Skip to main content

​
POST /api/chat/with-context

Streams responses from the AI assistant with the complete medical knowledge base as context, including all scraped medical data.

​
Request Body

​
model
string
required
The Ollama model to use (e.g., “llama2”, “mistral”, “medllama2”)
​
question
string
required
The user’s question to the AI
​
instruction
string
Custom system instruction. Defaults to: “Responda com base EXCLUSIVA no contexto. Se algo não estiver no contexto, diga que não há informação suficiente.”
​
maxItems
integer
default:"500"
Maximum number of scraped items to include in context
​
maxChars
integer
default:"12000"
Maximum total characters of context (prevents token overflow)

​
Response

The response is streamed using Server-Sent Events (SSE) format with Content-Type: application/octet-stream. Each line contains a JSON object with the AI’s response chunk.

​
Example Request

cURL
curl -X POST http://localhost:8080/api/chat/with-context \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama2",
    "question": "Quais são os sintomas da dengue segundo as últimas notícias?",
    "maxItems": 300,
    "maxChars": 10000
  }' \
  --no-buffer
JavaScript
const response = await fetch('http://localhost:8080/api/chat/with-context', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    model: 'llama2',
    question: 'Quais são os sintomas da dengue segundo as últimas notícias?',
    maxItems: 300,
    maxChars: 10000
  })
});

const reader = response.body.getReader();
const decoder = new TextDecoder();

while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  
  const chunk = decoder.decode(value);
  const lines = chunk.split('\n').filter(line => line.trim());
  
  for (const line of lines) {
    const data = JSON.parse(line);
    process.stdout.write(data.response || '');
  }
}

​
Example Response Stream

{"model":"llama2","created_at":"2024-03-15T10:30:00Z","response":"Segundo","done":false}
{"model":"llama2","created_at":"2024-03-15T10:30:00Z","response":" as informações","done":false}
{"model":"llama2","created_at":"2024-03-15T10:30:00Z","response":" do contexto fornecido","done":false}
...
{"model":"llama2","created_at":"2024-03-15T10:30:01Z","response":".","done":true}

​
How It Works

From OllamaChatController.java:108-176:
  1. Context Building: Retrieves ALL scraped medical data 
    String contexto = contextService.buildAllContext(maxItems, maxChars);
  2. Prompt Construction: Combines instruction + context + question 
    String prompt = instruction + "\n\n[CONTEXTO]\n" + contexto + 
                "\n\n[PERGUNTA]\n" + question + "\n\nResposta objetiva:";
  3. Ollama Streaming: Sends to Ollama and streams back the response

​
Context Sources

The context includes data from all scraped sources:
  • CID Codes: Disease classifications and descriptions
  • Health News: Latest medical news from government sources
  • Medical Skills: CFM-defined medical specialties and abilities
  • Vigilance Guides: Public health surveillance definitions

​
Differences from /api/chat

Feature/api/chat/api/chat/with-context
ContextAutomatic (keyword-based) or optionalAlways includes ALL context
Context SizeTargeted/filteredComplete knowledge base
Use CaseGeneral questionsQuestions needing comprehensive medical data
PerformanceFaster (less context)Slower (more context)

​
Use Cases

  • Complex medical queries requiring multiple data sources
  • Questions about current health news and guidelines
  • Comprehensive disease information lookup
  • Correlating symptoms with recent health alerts
  • Medical research and information gathering
Prerequisites:
  • Ollama must be running locally at http://localhost:11434
  • The specified model must support large context windows
  • Consider using models optimized for medical domains (e.g., “medllama2”)
For simpler questions or when you want targeted context, use POST /api/chat which automatically detects relevant context.

​
Performance Considerations

  • Context Size: The maxChars parameter prevents token overflow. Adjust based on your Ollama model’s context window
  • Response Time: Larger context = longer initial processing time
  • Memory: Streaming response uses minimal memory on the server
  • Ollama Model: Choose a model with appropriate context window size (e.g., 8K, 32K, 128K tokens)

​
Error Responses

Status CodeDescription
200Streaming response in progress
400Invalid request (missing model or question)
500Error connecting to Ollama, context building failed, or streaming error
This endpoint is ideal for the medical chatbot feature where comprehensive, accurate medical information is critical.

Build docs developers (and LLMs) love

Get started for freeTalk to us