stspeck

The Speck molecule viewer for Streamlit, with the same renderer and settings as the ipyspeck Jupyter widget: ambient occlusion, cartoons, molecular surfaces (optionally transparent), metallic and glossy materials, shadows, fog, outlines, highlighting, trajectories and unit cells.

pip install stspeck      # Python 3.8+, Streamlit 1.20+

import streamlit as st
import stspeck

# Any structure as PDB or (extended) XYZ text
stspeck.speck(data=open("1ubq.pdb").read(), cartoon=True, surface=True, surfaceOpacity=0.35)

# Loaders return keyword arguments to splat into speck()
stspeck.speck(**stspeck.fetch_alphafold("Q8W3K0"), preset="cover", height=500)
stspeck.speck(**stspeck.fetch_pdb("4HHB"), highlight={"resName": "HEM"}, ghost=0.6)
stspeck.speck(**stspeck.from_ase(atoms), preset="metal")   # also from_rdkit, from_pymatgen

• Settings: every ipyspeck setting is a keyword argument; see stspeck.SETTINGS for names and defaults, and stspeck.PRESETS for named looks (preset="cover").
• Toolbar: styles, cartoon / surface / ligand toggles, standard views, color schemes and a camera button that downloads a high-resolution PNG (export_width=3000, export_scale=...). Toolbar choices persist across reruns.
• State: with return_state=True the call returns the camera, the current settings and nframes; pass the camera back as camera= to restore a view.
• Trajectories: multi-frame XYZ or multi-model PDB; drive frame= with a slider (stspeck.count_frames(data) gives the range).

Run the demo with streamlit run example/app.py.

Gallery

For Jupyter, use ipyspeck, which has the same settings.

Development

The frontend (frontend/) bundles the viewer shared with ipyspeck (../core). Build it with cd frontend && npm install && npm run build, which also compiles ../core. pip install . runs the same build when stspeck/static is missing.