================================================================================
                        STREAMING CHAT DEMO CAPTURE
                     Gunicorn Dirty Workers + FastAPI SSE
================================================================================

$ curl -s http://127.0.0.1:8000/health
{"status":"ok"}

================================================================================
                           TEST 1: Hello Prompt
================================================================================

$ curl -N http://127.0.0.1:8000/chat -d '{"prompt": "hello"}'

data: {"token": "Hello! "}

data: {"token": "I'm "}

data: {"token": "a "}

data: {"token": "simulated "}

data: {"token": "AI "}

data: {"token": "assistant "}

data: {"token": "running "}

data: {"token": "on "}

data: {"token": "Gunicorn's "}

data: {"token": "dirty "}

data: {"token": "workers. "}

data: {"token": "I "}

data: {"token": "can "}

data: {"token": "demonstrate "}

data: {"token": "streaming "}

data: {"token": "responses "}

data: {"token": "just "}

data: {"token": "like "}

data: {"token": "a "}

data: {"token": "real "}

data: {"token": "LLM, "}

data: {"token": "but "}

data: {"token": "without "}

data: {"token": "the "}

data: {"token": "heavy "}

data: {"token": "ML "}

data: {"token": "dependencies. "}

data: {"token": "How "}

data: {"token": "can "}

data: {"token": "I "}

data: {"token": "help "}

data: {"token": "you "}

data: {"token": "today?"}

data: [DONE]

================================================================================
                        TEST 2: Explain Dirty Workers
================================================================================

$ curl -N http://127.0.0.1:8000/chat -d '{"prompt": "explain dirty workers"}'

data: {"token": "Dirty "}

data: {"token": "workers "}

data: {"token": "are "}

data: {"token": "separate "}

data: {"token": "processes "}

data: {"token": "that "}

data: {"token": "handle "}

data: {"token": "long-running "}

data: {"token": "tasks "}

data: {"token": "like "}

data: {"token": "ML "}

data: {"token": "inference. "}

data: {"token": "They "}

data: {"token": "keep "}

data: {"token": "models "}

data: {"token": "loaded "}

data: {"token": "in "}

data: {"token": "memory "}

data: {"token": "across "}

data: {"token": "requests, "}

data: {"token": "avoiding "}

data: {"token": "expensive "}

data: {"token": "reload "}

data: {"token": "times. "}

data: {"token": "HTTP "}

data: {"token": "workers "}

data: {"token": "remain "}

data: {"token": "lightweight "}

data: {"token": "and "}

data: {"token": "responsive "}

data: {"token": "while "}

data: {"token": "dirty "}

data: {"token": "workers "}

data: {"token": "handle "}

data: {"token": "the "}

data: {"token": "heavy "}

data: {"token": "computation. "}

data: {"token": "This "}

data: {"token": "architecture "}

data: {"token": "is "}

data: {"token": "inspired "}

data: {"token": "by "}

data: {"token": "Erlang's "}

data: {"token": "dirty "}

data: {"token": "schedulers."}

data: [DONE]

================================================================================
                         TEST 3: Sync Endpoint
================================================================================

$ curl -s http://127.0.0.1:8000/chat/sync -d '{"prompt": "hello"}'

{"response":"Hello! I'm a simulated AI assistant running on Gunicorn's dirty workers. I can demonstrate streaming responses just like a real LLM, but without the heavy ML dependencies. How can I help you today?"}

================================================================================
                         DEMO COMPLETE
================================================================================

Browser UI available at: http://localhost:8000/

Features demonstrated:
  - Token-by-token SSE streaming
  - Async generators via dirty workers
  - Different responses based on keywords
  - Sync endpoint for comparison
  - Health check endpoint

Server Logs:
[INFO] Starting gunicorn 24.1.0
[INFO] Listening at: http://0.0.0.0:8000 (1)
[INFO] Using worker: asgi
[INFO] Spawned dirty arbiter (pid: 7)
[INFO] Dirty arbiter starting (pid: 7)
[INFO] Booting worker with pid: 8
[INFO] Dirty arbiter listening on /tmp/gunicorn-dirty-.../arbiter.sock
[INFO] Spawned dirty worker (pid: 9)
[INFO] Initialized dirty app: streaming_chat.chat_app:ChatApp
[INFO] Dirty worker 9 listening on /tmp/gunicorn-dirty-.../worker-1.sock
[INFO] ASGI server listening on http://0.0.0.0:8000