---

name: visualization description: "Create and troubleshoot ProLIF visual outputs from completed fingerprints and molecules." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Visualization

Use this sub-skill when an agent needs plots or visual exports from an already executed ProLIF fingerprint: ligand interaction networks, barcode plots, 3D complex views, or residue grids. The plotting APIs consume Fingerprint.ifp, Fingerprint.to_dataframe(), and molecule objects; they do not run fingerprint analysis themselves.

Preconditions

• Start from a prolif.Fingerprint that has already run through fp.run(...), fp.run_from_iterable(...), or Fingerprint.from_pickle(...) with saved results.
• Keep the ligand/protein/water molecules that match the fingerprint frame and residue identifiers, especially for LigNetwork and Complex3D.
• Install the optional plotting backend needed by the output: matplotlib for barcode plots, py3Dmol for 3D views, and notebook/IPython support for inline display helpers.
• For fingerprint generation and molecule conversion, route to ../fingerprints/ first. For changing interaction definitions or water bridges, route to ../interactions/ first.

Fast Routing

Goal	Use	Notes
Save a standalone 2D network HTML	LigNetwork.from_fingerprint(...).save(path)	Preferred for scripts and non-notebook execution.
Display a 2D network inline	fp.plot_lignetwork(...) or LigNetwork.display()	Requires display-capable IPython context.
Plot per-frame interaction barcode	fp.plot_barcode(...) or Barcode.from_fingerprint(fp).display(...)	Returns a matplotlib Axes; set a non-interactive backend in batch jobs.
Show a 3D complex	fp.plot_3d(lig_mol, prot_mol, frame=...)	Requires py3Dmol and molecule coordinates for the chosen frame.
Inspect residue drawings	prolif.display_residues(mol, slice(...))	Useful for checking residue IDs, sanitization, and display readiness.

Workflow

1. Validate that the fingerprint has results: hasattr(fp, "ifp"), non-empty fp.ifp, and the requested frame exists.
2. Choose the output path: use LigNetwork.save() for durable HTML, matplotlib savefig() for barcode images, and notebook display only when the user explicitly wants inline rendering.
3. Pass matching molecule objects: lig_mol for networks; lig_mol, prot_mol, and optional water_mol for 3D views.
4. Tune display-only parameters without recomputing interactions: kind, threshold, display_all, use_coordinates, flatten_coordinates, sanitize, and remove_hydrogens.
5. If output is empty or raises RunRequiredError, verify fingerprint execution and molecule inputs before changing plotting options.

References

• API details: references/api-reference.md
• Plotting workflows: references/workflows.md
• Failure recovery: references/troubleshooting.md
• Batch smoke helper: scripts/plotting_smoke.py