from chatjimmy import ChatJimmy

client = ChatJimmy()

answer = client.ask("What is the capital of France?")
print(answer)

from chatjimmy import ChatJimmy

client = ChatJimmy()
print(client.ask("Explain quantum computing in one sentence."))

from chatjimmy import ChatJimmy

client = ChatJimmy()

response = client.chat(
    messages=[{"role": "user", "content": "Explain recursion"}],
    system_prompt="You are a computer science tutor.",
    top_k=4,
)

print(response.text)
print(f"Output tokens: {response.stats.decode_tokens}")
print(f"Speed: {response.stats.decode_rate:.0f} tokens/sec")

from chatjimmy import ChatJimmy

client = ChatJimmy()

messages = [
    {"role": "user", "content": "My name is Mohamed."},
]
resp = client.chat(messages)
print(resp.text)

messages.append({"role": "assistant", "content": resp.text})
messages.append({"role": "user", "content": "What's my name?"})
resp = client.chat(messages)
print(resp.text)

from chatjimmy import ChatJimmy

client = ChatJimmy()

for chunk in client.chat_stream(
    messages=[{"role": "user", "content": "Write a haiku about code"}]
):
    print(chunk, end="", flush=True)
print()

from chatjimmy import ChatJimmy

client = ChatJimmy()
health = client.health()

print(health.healthy)       # True/False
print(health.backend)       # "healthy"
print(health.timestamp)     # ISO timestamp

from chatjimmy import ChatJimmy

client = ChatJimmy()

for model in client.models():
    print(f"{model.id} (by {model.owned_by})")

from chatjimmy import ChatJimmy, Message

client = ChatJimmy()

response = client.chat(
    messages=[Message(role="user", content="Hello!")],
    system_prompt="Reply in French.",
)
print(response.text)

from chatjimmy import ChatJimmy, Attachment

client = ChatJimmy()

attachment = Attachment(name="data.txt", size=11, content="hello world")
response = client.chat(
    messages=[{"role": "user", "content": "Summarize this file"}],
    attachment=attachment,
)
print(response.text)

response = client.chat(messages=[{"role": "user", "content": "hi"}])
stats = response.stats

stats.prefill_tokens    # input tokens processed
stats.prefill_rate      # input processing speed (tokens/sec)
stats.decode_tokens     # output tokens generated
stats.decode_rate       # output generation speed (tokens/sec)
stats.total_tokens      # prefill + decode
stats.ttft              # time to first token (seconds)
stats.total_time        # total inference time (seconds)
stats.roundtrip_time    # network round trip (ms)
stats.done_reason       # "stop" (natural end)