Streamlit

Learn how to build Streamlit dashboards in TileDB with Python.

Streamlit turns Python code into shareable web applications, without the need for front end knowledge. With some extra configuration, you can create Streamlit dashboards based on Jupyter notebooks in TileDB. The following steps will guide you through creating a basic Streamlit dashboard.

Launch a new notebook server

To start, launch a new notebook server. Select the image and server size based on your needs. After opening the server, open the launcher. Under the Other section, select Terminal.

Create the streamlit_jupyter plugin

Create a new file named streamlit_jupyter.py in your home directory:

touch ~/streamlit_jupyter.py
nano ~/streamlit_jupyter.py

Paste the following code into the file:

Python

streamlit_jupyter.py

import os
import socket
import subprocess
import tempfile
import threading
import time

from IPython.core.magic import Magics, cell_magic, magics_class


def find_free_port():
    """Find an available port for the Streamlit app."""
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
        s.bind(("", 0))
        s.listen(1)
        port = s.getsockname()[1]
    return port


def start_streamlit_app(script_path, port):
    """Start a Streamlit app in a separate thread."""

    def run_streamlit():
        subprocess.run(
            [
                os.path.expanduser("~/.local/bin/streamlit"),
                "run",
                script_path,
                "--server.port",
                str(port),
                "--browser.gatherUsageStats",
                "false",
                ">/dev/null",
                "2>&1",
            ]
        )

    thread = threading.Thread(target=run_streamlit, daemon=True)
    thread.start()
    time.sleep(3)  # Give the app a moment to start
    return thread


@magics_class
class StreamlitMagics(Magics):
    @cell_magic
    def streamlit(self, line, cell):
        """
        Jupyter cell magic to run Streamlit apps.

        Args:
            line (str): Magic command line arguments (unused)
            cell (str): Cell content containing Streamlit code

        Returns:
            IPython.display.IFrame: An iframe displaying the Streamlit app
        """
        from IPython.display import IFrame

        # Create a temporary file
        with tempfile.NamedTemporaryFile(
            mode="w", suffix=".py", delete=False
        ) as temp_file:
            # Ensure necessary imports are present
            imports = "import streamlit as st\n"

            # Write imports and cell content to the temp file
            temp_file.write(imports + cell)
            temp_file_path = temp_file.name

        # Find an available port
        port = find_free_port()

        # Start the Streamlit app
        start_streamlit_app(temp_file_path, port)

        # Create an iframe pointing to the Streamlit app
        # Create an iframe pointing to the Streamlit app
        if os.getenv("JUPYTERHUB_API_URL"):
            user = os.getenv("JUPYTERHUB_USER")
            url = f"/user/{user}/proxy/{port}/"
        else:
            url = f"http://localhost:{port}/"
        return IFrame(src=url, width="100%", height="600px")


def load_ipython_extension(ipython):
    """
    Register the Streamlit magic command.

    Args:
        ipython (InteractiveShell): IPython shell instance
    """
    ipython.register_magics(StreamlitMagics)


def unload_ipython_extension(ipython):
    """
    Unregister the Streamlit magic command.

    Args:
        ipython (InteractiveShell): IPython shell instance
    """
    # No direct way to unregister, but this prevents further use
    try:
        del ipython.magics_manager.magics["cell"]["streamlit"]
    except KeyError:
        pass

Now save the file:

1. Press Ctrl-OCtrl-O to write out the changes.
2. Press EnterEnter or ReturnReturn to use the existing filename.
3. Press Ctrl-XCtrl-X to exit the editor.

Create a new notebook

Start by creating a new notebook in TileDB. In the first cell, you’ll load the streamlit_jupyter plugin you created earlier.

Python

%load_ext streamlit_jupyter

Next, install Streamlit into your user directory:

Python

%pip install --quiet --user -U --no-warn-script-location streamlit

Finally, add the %%streamlit cell magic, and add the rest of your code to build your dashboard:

Python

%%streamlit
st.title("My Streamlit App")
x = st.slider("Select a number")
st.write(f"You selected: {x}")

Convert notebook to dashboard

Convert the notebook to a dashboard, and launch the dashboard in the existing server.

Your dashboard should look something like this: