feat (interface/model): interface enhancements and biomistral again

Dockerfile

@@ -45,6 +45,9 @@ sleep 10\n\
 echo "Pulling phi3 model..."\n\
 ollama pull phi3\n\
 \n\
+echo "Pulling biomistral model..."\n\
+ollama pull cniongolo/biomistral\n\
+\n\
 echo "Models ready! Starting FastAPI server..."\n\
 exec uvicorn server:app --host 0.0.0.0 --port 7860\n\
 ' > /app/start.sh && chmod +x /app/start.sh

EXAMPLES.md

@@ -1,6 +1,8 @@
 # Pub/Sub Multi-Agent System - Example Configurations
 
-**Note**: This deployment uses **phi3** for all agents. phi3 is a capable general-purpose model suitable for most tasks.
+**Note**: This deployment includes **phi3** (general purpose) and **cniongolo/biomistral** (medical/scientific) models.
+
+**Important**: Use `{question}` in prompts to access the user's question, `{schema}` for database schema, and `{input}` for messages from subscribed topics.
 
 ## Example 1: Natural Language to SQL Pipeline (using phi3)
 
@@ -11,7 +13,7 @@
 Given this database schema:
 {schema}
 
-User question: {input}
+User question: {question}
 
 Analyze the question and identify:
 1. What tables are needed
@@ -34,6 +36,8 @@ Based on this analysis:
 Database schema:
 {schema}
 
+Original question: {question}
+
 Generate a SQL query that answers the user's question.
 Return ONLY the SQL query, no explanation.
 ```
@@ -56,22 +60,24 @@ If invalid, return "INVALID: " followed by the corrected query.
 ```
 - **Model**: phi3
 - **Subscribe Topic**: SQL_GENERATED
-- **Publish Topic**: RESULT
+- **Publish Topic**: FINAL
 
 ---
 
 ## Example 2: Medical Symptom Analysis (Using BioMistral)
 
+**User Question**: "Patient has fever, cough, and shortness of breath for 3 days"
+
 ### Agent 1: Symptom Extractor
 - **Title**: Symptom Extractor
 - **Prompt**:
 ```
-Patient report: {input}
+Patient report: {question}
 
 Extract all symptoms mentioned in the patient report.
 List them clearly and concisely.
 ```
-- **Model**: phi3
+- **Model**: cniongolo/biomistral
 - **Subscribe Topic**: START
 - **Publish Topic**: SYMPTOMS_EXTRACTED
 
@@ -89,7 +95,9 @@ Provide a preliminary medical assessment including:
 
 Note: This is for educational purposes only, not medical advice.
 ```
-- **Model**: phi3
+- **Model**: cniongolo/biomistral
+- **Subscribe Topic**: SYMPTOMS_EXTRACTED
+- **Publish Topic**: FINAL
 - **Subscribe Topic**: SYMPTOMS_EXTRACTED
 - **Publish Topic**: ASSESSMENT_COMPLETE
 
@@ -227,12 +235,18 @@ Note: This is for educational purposes only.
 1. **Clear Subscribe/Publish Topics**: Use descriptive names like "SQL_GENERATED" not just "STEP2"
 
 2. **Prompt Placeholders**: 
+   - Use `{question}` for the user's input question
    - Use `{input}` for the message from the subscribed topic
    - Use `{schema}` to access the database schema
-   - Both can be used together
+   - All three can be used together in any prompt
 
 3. **Model Selection**:
-   - **phi3**: Versatile general-purpose model suitable for all tasks in this system
+   - **phi3**: Versatile general-purpose model - text analysis, SQL, reasoning, classification
+   - **cniongolo/biomistral**: Medical and scientific tasks - diagnosis, clinical reasoning, biomedical analysis
+
+4. **Final Results**:
+   - Publish to topic "FINAL" to display results in the Final Result box
+   - Only the last agent in your chain should publish to FINAL
 
 4. **Chain Design**:
    - Start simple with 2-3 agents

QUICK_START.md

@@ -0,0 +1,142 @@
+# Quick Start Guide - Enhanced Pub/Sub System
+
+## New Features
+
+### 1. User Question Input
+- Located above the agent configuration
+- Enter your question here
+- Available in prompts using `{question}` placeholder
+- Gets published to START topic when you click Execute
+
+### 2. Final Result Display
+- Green-highlighted box at the top right
+- Automatically shows results from agents publishing to "FINAL" topic
+- **Important**: Make your last agent publish to "FINAL" topic to see results here
+
+### 3. BioMistral Model Support
+- Now includes `cniongolo/biomistral` for medical/scientific tasks
+- Use for healthcare, clinical reasoning, biomedical analysis
+
+## How to Use
+
+### Step 1: Enter Your Question
+In the "User Question" box, type your question:
+```
+What were the top 5 products by revenue in 2024?
+```
+
+### Step 2: Configure Agents
+Click "Add Agent" and set up your pipeline. Example:
+
+**Agent 1:**
+- Title: `SQL Generator`
+- Prompt: `Generate SQL for: {question}\nSchema: {schema}`
+- Model: `phi3`
+- Subscribe: `START`
+- Publish: `SQL_GENERATED`
+
+**Agent 2:**
+- Title: `SQL Formatter`
+- Prompt: `Format this SQL nicely: {input}`
+- Model: `phi3`
+- Subscribe: `SQL_GENERATED`
+- Publish: `FINAL` ← **Important!**
+
+### Step 3: Execute
+Click "Execute Pipeline" and watch:
+1. **Execution Log** shows each step
+2. **Final Result** displays the output from the agent that published to "FINAL"
+
+## Prompt Placeholders
+
+You can use three placeholders in agent prompts:
+
+| Placeholder | Contains | Example Use |
+|------------|----------|-------------|
+| `{question}` | User's input question | First agent in chain |
+| `{input}` | Message from subscribed topic | Middle/end agents |
+| `{schema}` | Database schema | Any agent needing schema |
+
+### Example Prompt Using All Three:
+```
+User asked: {question}
+
+Previous analysis: {input}
+
+Database schema:
+{schema}
+
+Now generate the final SQL query.
+```
+
+## Medical Analysis Example
+
+**User Question:** "Patient has fever 39°C, persistent cough, difficulty breathing"
+
+**Agent 1: Symptom Analyzer**
+- Prompt: `Extract symptoms from: {question}`
+- Model: `cniongolo/biomistral`
+- Subscribe: `START`
+- Publish: `SYMPTOMS`
+
+**Agent 2: Diagnosis**
+- Prompt: `Based on symptoms: {input}\nProvide differential diagnosis`
+- Model: `cniongolo/biomistral`
+- Subscribe: `SYMPTOMS`
+- Publish: `FINAL`
+
+Result appears in the Final Result box automatically!
+
+## Tips for Best Results
+
+1. **Always end with FINAL**: Your last agent should publish to "FINAL" topic to display results
+
+2. **Use {question} in first agent**: The START message contains the user's question
+
+3. **Chain logically**: Each agent builds on the previous:
+   ```
+   START → Analyze → Generate → Validate → FINAL
+   ```
+
+4. **Choose right model**:
+   - General tasks (SQL, text): use `phi3`
+   - Medical/scientific: use `cniongolo/biomistral`
+
+5. **Test incrementally**: Start with 2 agents, then add more
+
+## Common Patterns
+
+### Pattern 1: Analysis → Generation → Validation
+```
+Agent 1: Analyze {question} → ANALYZED
+Agent 2: Generate from {input} → GENERATED  
+Agent 3: Validate {input} → FINAL
+```
+
+### Pattern 2: Extract → Process → Summarize
+```
+Agent 1: Extract data from {question} → EXTRACTED
+Agent 2: Process {input} → PROCESSED
+Agent 3: Summarize {input} → FINAL
+```
+
+### Pattern 3: Medical Workflow
+```
+Agent 1: Categorize symptoms from {question} → CATEGORIZED
+Agent 2: Diagnose based on {input} → DIAGNOSIS
+Agent 3: Recommend treatment for {input} → FINAL
+```
+
+## Troubleshooting
+
+**Q: Final Result box is empty**
+- Check if your last agent publishes to "FINAL" (exact capitalization)
+
+**Q: Nothing happens when I click Execute**
+- Ensure at least one agent subscribes to "START"
+
+**Q: Agent doesn't receive the question**
+- Use `{question}` in the prompt, not `{input}` for the first agent
+
+**Q: Chain stops early**
+- Check that each agent's Publish Topic matches the next agent's Subscribe Topic

README.md

@@ -35,10 +35,10 @@ This system allows you to create and orchestrate multiple AI agents that communi
 ### 1. Agent Configuration
 Each agent has:
 - **Title**: A descriptive name
-- **Prompt**: Template for processing (can use `{schema}` and `{input}` placeholders)
+- **Prompt**: Template for processing (can use `{schema}`, `{input}`, and `{question}` placeholders)
 - **Model**: Which LLM to use
 - **Subscribe Topic**: Which topic triggers this agent
-- **Publish Topic**: Where to publish results
+- **Publish Topic**: Where to publish results (use "FINAL" for end results)
 
 ### 2. Message Bus
 - Agents subscribe to topics during initialization
@@ -48,9 +48,15 @@ Each agent has:
 
 ### 3. Execution Flow
 ```
-START → Agent1 (subscribes to START) → publishes to TOPIC_A
-     → Agent2 (subscribes to TOPIC_A) → publishes to TOPIC_B
-     → Agent3 (subscribes to TOPIC_B) → publishes to DONE
+User enters question → Published to START topic
+     ↓
+Agent1 (subscribes to START) → publishes to TOPIC_A
+     ↓
+Agent2 (subscribes to TOPIC_A) → publishes to TOPIC_B
+     ↓
+Agent3 (subscribes to TOPIC_B) → publishes to FINAL
+     ↓
+Final result displayed to user
 ```
 
 ## Example Use Case: NL to SQL Pipeline
@@ -86,14 +92,12 @@ START → Agent1 (subscribes to START) → publishes to TOPIC_A
 
 ## Supported Models
 
-This deployment uses **phi3** (3.8B parameters) - a general-purpose model that handles:
-- Text analysis and classification
-- SQL query generation
-- Summarization and extraction
-- Question answering
-- General reasoning tasks
+This deployment includes two models optimized for different tasks:
 
-The single-model approach ensures fast startup and reliable performance on free-tier resources.
+- **phi3**: General-purpose model (3.8B parameters) - Great for text analysis, SQL generation, summarization, reasoning
+- **cniongolo/biomistral**: Medical/scientific domain model - Specialized for healthcare and biomedical tasks
+
+These models provide a good balance between general and specialized capabilities.
 
 ## Architecture
 
@@ -122,4 +126,3 @@ Perfect for learning:
 - Pub/Sub messaging patterns
 - Asynchronous agent orchestration
 - Dynamic workflow composition
-

server.py

@@ -37,6 +37,7 @@ class Agent(BaseModel):
 
 class ExecutionRequest(BaseModel):
     schema: str
+    user_question: str = ""
     agents: List[Agent]
 
 # Pub/Sub Bus
@@ -78,12 +79,15 @@ def get_llm(model_name: str):
     return Ollama(model=model_name, temperature=0.1)
 
 # Execute agent
-async def execute_agent(agent: Agent, input_content: str, schema: str) -> str:
+async def execute_agent(agent: Agent, input_content: str, schema: str, user_question: str) -> str:
     """Execute a single agent with the given input"""
     llm = get_llm(agent.model)
     
     # Replace placeholders in prompt
-    prompt_text = agent.prompt.replace("{schema}", schema).replace("{input}", input_content)
+    prompt_text = (agent.prompt
+                   .replace("{schema}", schema)
+                   .replace("{input}", input_content)
+                   .replace("{question}", user_question))
     
     # Create and invoke prompt
     prompt = PromptTemplate(
@@ -112,8 +116,9 @@ async def execute_pipeline(request: ExecutionRequest) -> AsyncGenerator[str, Non
                 yield create_event("agent_subscribed", agent=agent.title, topic=agent.subscribe_topic)
         
         # Publish START message
-        bus.publish("START", "System initialized")
-        yield create_event("message_published", topic="START", content="System initialized")
+        start_message = request.user_question if request.user_question else "System initialized"
+        bus.publish("START", start_message)
+        yield create_event("message_published", topic="START", content=start_message)
         
         # Process messages in the bus
         processed_topics = set()
@@ -146,13 +151,17 @@ async def execute_pipeline(request: ExecutionRequest) -> AsyncGenerator[str, Non
                     
                     # Execute agent
                     try:
-                        result = await execute_agent(agent, message_content, request.schema)
+                        result = await execute_agent(agent, message_content, request.schema, request.user_question)
                         yield create_event("agent_output", content=result)
                         
                         # Publish result to agent's publish topic
                         if agent.publish_topic:
                             bus.publish(agent.publish_topic, result)
                             yield create_event("message_published", topic=agent.publish_topic, content=result)
+                            
+                            # If publishing to FINAL, send special event to frontend
+                            if agent.publish_topic == "FINAL":
+                                yield create_event("final_result", content=result)
                         
                         yield create_event("agent_completed", agent=agent.title)
                         

static/index.html

@@ -21,11 +21,14 @@
 - products (id, name, category, price)
 - orders (id, customer_id, order_date, total)
 - order_items (id, order_id, product_id, quantity, price)`);
+          const [userQuestion, setUserQuestion] = useState('');
+          const [finalResult, setFinalResult] = useState('');
           const [logs, setLogs] = useState('');
           const [isExecuting, setIsExecuting] = useState(false);
 
           const models = [
-            "phi3"
+            "phi3",
+            "cniongolo/biomistral"
           ];
 
           const addLog = (message, type = 'info') => {
@@ -59,9 +62,13 @@
           const executeSystem = async () => {
             setIsExecuting(true);
             setLogs('');
+            setFinalResult('');
             
             addLog('Initializing Pub/Sub Agent System...', 'info');
             addLog(`Total agents configured: ${agents.length}`, 'info');
+            if (userQuestion) {
+              addLog(`User Question: ${userQuestion}`, 'info');
+            }
             
             try {
               const response = await fetch('/execute', {
@@ -71,6 +78,7 @@
                 },
                 body: JSON.stringify({
                   schema,
+                  user_question: userQuestion,
                   agents: agents.map(a => ({
                     title: a.title,
                     prompt: a.prompt,
@@ -116,6 +124,9 @@
                         addLog(`Output: ${data.content}`, 'info');
                       } else if (data.type === 'agent_completed') {
                         addLog(`Agent "${data.agent}" completed`, 'success');
+                      } else if (data.type === 'final_result') {
+                        addLog(`\n=== FINAL RESULT ===`, 'success');
+                        setFinalResult(data.content);
                       } else if (data.type === 'no_subscribers') {
                         addLog(`No subscribers for topic "${data.topic}"`, 'error');
                       } else if (data.type === 'execution_complete') {
@@ -167,6 +178,19 @@
                       />
                     </div>
 
+                    {/* User Question */}
+                    <div className="bg-white rounded-lg shadow p-4">
+                      <label className="block text-sm font-semibold text-gray-700 mb-2">
+                        💬 User Question
+                      </label>
+                      <textarea
+                        value={userQuestion}
+                        onChange={(e) => setUserQuestion(e.target.value)}
+                        className="w-full h-24 p-3 border border-gray-300 rounded-lg text-sm focus:ring-2 focus:ring-purple-500 focus:border-transparent"
+                        placeholder="Enter your question (available as {question} in prompts)..."
+                      />
+                    </div>
+
                     {/* Agents */}
                     <div className="bg-white rounded-lg shadow p-4">
                       <div className="flex justify-between items-center mb-4">
@@ -212,7 +236,7 @@
                                   value={agent.prompt}
                                   onChange={(e) => updateAgent(agent.id, 'prompt', e.target.value)}
                                   className="w-full h-24 p-2 border border-gray-300 rounded text-sm font-mono focus:ring-2 focus:ring-purple-500"
-                                  placeholder="Enter prompt template (use {schema}, {input}, etc.)"
+                                  placeholder="Use {schema}, {input}, {question} as placeholders"
                                 />
                               </div>
 
@@ -283,17 +307,33 @@
                     </button>
                   </div>
 
-                  {/* Right Column - Execution Log */}
-                  <div className="bg-white rounded-lg shadow p-4">
-                    <label className="block text-sm font-semibold text-gray-700 mb-2">
-                      📋 Execution Log
-                    </label>
-                    <textarea
-                      value={logs}
-                      readOnly
-                      className="w-full h-[calc(100vh-200px)] p-3 border border-gray-300 rounded-lg font-mono text-xs bg-gray-50 focus:outline-none overflow-auto"
-                      placeholder="Execution logs will appear here when you run the pipeline..."
-                    />
+                  {/* Right Column - Logs and Results */}
+                  <div className="space-y-4">
+                    {/* Final Result */}
+                    <div className="bg-white rounded-lg shadow p-4">
+                      <label className="block text-sm font-semibold text-gray-700 mb-2">
+                        ✨ Final Result (from FINAL topic)
+                      </label>
+                      <textarea
+                        value={finalResult}
+                        readOnly
+                        className="w-full h-48 p-3 border border-gray-300 rounded-lg text-sm bg-green-50 focus:outline-none overflow-auto"
+                        placeholder="The result from agents publishing to 'FINAL' topic will appear here..."
+                      />
+                    </div>
+
+                    {/* Execution Log */}
+                    <div className="bg-white rounded-lg shadow p-4">
+                      <label className="block text-sm font-semibold text-gray-700 mb-2">
+                        📋 Execution Log
+                      </label>
+                      <textarea
+                        value={logs}
+                        readOnly
+                        className="w-full h-[calc(100vh-500px)] p-3 border border-gray-300 rounded-lg font-mono text-xs bg-gray-50 focus:outline-none overflow-auto"
+                        placeholder="Execution logs will appear here when you run the pipeline..."
+                      />
+                    </div>
                   </div>
                 </div>
               </div>