Examples

The examples are designed to be read top to bottom. Each notebook keeps input paths near the top, writes into a dedicated output folder, and pauses after important stages so you can inspect the data before continuing.

Notebook Guide

Notebook	Start here when	Main outputs
video_to_trc_quickstart.ipynb	You have a single-camera video and want CSV/TRC exports.	pose CSV files, global TRC
marker_trc_cleanup.ipynb	You already have a TRC file and want a cleaned version.	marker summary, cleaned TRC
opensim_scale_ik_template.ipynb	You have a TRC and want OpenSim scale and IK.	scale setup, scaled model, IK motion
video_to_inverse_dynamics_pipeline.ipynb	You want video to TRC, scale, IK, external loads, and ID.	pose outputs, OpenSim files, estimated loads, ID storage
run_video_smoke.py	You want a repeatable command-line check for a real video.	JSON report, CSV, TRC, optional OpenSim outputs

Suggested Order

Verify import

Run the first cell before pointing at large data files.

Set paths once

Edit the input path and output directory variables near the top.

Inspect before export

Use summaries and DataFrame previews before writing files.

Move downstream slowly

Only run OpenSim after marker names, units, and time ranges look sensible.

What A Complete Example Shows

The full video-to-ID notebook is organized around the same checkpoints used by the docs tour. You should be able to see the 2D pose on the source image, the root-centered 3D pose, the global pose after PnP and floor alignment, key IK angles, estimated force signals, key ID kinetics, and the synchronized animation viewer.

Full Video To ID Example

import monomech as mm

pose = mm.estimate_pose("data/subject01.mp4")
pose = mm.gap_fill(mm.smooth(pose))

scale = mm.run_scaling(
    pose,
    model="pose",
    output_dir="outputs/subject01/scale",
)

ik = mm.run_ik(scale, output_dir="outputs/subject01/ik")

estimated_loads = mm.estimate_grf(pose, body_mass_kg=75.0)

id_result = mm.run_id(
    ik=ik,
    external_forces=estimated_loads,
    output_dir="outputs/subject01/id",
)

print(id_result.path)
print(id_result.metadata["external_loads_mot_path"])

Command-Line Smoke Test

python examples/run_video_smoke.py "data/subject01.mp4" --output-dir outputs/subject01_smoke

Run OpenSim too when the bindings are installed:

python examples/run_video_smoke.py "data/subject01.mp4" --output-dir outputs/subject01_smoke --opensim

Minimal Video Example

from pathlib import Path
import monomech as mm

video_path = Path("data/subject01.mp4")
output_dir = Path("outputs/subject01")
output_dir.mkdir(parents=True, exist_ok=True)

pose3d_global = mm.estimate_pose(video_path)
pose3d_global = mm.gap_fill(mm.smooth(pose3d_global))

pose3d_global.to_csv(output_dir / "subject01_global.csv")
pose3d_global.to_trc(output_dir / "subject01_global.trc")

Minimal Marker Example

from pathlib import Path
import monomech as mm

trc_path = Path("data/walk.trc")
output_dir = Path("outputs/walk")
output_dir.mkdir(parents=True, exist_ok=True)

markers = mm.gap_fill(trc_path, max_gap_frames=20)
markers = mm.smooth(markers, cutoff_hz=6.0)

display(markers.summary())
markers.to_trc(output_dir / "walk_clean.trc")

Good Notebook Hygiene

• Keep raw input files in data/ and generated files in outputs/.
• Save intermediate CSV files while tuning pose, smoothing, or marker mapping.
• Record the exact install command and package version in the first notebook cell.
• Use one notebook per subject when parameter choices differ between trials.
• Keep OpenSim scale, IK, external-load, and ID sections separate so failures are easier to debug.