from dotenv import load_dotenv

load_dotenv()

True

import os
from langchain.chat_models import init_chat_model
from langchain.agents import create_agent
from langchain.messages import HumanMessage, AIMessage
from langchain.tools import tool
from typing import Dict, Any
from tavily import TavilyClient
from pprint import pprint


DEEPSEEK_MODEL = "deepseek-chat"
DEEPSEEK_BASE_URL = "https://api.deepseek.com"

def deepseek_model(model: str = DEEPSEEK_MODEL, max_tokens=1000, **kwargs):
    return init_chat_model(
        model=model,
        # DeepSeek uses LangChain's OpenAI-compatible transport.
        model_provider="openai",
        api_key=os.environ["DEEPSEEK_API_KEY"],
        base_url=DEEPSEEK_BASE_URL,
        max_tokens=max_tokens,
        **kwargs,
    )

agent = create_agent(
    model=deepseek_model(),
    system_prompt="You are a science fiction writer, create a capital city at the users request.",
)

question = HumanMessage(content=[
    {"type": "text", "text": "What is the capital of The Moon?"}
])

response = agent.invoke(
    {"messages": [question]}
)

print(response['messages'][-1].content)

from ipywidgets import FileUpload
from IPython.display import display

uploader = FileUpload(accept='.png', multiple=False)
display(uploader)

FileUpload(value=(), accept='.png', description='Upload')

print(uploader.value)

({'name': 'i1.png', 'type': 'image/png', 'size': 5769586, 'content': <memory at 0x1184bd840>, 'last_modified': datetime.datetime(2026, 4, 22, 6, 29, 19, 414000, tzinfo=datetime.timezone.utc)},)

print(uploader.value[0])

{'name': 'i1.png', 'type': 'image/png', 'size': 5769586, 'content': <memory at 0x1184bd840>, 'last_modified': datetime.datetime(2026, 4, 22, 6, 29, 19, 414000, tzinfo=datetime.timezone.utc)}

import base64

# Get the first (and only) uploaded file dict
uploaded_file = uploader.value[0]

# This is a memoryview
content_mv = uploaded_file["content"]

# Convert memoryview -> bytes
img_bytes = bytes(content_mv)  # or content_mv.tobytes()

# Now base64 encode
img_b64 = base64.b64encode(img_bytes).decode("utf-8")


question = HumanMessage(content=[
    {
        "type": "text",
        "text": "What is in this image?"
    },
    {
        "type": "image_url",
        "image_url": {"url": f"data:image/png;base64,{img_b64}"}
    }
])

print(
    "DeepSeek's current public chat API is text/tool-call oriented and does not expose "
    "an image input model in the official API docs. Keep this upload/base64 cell for "
    "practice, but use a multimodal provider if you need image understanding."
)

import sounddevice as sd
from scipy.io.wavfile import write
import base64
import io
import time
from tqdm import tqdm

# Recording settings
duration = 5  # seconds
sample_rate = 44100

print("Recording...")
audio = sd.rec(int(duration * sample_rate), samplerate=sample_rate, channels=1)
# Progress bar for the duration
for _ in tqdm(range(duration * 10)):   # update 10× per second
    time.sleep(0.1)
sd.wait()
print("Done.")

# Write WAV to an in-memory buffer
buf = io.BytesIO()
write(buf, sample_rate, audio)
wav_bytes = buf.getvalue()

aud_b64 = base64.b64encode(wav_bytes).decode("utf-8")

print(
    "DeepSeek's current public chat API is text/tool-call oriented and does not expose "
    "an audio input model in the official API docs. Keep the recording cell for practice, "
    "but use a multimodal audio provider if you need audio understanding."
)

Text input¶

Image input¶

Audio input¶