Visualization Functions

Functions:

crop_image(image, centroid, crop_size)	Crop an image around a centroid.
write_video_clip(frames, path[, fps, quality])	Write a video clip to a file using imageio.
grid_movie(instances, rows, cols, videos, ...)	Generate a grid movie from video instances.
generate_grid_movies(states_dict, ...[, ...])	Generate and save grid movies for high-level states.
plot_sankey(states_dict1, states_dict2)	Create a Sankey diagram representing the relationship between two sets of states.

state_moseq.viz.crop_image(image, centroid, crop_size)

Crop an image around a centroid.

Parameters:

• image (Array) – Image to crop, shape (height, width, channels).

• centroid (Tuple[int, int]) – (x, y) coordinates of the centroid.

• crop_size (int | Tuple[int, int]) – Size of crop around centroid (int or (width, height)).

Returns:

Cropped (and zero-padded if needed) image.

Return type:

cropped_image

state_moseq.viz.write_video_clip(frames, path, fps=30, quality=7)

Write a video clip to a file using imageio.

Parameters:

• frames (Array) – Video frames of shape (n_frames, height, width, channels).

• path (str) – Output file path.

• fps (int) – Frames per second.

• quality (int) – Video encoding quality.

Return type:

None

state_moseq.viz.grid_movie(instances, rows, cols, videos, centroids, window_size, max_duration=1200, dimming_factor=0.5)

Generate a grid movie from video instances.

Parameters:

• instances (List[Tuple[str, int, int]]) – List of (key, start, end) tuples.

• rows (int) – Number of rows in grid.

• cols (int) – Number of columns in grid.

• videos (Dict) – Dictionary of video readers keyed by name.

• centroids (Dict[str, Float[Array, 'n_timesteps 2']]) – Dictionary of centroids for each frame.

• window_size (int) – Size of cropped window around centroid.

• max_duration (int) – Maximum duration of the grid movie.

• dimming_factor (float) – Multiplicative factor for dimming after state instance has ended.

Returns:

Combined movie frames as a single array.

Return type:

frames

state_moseq.viz.generate_grid_movies(states_dict, video_paths, centroids, output_dir='grid_movies', rows=3, cols=4, max_duration=1200, dimming_factor=0.25, quality=7, window_size=256)

Generate and save grid movies for high-level states.

Parameters:

• states_dict (Dict[str, Int[Array, 'n_timesteps']]) – Dictionary of state sequences.

• video_paths (Dict[str, str]) – Dictionary of paths to video files.

• centroids (Dict[str, Float[Array, 'n_timesteps 2']]) – Dictionary of centroids for each frame.

• output_dir (str) – Directory to save output movies.

• rows (int) – Number of rows in each grid.

• cols (int) – Number of columns in each grid.

• max_duration (int) – Maximum duration of each movie.

• dimming_factor (float) – Factor to dim frames after state instance has ended.

• quality (int) – Video encoding quality.

• window_size (int) – Size of window around centroid.

Returns:

Sampled instances used for grid movies.

Return type:

instances

state_moseq.viz.plot_sankey(states_dict1, states_dict2)

Create a Sankey diagram representing the relationship between two sets of states.

Parameters:

• states_dict1 (Dict[str, Int[Array, 'n_timesteps']]) – First dictionary of state sequences.

• states_dict2 (Dict[str, Int[Array, 'n_timesteps']]) – Second dictionary of state sequences.

Returns:

Plotly figure object containing the Sankey diagram.

Return type:

fig