all
+```
+
+**Individual test suites:**
+```bash
+./test/run_tests.sh backend
+./test/run_tests.sh frontend
+./test/run_tests.sh e2e
+```
+
+## Development Conventions
+
+- **Python Package Manager**: **ALWAYS use `uv`**, never pip or conda.
+- **Frontend Development**: **NEVER use `npm run dev`**, it has WebSocket connection problems. Always use `npm run build`.
+- **Backend Development**: **NEVER use `uvicorn --reload`**, it causes problems.
+- **File Naming**: Do not use generic names like `utils.py` or `helpers.py`. Use descriptive names that reflect the file's purpose.
+- **No Emojis**: No emojis should ever be added in this repo.
+- **Linting**: Run `ruff check backend/` for Python and `npm run lint` for the frontend before committing.
+
+
+Also read. 
+/workspaces/atlas-ui-3/.github/copilot-instructions.md
+
+and CLAUDE.md

agent_start.sh

@@ -20,14 +20,26 @@ done
 # Configuration
 USE_NEW_FRONTEND=${USE_NEW_FRONTEND:-true}
 
-# Check if MinIO is running
-if ! docker ps | grep -q atlas-minio; then
-    echo "⚠️  MinIO is not running. Starting MinIO with docker-compose..."
-    docker-compose up -d minio minio-init
-    echo "✅ MinIO started successfully"
-    sleep 3
+# Read USE_MOCK_S3 from .env file
+if [ -f .env ]; then
+    USE_MOCK_S3=$(grep -E "^USE_MOCK_S3=" .env | cut -d '=' -f2)
 else
-    echo "✅ MinIO is already running"
+    USE_MOCK_S3="true"  # Default to mock if no .env
+fi
+
+# Only start MinIO if not using mock S3
+if [ "$USE_MOCK_S3" = "true" ]; then
+    echo "✅ Using Mock S3 (no Docker required)"
+else
+    # Check if MinIO is running
+    if ! docker ps | grep -q atlas-minio; then
+        echo "⚠️  MinIO is not running. Starting MinIO with docker-compose..."
+        docker-compose up -d minio minio-init
+        echo "✅ MinIO started successfully"
+        sleep 3
+    else
+        echo "✅ MinIO is already running"
+    fi
 fi
 
 # Kill any running uvicorn processes (skip if only rebuilding frontend)

backend/application/chat/preprocessors/message_builder.py

@@ -41,22 +41,27 @@ async def build_messages(
     ) -> List[Dict[str, Any]]:
         """
         Build messages array from session history and context.
-        
+
         Args:
             session: Current chat session
             include_files_manifest: Whether to append files manifest
-            
+
         Returns:
             List of messages ready for LLM call
         """
         # Get conversation history from session
         messages = session.history.get_messages_for_llm()
-        
+
         # Optionally add files manifest
         if include_files_manifest:
             session_context = build_session_context(session)
+            files_in_context = session_context.get("files", {})
+            logger.info(f"DEBUG: Session has {len(files_in_context)} files: {list(files_in_context.keys())}")
             files_manifest = file_utils.build_files_manifest(session_context)
             if files_manifest:
+                logger.info(f"DEBUG: Adding files manifest to messages: {files_manifest['content'][:100]}")
                 messages.append(files_manifest)
-
+            else:
+                logger.warning("DEBUG: No files manifest generated despite include_files_manifest=True")
+
         return messages

backend/infrastructure/app_factory.py

@@ -7,6 +7,7 @@
 from interfaces.transport import ChatConnectionProtocol
 from modules.config import ConfigManager
 from modules.file_storage import S3StorageClient, FileManager
+from modules.file_storage.mock_s3_client import MockS3StorageClient
 from modules.llm.litellm_caller import LiteLLMCaller
 from modules.mcp_tools import MCPToolManager
 from modules.rag import RAGClient
@@ -37,7 +38,12 @@ def __init__(self) -> None:
         )
 
         # File storage & manager
-        self.file_storage = S3StorageClient()
+        if self.config_manager.app_settings.use_mock_s3:
+            logger.info("Using MockS3StorageClient (in-process, no Docker required)")
+            self.file_storage = MockS3StorageClient()
+        else:
+            logger.info("Using S3StorageClient (MinIO/AWS S3)")
+            self.file_storage = S3StorageClient()
         self.file_manager = FileManager(self.file_storage)
 
         logger.info("AppFactory initialized")

backend/modules/config/manager.py

@@ -127,6 +127,7 @@ def agent_mode_available(self) -> bool:
     test_user: str = "[email protected]"  # Test user for development
 
     # S3/MinIO storage settings
+    use_mock_s3: bool = False  # Use in-process S3 mock (no Docker required)
     s3_endpoint: str = "http://localhost:9000"
     s3_bucket_name: str = "atlas-files"
     s3_access_key: str = "minioadmin"