copick convert picks2mesh

utils

Convert picks to meshes using convex hull or alpha shapes.

Plugin command — copick-utils

This command is provided by the copick-utils plugin, not copick core. Install it to make this command available:

pip install copick-utils

See the plugin system guide for details.

copick convert picks2mesh output — Output

Convert picks to meshes using convex hull or alpha shapes.

Usage

copick convert picks2mesh [OPTIONS]

Description

For each matched pick set, builds a single surface mesh that encloses the pick coordinates. Two methods are supported: convex_hull wraps the points in their tightest convex envelope, while alpha_shape (requires --alpha) yields a tighter, possibly non-convex surface. Optional clustering (DBSCAN or KMeans) first splits the picks into groups, and --individual-meshes writes one mesh per cluster/instance via the {instance_id} placeholder.

Input and output URIs select picks and meshes by object name, user id and session id. They support exact, glob, regex (re: prefix) and bare-wildcard matching, enabling one-to-one, one-to-many and many-to-many conversions through {input_session_id} and {instance_id} placeholders in the output URI.

URI Format

Picks: object_name:user_id/session_id
Meshes: object_name:user_id/session_id

Options

Option	Type	Default	Description
`-c, --config`	path	—	Path to the configuration file.
`--debug / --no-debug`	boolean flag	`False`	Enable debug logging.

Input Options

Option	Type	Default	Description
`--run-names, -r`	text · multiple	—	Specific run names to process (default: all runs).
`--input, -i`	COPICK_URI	required	Input picks URI (format: object_name:user_id/session_id). Supports glob patterns.

Tool Options

Option	Type	Default	Description
`--mesh-type, -t`	choice (convex_hull \| alpha_shape)	`convex_hull`	Type of mesh to create.
`--alpha, -a`	float	—	Alpha parameter for alpha shapes (required if mesh-type=alpha_shape).
`--use-clustering, -cl / --no-use-clustering`	boolean flag	`False`	Cluster points before mesh creation.
`--clustering-method`	choice (dbscan \| kmeans)	`dbscan`	Clustering method.
`--clustering-eps`	float	`1.0`	DBSCAN eps parameter - maximum distance between points in a cluster (in angstroms).
`--clustering-min-samples`	integer	`3`	DBSCAN min_samples parameter.
`--clustering-n-clusters`	integer	`1`	K-means n_clusters parameter.
`--all-clusters, -mm / --largest-cluster-only`	boolean flag	`True`	Use all clusters (True) or only the largest cluster (False).
`--workers, -w`	integer	`8`	Number of worker processes.

Output Options

Option	Type	Default	Description
`--output, -o`	COPICK_URI	required	Output mesh URI. Supports smart defaults (e.g., "membrane", "membrane/my-session", or "/my-session"). Full format: object_name:user_id/session_id.
`--individual-meshes, -im / --no-individual-meshes`	boolean flag	`False`	Create individual meshes for each instance (enables {instance_id} placeholder).

Examples

# Convert a single pick set to a single convex-hull mesh
copick convert picks2mesh -i "ribosome:user1/manual-001" -o "ribosome:picks2mesh/mesh-001"

# Create individual meshes from clusters
copick convert picks2mesh -i "ribosome:user1/manual-001" \
    -o "ribosome:picks2mesh/mesh-{instance_id}" --individual-meshes

# Build alpha-shape surfaces for all manual picks via regex matching
copick convert picks2mesh --mesh-type alpha_shape --alpha 150 \
    -i "re:ribosome:user\d+/manual-.*" -o "ribosome:picks2mesh/mesh-{input_session_id}"