refactor(app): streamline chatbot architecture and add tool integration

.github/workflows/verify.yml

@@ -0,0 +1,30 @@
+name: Core Functionality Check
+
+on: [push, pull_request]
+
+jobs:
+  verify:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python 3.11
+      uses: actions/setup-python@v4
+      with:
+        python-version: "3.11"
+    - name: Install requirements
+      run: |
+        python -m pip install --upgrade pip
+        pip install -r requirements.txt
+    - name: Run basic functionality test
+      run: |
+        python -c "
+        try:
+            from app import OllamaChat
+            print('✅ OllamaChat imported successfully')
+
+            # Skip actual initialization to avoid API calls
+            print('✅ Test passed (initialization skipped to avoid API calls)')
+        except ImportError as e:
+            print(f'❌ Import failed: {str(e)}')
+            raise
+        "

.gitignore

@@ -33,3 +33,6 @@ Thumbs.db
 # Project specific
 temp/
 uploads/
+
+# Text-to-speech generated files
+temp_*.mp3

README.md

@@ -1,236 +1,31 @@
-# 🤖 RAG-Enabled Local Chatbot
-
-A Retrieval-Augmented Generation (RAG) chatbot using local LLMs and vector storage for document-aware conversations.
-
-## 📸 Screenshots
-
-![Document-Aware Chat Interface](./assets/chat-interface.png)
-*The chatbot analyzing a PDF about DeFi and blockchain technology, demonstrating document-aware responses*
-
-## Key Features
-- 📄 Process and understand multiple document formats
-- 🔍 Retrieve relevant context from documents
-- 💬 Natural conversation with document awareness
-- 🏃 Fast local processing with Ollama
-- 🔒 Privacy-focused (all data stays local)
-
-## 🏗️ System Architecture
-
+# Ollama Chatbot
+
+## Project Overview
+A Streamlit-based chatbot interface powered by Ollama with:
+- Natural language conversations
+- Document processing (PDFs, images)
+- Built-in tools (time, date, document summary)
+- Text-to-speech functionality
+
+## Latest Features
+### 🆕 Enhanced Features
+- **Tool Integration**: Built-in tools for common queries
+  - Date lookup
+  - Time lookup (with timezone support)
+  - Document summarization
+- **Audio Features**: Text-to-speech with playback controls
+- **Document Processing**: Support for PDFs and images
+- **Clean File Management**: No temporary files left behind
+
+## Architecture
 ```mermaid
 graph TD
-    subgraph UserInterface["Streamlit Interface"]
-        Upload[Document Upload]
-        Chat[Chat Interface]
-        History[Chat History]
-    end
-
-    subgraph RAGPipeline["RAG Pipeline"]
-        subgraph DocumentProcessing["Document Processing"]
-            DP[Document Processor]
-            TS[Text Splitter]
-            OCR[Tesseract OCR]
-        end
-
-        subgraph Embeddings["Embedding Layer"]
-            OE[Ollama Embeddings]
-            VDB[Vector Database]
-        end
-
-        subgraph Retrieval["Context Retrieval"]
-            CS[Context Search]
-            CR[Context Ranking]
-        end
-    end
-
-    subgraph LLMLayer["Local LLM Layer"]
-        OL[Ollama - Mistral]
-        PM[Prompt Management]
-    end
-
-    Upload --> DP
-    DP --> |PDFs/Text| TS
-    DP --> |Images| OCR
-    OCR --> TS
-    TS --> OE
-    OE --> VDB
-    Chat --> CS
-    CS --> VDB
-    VDB --> CR
-    CR --> PM
-    PM --> OL
-    OL --> History
+    A[User Interface] -->|Query| B[Streamlit App]
+    B -->|Tool Trigger| C[Tool Registry]
+    B -->|LLM Query| D[Ollama API]
+    C --> E[Date Tool]
+    C --> F[Time Tool]
+    C --> G[Document Tool]
+    B -->|Process| H[Documents]
+    B -->|Convert| I[Text-to-Speech]
 ```
-
-## 🔄 RAG Implementation Flow
-
-```mermaid
-sequenceDiagram
-    participant U as User
-    participant I as Interface
-    participant R as RAG System
-    participant L as Local LLM
-
-    U->>I: Upload Document
-    I->>R: Process Document
-    activate R
-    R->>R: Extract Text
-    R->>R: Split into Chunks
-    R->>R: Generate Embeddings
-    R->>R: Store in Vector DB
-    deactivate R
-
-    U->>I: Ask Question
-    I->>R: Process Query
-    activate R
-    R->>R: Generate Query Embedding
-    R->>R: Search Similar Chunks
-    R->>R: Rank Relevance
-    R->>R: Build Context
-    R->>L: Context + Query
-    activate L
-    L->>L: Generate Response
-    L->>I: Return Answer
-    deactivate L
-    deactivate R
-    I->>U: Display Response
-```
-
-## 🛠️ Technical Implementation
-
-### Local Models
-- **LLM**: Ollama (Mistral)
-  - Local inference
-  - No data leaves system
-  - Customizable parameters
-
-### RAG Components
-1. **Document Processing**
-   ```python
-   # Text splitting configuration
-   text_splitter = RecursiveCharacterTextSplitter(
-       chunk_size=1000,
-       chunk_overlap=200,
-       length_function=len,
-   )
-   ```
-
-2. **Embedding Generation**
-   ```python
-   embeddings = OllamaEmbeddings(
-       model="nomic-embed-text",
-       base_url="http://localhost:11434"
-   )
-   ```
-
-3. **Vector Storage**
-   ```python
-   vectorstore = Chroma(
-       persist_directory="./chroma_db",
-       embedding_function=embeddings
-   )
-   ```
-
-### Supported Formats
-- 📄 PDF Documents
-- 📝 Text Files
-- 🖼️ Images (OCR-enabled)
-- 📊 Markdown Files
-
-## 🚀 Quick Start
-
-1. **System Requirements**
-```bash
-# Core dependencies
-brew install ollama
-brew install tesseract
-```
-
-2. **Environment Setup**
-```bash
-# Initialize project
-poetry install
-
-# Run setup script
-poetry run python setup.py
-
-# Start Ollama
-ollama serve
-```
-
-3. **Launch Application**
-```bash
-poetry run streamlit run app.py
-```
-
-## 🔧 Configuration
-
-### Environment Variables
-```env
-OLLAMA_BASE_URL=http://localhost:11434
-CHUNK_SIZE=1000
-CHUNK_OVERLAP=200
-```
-
-### LLM Settings
-```python
-llm = ChatOllama(
-    model="mistral",
-    temperature=0.7,
-    base_url="http://localhost:11434"
-)
-```
-
-## 📊 Performance Considerations
-
-1. **Memory Usage**
-   - Vector DB scaling
-   - Document chunk size
-   - Embedding cache
-
-2. **Processing Speed**
-   - OCR optimization
-   - Batch processing
-   - Concurrent operations
-
-3. **Response Quality**
-   - Context window size
-   - Chunk overlap
-   - Relevance threshold
-
-## 🔍 Debugging
-
-```bash
-# Check Ollama status
-curl http://localhost:11434/api/version
-
-# Verify vector store
-poetry run python -c "import chromadb; print(chromadb.__version__)"
-
-# Test OCR
-poetry run python -c "import pytesseract; print(pytesseract.get_tesseract_version())"
-```
-
-## 🐛 Known Issues
-
-1. **Image Processing**
-   - OCR quality varies with image clarity
-   - Large images may require preprocessing
-   - PNG transparency can affect OCR
-
-2. **Vector Storage**
-   - ChromaDB requires periodic optimization
-   - Large collections need index management
-   - Memory usage scales with document count
-
-## 🔒 Security
-
-- All processing done locally
-- No external API calls
-- Data remains on system
-- Configurable access controls
-
-## 📚 References
-
-- [LangChain Documentation](https://python.langchain.com/docs/get_started/introduction)
-- [Ollama GitHub](https://github.com/ollama/ollama)
-- [ChromaDB Documentation](https://docs.trychroma.com/)