bitjet

Binary data visualization as a Jupyter anywidget. Drop bytes in, see structure come out.

from bitjet import BitView, ByteView

BitView(data=open("/bin/ls", "rb").read(), datawidth=64)

What it does

Every bit (or byte) in your payload becomes a fixed-size block on a canvas. The datawidth slider changes how many units are drawn per row — not how big the canvas gets — so periodic structures in the data (byte boundaries in ASCII text, alignment padding, null runs in executables, LZ-compressed block headers) line up into columns and patterns pop out as you drag.

Install

pip install bitjet

That's it. anywidget handles the JS — no lab extension, no server extension, no rebuild step.

Use

from bitjet import BitView, ByteView

# Bit mode: each bit is a white (1) or black (0) block. Good for ASCII text,
# checksums, bitfields, anywhere the individual bits matter.
BitView(data=b"The quick brown fox jumps over the lazy dog. " * 128, datawidth=64)

# Byte mode: each byte is a grayscale block (0=black, 255=white). Good for
# pixel data, pickled objects, compressed blobs, anything where byte-level
# entropy matters.
ByteView(data=bytes(range(256)) * 16, datawidth=32, block_size=6)

Assign to data to live-update:

import os

view = BitView(data=b"")
view  # renders

# In a later cell
view.data = os.urandom(8192)

Traitlets

All sync bidirectionally. Drag the slider and the Python side sees the update; assign view.datawidth = 128 and the canvas redraws.

Compatible with

• JupyterLab 3+
• Jupyter Notebook 7+
• VS Code notebooks
• nteract desktop (tested during development)
• Anything else that runs anywidget

History

Replaces the original 2014-era bitjet that depended on ipywidgets 1.x-style nbextensions. Same idea, modern plumbing — anywidget loads the ESM directly instead of requiring a notebook extension install.

License

BSD-3-Clause. See LICENSE.