Claude/add tutorials docs aa mpp#47
Merged
aylward merged 6 commits intoMay 8, 2026
Merged
Conversation
- tutorials/: six end-to-end tutorial scripts (one per workflow class) covering HeartGatedCTToUSD, CTToVTK, FitStatisticalModelToPatient, CreateStatisticalModel, VTKToUSD, and ReconstructHighres4DCT. Each script includes a run_tutorial() function (used by tests), a standalone argparse CLI, and structured docstrings documenting Inputs/Outputs/Strengths/Weaknesses/Classes/CLI. - tutorials/README.md: tutorial index, dataset requirements, run order, and experiment-test instructions. - tutorials/__init__.py: makes tutorials/ a package so tests can import run_tutorial() with a clean `from tutorials.tutorial_XX import …`. - test_tools.py: add save_screenshot_mesh() (PyVista off-screen PNG) and save_screenshot_image_slice() (matplotlib axial/coronal/sagittal PNG) to TestTools; both save under results_dir/class_name/. - tests/test_tutorials.py: six @pytest.mark.experiment test classes wired to the corresponding run_tutorial() functions; compare screenshots against baselines via the existing ITK comparison infrastructure. - pyproject.toml: add pythonpath = ["."] so pytest discovers tutorials/. - README.md: add "Getting Started: Tutorials" section before Quick Start. - docs/quickstart.rst: add Tutorials section with table and run examples. https://claude.ai/code/session_01PmHfz2ntnAAFkyrGwSx7Ui
- Correct tutorial documentation so Slicer-Heart-CT is described as data to prepare first, not data auto-downloaded by the tutorial scripts. - Fix Tutorial 1 to return the actual USD path under the output directory. - Tighten tutorial typing for PyVista-loaded meshes and screenshot helpers. - Make screenshot regression tolerance use the intended pixel-count criterion. - Regenerate docs/API_MAP.md after adding public TestTools screenshot helpers. - Author default USD mesh points, extent, and normals when writing time samples so single-frame USD meshes work with readers that query attributes without an explicit time code.
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
✅ Files skipped from review due to trivial changes (1)
WalkthroughThis PR standardizes CUDA support to 13.0 across packaging, CI, and docs; adds six runnable tutorial scripts with ChangesCUDA 13.0 Migration
Tutorial System, Tests & TestTools
Sequence Diagram(s)sequenceDiagram
participant User
participant TutorialScript as Tutorial<br/>(e.g. tutorial_01)
participant Workflow as Workflow<br/>(e.g. HeartGatedCT)
participant TestTools
participant FileSystem as Output<br/>Files
User->>TutorialScript: run_tutorial(data_dir, output_dir)
TutorialScript->>Workflow: instantiate + process()
Workflow->>FileSystem: write USD / meshes / volumes
TutorialScript->>TestTools: save_screenshot_mesh/image_slice()
TestTools->>FileSystem: write PNG screenshots
TutorialScript->>User: return dict with artifact paths
sequenceDiagram
participant PytestRunner as Pytest<br/>Runner
participant TestClass as TestTutorialNN<br/>(Experiment)
participant TutorialScript as Tutorial<br/>Script
participant TestTools
participant Baseline as Baseline<br/>Images
PytestRunner->>TestClass: test_run()
TestClass->>TutorialScript: run_tutorial()
TutorialScript->>TestTools: save_screenshot_mesh/image_slice()
TestTools->>FileSystem: PNG screenshots
TestClass->>TestTools: compare_result_to_baseline_image()
TestTools->>Baseline: load baseline PNG
TestTools->>TestTools: pixel comparison with tolerances
TestTools->>TestClass: pass/fail result
TestClass->>PytestRunner: assert pass
Estimated code review effort🎯 4 (Complex) | ⏱️ ~75 minutes Poem
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (2 warnings)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Signed-off-by: Stephen R. Aylward <stephen@aylward.org>
There was a problem hiding this comment.
Pull request overview
Adds a tutorials suite (scripts + docs) and an experiment-style pytest harness to run them end-to-end with screenshot baselines, alongside a small USD time-sample robustness improvement.
Changes:
- Added 6 runnable tutorial scripts under
tutorials/plus a tutorials index README. - Added
tests/test_tutorials.pyexperiment tests and extendedTestToolswith screenshot helpers for meshes and ITK image slices. - Improved USD time-varying mesh authoring by also writing default (non-time-sampled) points/extents/normals.
Reviewed changes
Copilot reviewed 14 out of 15 changed files in this pull request and generated 10 comments.
Show a summary per file
| File | Description |
|---|---|
tutorials/tutorial_01_heart_gated_ct_to_usd.py |
New tutorial script for heart-gated CT → animated USD with screenshots. |
tutorials/tutorial_02_ct_to_vtk.py |
New tutorial script for CT segmentation → VTK surfaces/meshes with screenshots. |
tutorials/tutorial_03_fit_statistical_model_to_patient.py |
New tutorial script for fitting a statistical model to patient (demo) meshes + screenshots. |
tutorials/tutorial_04_create_statistical_model.py |
New tutorial script for building a PCA shape model + screenshots. |
tutorials/tutorial_05_vtk_to_usd.py |
New tutorial script for VTK → USD conversion + screenshot. |
tutorials/tutorial_06_reconstruct_highres_4d_ct.py |
New tutorial script for high-res 4D CT reconstruction + screenshots. |
tutorials/README.md |
Tutorial index, run instructions, and experiment-test instructions. |
tests/test_tutorials.py |
New experiment tests that run tutorials and compare screenshots to baselines. |
src/physiomotion4d/test_tools.py |
Added screenshot helpers for PyVista meshes and ITK image slices. |
src/physiomotion4d/vtk_to_usd/usd_mesh_converter.py |
Author default values for time-varying points/extents/normals for better reader compatibility. |
README.md |
Added “Getting Started: Tutorials” section. |
docs/quickstart.rst |
Added tutorials section and updated quickstart text. |
docs/API_MAP.md |
Updated API map entries for new tutorial/test helpers. |
pyproject.toml |
Added pytest pythonpath to allow importing tutorials package in tests. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+151
to
+154
| # Reference frame: the workflow caches 3D frames in output_dir | ||
| ref_frames = sorted(output_dir.glob("slice_???.mha")) | ||
| if ref_frames: | ||
| ref_image = itk.imread(str(ref_frames[0])) |
Comment on lines
+167
to
+180
| # Segmentation overlay: look for cached labelmap | ||
| label_files = sorted(output_dir.glob("slice_???_labelmap*.mha")) | ||
| overlay = itk.imread(str(label_files[0])) if label_files else None | ||
| screenshots.append( | ||
| tt.save_screenshot_image_slice( | ||
| ref_image, | ||
| "segmentation_overlay.png", | ||
| axis=0, | ||
| slice_fraction=0.5, | ||
| colormap="gray", | ||
| vmin=-200, | ||
| vmax=600, | ||
| overlay_mask=overlay, | ||
| ) |
Comment on lines
+25
to
+29
| - ``output_dir/<phase>_*.vtp`` - per-frame surface meshes (VTK PolyData) | ||
| - Screenshots (PNG) for documentation and regression testing: | ||
| - ``reference_frame_axial.png`` - axial slice of the reference CT frame | ||
| - ``segmentation_overlay.png`` - segmentation mask overlaid on reference | ||
| - ``contours_3d.png`` - 3-D isometric view of the reference-frame contours |
Comment on lines
+123
to
+126
| template_model = pv.read(str(template_file)) | ||
| if not isinstance(template_model, pv.PolyData): | ||
| template_model = template_model.extract_surface() | ||
| template_model = cast(pv.PolyData, template_model) |
Comment on lines
+139
to
+142
| sample_model = pv.read(str(sample_file)) | ||
| if not isinstance(sample_model, pv.PolyData): | ||
| sample_model = sample_model.extract_surface() | ||
| patient_models.append(cast(pv.PolyData, sample_model)) |
Comment on lines
+185
to
+187
| before_path = tt._results_dir / "model_before_registration.png" | ||
| before_path.parent.mkdir(parents=True, exist_ok=True) | ||
| plotter.screenshot(str(before_path)) |
Comment on lines
+125
to
+133
| # Discover phase images (MetaImage .mhd or .mha) | ||
| phase_files = sorted(case_dir.glob("*.mhd")) + sorted(case_dir.glob("*.mha")) | ||
| if not phase_files: | ||
| raise FileNotFoundError( | ||
| f"No .mhd / .mha files found under {case_dir}.\n" | ||
| "See data/README.md for manual download instructions." | ||
| ) | ||
|
|
||
| phase_files = phase_files[:max_frames] |
Comment on lines
+426
to
+432
| import matplotlib.pyplot as plt | ||
| import numpy as np | ||
|
|
||
| arr = np.asarray(itk.array_view_from_image(image), dtype=np.float64) | ||
| idx = int(arr.shape[axis] * slice_fraction) | ||
| idx = max(0, min(idx, arr.shape[axis] - 1)) | ||
|
|
| plotter.camera_position = "iso" | ||
|
|
||
| png_name = f"pca_mode_{mode_idx + 1:02d}.png" | ||
| png_path = tt._results_dir / png_name |
Comment on lines
+133
to
+136
| phase_files = phase_files[:max_frames] | ||
| time_series = [itk.imread(str(f)) for f in phase_files] | ||
| fixed_image = time_series[0] # use first phase as high-res reference | ||
|
|
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #47 +/- ##
==========================================
+ Coverage 21.72% 22.09% +0.36%
==========================================
Files 46 46
Lines 6268 6272 +4
==========================================
+ Hits 1362 1386 +24
+ Misses 4906 4886 -20
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
- Make the Sphinx index page the NVIDIA-styled tutorial landing page - Add tutorial cards, documentation topic cards, and NVIDIA logo styling - Configure uv-managed PyTorch packages to resolve from the cu130 index by default - Remove cuda12/cu128 dependency configuration and update install docs accordingly - Align GitHub workflow docs with CUDA 13.0 setup and cuda13 test installs - Verify default uv torch resolves as cu130 and Sphinx docs build succeeds
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 12
🧹 Nitpick comments (3)
src/physiomotion4d/test_tools.py (1)
444-445: 💤 Low valueRedundant
import numpy as npinside method.
numpyis already imported at module level (Line 17). The local import is a no-op but adds noise.♻️ Proposed fix
- import matplotlib.pyplot as plt - import numpy as np + import matplotlib.pyplot as plt🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/physiomotion4d/test_tools.py` around lines 444 - 445, Remove the redundant local import "import numpy as np" inside the method where "import matplotlib.pyplot as plt" is present; numpy is already imported at module scope, so delete the local "import numpy as np" statement (leave the plt import if it's only used locally) to avoid no-op duplicate imports and reduce noise.docs/index.rst (1)
105-248: ⚡ Quick winDuplicate tutorial content across
index.rstandtutorials.rstcreates a maintenance burdenLines 105–248 inline the complete tutorial documentation (Recommended Run Order + all six Tutorial sections + Dataset Notes) that is also fully present in
docs/tutorials.rst. Any future change to commands, dataset paths, or outputs must be applied in both files, and they will inevitably drift.Consider replacing the inlined content with a Sphinx
.. include::directive or replacing the in-page cardhrefanchors with cross-document links (:doc:tutorials`` + section anchors), keeping the full content only intutorials.rst.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/index.rst` around lines 105 - 248, Index.rst duplicates the entire tutorial content (the "Recommended Run Order" and "Tutorial 1" through "Tutorial 6" sections and "Dataset Notes") that already lives in tutorials.rst, causing maintenance drift; remove the inlined blocks in index.rst (the sections headed "Recommended Run Order", "Tutorial 1: Heart-Gated CT to Animated USD" … "Tutorial 6: Reconstruct High-Resolution 4D CT", and "Dataset Notes") and replace them with a single Sphinx-level reference to the canonical tutorials.rst content — either use a Sphinx include of the tutorials document or replace the inline cards/anchors with cross-document links to the tutorials document and its section anchors (so updates live only in tutorials.rst).tutorials/tutorial_03_fit_statistical_model_to_patient.py (1)
169-174: 💤 Low valueRedundant
_extract_surfacecall on an already-converted mesh.
template_modelwas already extracted to apv.PolyDataon Line 130, so the_extract_surfacewrapping on Line 169 is a no-op (the function returns apv.PolyDataunchanged).♻️ Proposed fix
- plotter.add_mesh( - _extract_surface(template_model), - color="dodgerblue", - opacity=0.6, - label="Template", - ) + plotter.add_mesh( + template_model, + color="dodgerblue", + opacity=0.6, + label="Template", + )🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@tutorials/tutorial_03_fit_statistical_model_to_patient.py` around lines 169 - 174, The call to _extract_surface is redundant because template_model is already a pv.PolyData; update the plotter.add_mesh invocation to pass template_model directly instead of wrapping it in _extract_surface. Locate the plotter.add_mesh call that currently uses _extract_surface(template_model) and replace the argument with template_model (keeping color, opacity, and label unchanged) so you avoid the no-op wrapper around the already-converted mesh.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/workflows/README.md:
- Around line 110-114: Add a note to the "Install CUDA toolkit 13.0" section
warning that CUDA 13.0 requires Ubuntu 22.04+ so installs of the
wget/dpkg/cuda-keyring and apt-get install cuda-toolkit-13-0 commands will fail
on Ubuntu 20.04; update the README near the "Install CUDA toolkit 13.0" heading
and mention self-hosted runners should be upgraded to Ubuntu 22.04 LTS or later
(reference the "cuda-keyring_1.1-1_all.deb" and "cuda-toolkit-13-0" package
names in the note).
- Around line 110-114: The apt install command uses the wrong package name;
change the APT package in the workflow from "cuda-toolkit-13-0" to the correct
meta-package "cuda-toolkit-13" so the CUDA toolkit installs successfully (update
the sudo apt-get install line that currently references cuda-toolkit-13-0).
In `@docs/_static/custom.css`:
- Around line 303-310: Add a distinct keyboard focus indicator for the card
components so focus is visually different from hover: update the .pm4d-card and
.pm4d-topic-card selectors to keep the existing hover styles for :hover but add
a separate :focus-visible rule that adds a clear, high-contrast outline or an
adjusted box-shadow (and does not rely solely on the hover transform), so
keyboard users can distinguish focus from hover and meet WCAG 2.4.11; target
.pm4d-card:focus-visible and .pm4d-topic-card:focus-visible (instead of :focus)
and ensure the indicator uses accessible color/contrast and does not get hidden
by outline: none elsewhere.
In `@src/physiomotion4d/test_tools.py`:
- Around line 392-404: The off-screen pv.Plotter created in this function can
leak resources if screenshot or other calls raise; wrap the Plotter usage in a
try/finally so plotter.close() is always called (e.g., create plotter =
pv.Plotter(...); try: add_mesh/ set camera/ screenshot; finally:
plotter.close()). Also ensure that if you successfully started pv.start_xvfb(),
you call pv.stop_xvfb() in a finally block (or after plotter.close()) to clean
up Xvfb; reference the pv.start_xvfb(), pv.stop_xvfb() calls and the plotter
variable/plotter.close() to locate where to add the try/finally.
- Around line 455-473: The plotting code creates fig and ax then calls ax.imshow
and fig.savefig but never guarantees plt.close(fig) on exceptions; wrap the
plotting and save calls in a try/finally (or use try/except/finally) around the
block that uses fig/ax (the ax.imshow(...) and fig.savefig(...) calls) and move
plt.close(fig) into the finally so the figure is always closed even if ax.imshow
or fig.savefig raises; reference the local symbols fig, ax, ax.imshow,
fig.savefig, and plt.close to locate and update the code.
In `@src/physiomotion4d/workflow_convert_heart_gated_ct_to_usd.py`:
- Around line 181-183: The code currently indexes self._time_series_images using
reference_frame computed from self._num_time_points without guarding against an
empty series; add a guard in the conversion routine (around the reference_frame
/ self._fixed_image assignment) that checks self._num_time_points > 0 and
self._time_series_images is non-empty, and if not either raise a clear
ValueError/RuntimeError mentioning no time-series images were produced or set
self._fixed_image to None and handle that downstream; update any callers of the
conversion method to expect and handle the new error/None return accordingly so
IndexError is avoided.
In `@tests/test_tutorials.py`:
- Around line 59-72: The test helper _compare_screenshots currently silently
succeeds when the screenshots list is empty; add an explicit guard at the start
of _compare_screenshots to fail the test when screenshots is empty (e.g., if not
screenshots: pytest.fail("No screenshots produced by run_tutorial")) so missing
outputs cause a test failure; reference the screenshots parameter and keep the
existing per-image checks/loop intact so existing comparisons still run when
screenshots are present.
In `@tutorials/README.md`:
- Around line 54-57: The "Tutorials 3 and 4" line is misleading because Tutorial
3 (WorkflowFitStatisticalModelToPatient) requires the model produced by Tutorial
4 (WorkflowCreateStatisticalModel); update the README to explicitly state that
Tutorial 4 must be run before Tutorial 3 (either swap their mention or change
the text to "Run Tutorial 4 (WorkflowCreateStatisticalModel) before Tutorial 3
(WorkflowFitStatisticalModelToPatient)"), so readers know the required execution
order.
In `@tutorials/tutorial_04_create_statistical_model.py`:
- Around line 193-199: pv.start_xvfb() is being called inside the PCA loop (the
for mode_idx in range(min(2, pca_components)) block) which can start multiple
Xvfb instances; move the pv.start_xvfb() call (with its existing try/except
guard) to run once before the loop (for example immediately before entering the
TestTools block or before the for loop) so Xvfb is started only once; update the
code around pv.start_xvfb() and the loop to ensure the virtual framebuffer is
started once and still protected by the same exception handling.
- Around line 129-137: The fallback glob currently includes the reference mesh
pca_mean.vtu (via kcl_dir.glob("*.vtu")), biasing the PCA; update the sample
selection after both globs so sample_files = [f for f in sample_files if f.name
!= "pca_mean.vtu"] (or filter out kcl_dir / "pca_mean.vtu") to exclude the
reference mesh from training, and ensure the length check (len(sample_files) <
3) counts only non-reference files and raises the FileNotFoundError if there are
fewer than 3 valid samples; adjust the error message to mention the directory
actually searched (sample_dir/kcl_dir) if desired.
In `@tutorials/tutorial_05_vtk_to_usd.py`:
- Around line 103-116: The candidate lookup for the Tutorial 2 VTK file is using
Path("output") which is CWD-relative and ignores the configured data_dir; update
the auto-discovery so candidate is resolved relative to the project/data root
(use data_dir as the base) instead of the current working directory — e.g.,
compute the project root from data_dir (or use an existing project_root
variable) and build candidate from that root (referencing the variables
vtk_file, candidate and data_dir in the existing block) so the script reliably
finds output/tutorial_02/patient_surfaces.vtp regardless of CWD or configured
output locations.
In `@tutorials/tutorial_06_reconstruct_highres_4d_ct.py`:
- Around line 129-135: The current concatenation of sorted lists (phase_files =
sorted(case_dir.glob("*.mhd")) + sorted(case_dir.glob("*.mha")) and similarly
for dirlab_dir) forces all .mhd before .mha and breaks temporal order; instead,
collect both extensions into one list then sort the combined list so filenames
are ordered temporally (e.g., merge results from case_dir.glob("*.mhd") and
case_dir.glob("*.mha") into a single list and call sorted(...) once, or sort by
extracted frame index from filenames), and apply the same fix for the fallback
that uses phase_pattern and dirlab_dir so phase_files holds a correctly ordered
sequence.
---
Nitpick comments:
In `@docs/index.rst`:
- Around line 105-248: Index.rst duplicates the entire tutorial content (the
"Recommended Run Order" and "Tutorial 1" through "Tutorial 6" sections and
"Dataset Notes") that already lives in tutorials.rst, causing maintenance drift;
remove the inlined blocks in index.rst (the sections headed "Recommended Run
Order", "Tutorial 1: Heart-Gated CT to Animated USD" … "Tutorial 6: Reconstruct
High-Resolution 4D CT", and "Dataset Notes") and replace them with a single
Sphinx-level reference to the canonical tutorials.rst content — either use a
Sphinx include of the tutorials document or replace the inline cards/anchors
with cross-document links to the tutorials document and its section anchors (so
updates live only in tutorials.rst).
In `@src/physiomotion4d/test_tools.py`:
- Around line 444-445: Remove the redundant local import "import numpy as np"
inside the method where "import matplotlib.pyplot as plt" is present; numpy is
already imported at module scope, so delete the local "import numpy as np"
statement (leave the plt import if it's only used locally) to avoid no-op
duplicate imports and reduce noise.
In `@tutorials/tutorial_03_fit_statistical_model_to_patient.py`:
- Around line 169-174: The call to _extract_surface is redundant because
template_model is already a pv.PolyData; update the plotter.add_mesh invocation
to pass template_model directly instead of wrapping it in _extract_surface.
Locate the plotter.add_mesh call that currently uses
_extract_surface(template_model) and replace the argument with template_model
(keeping color, opacity, and label unchanged) so you avoid the no-op wrapper
around the already-converted mesh.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 2b59bb8e-1847-443d-89ed-67d3aa768356
⛔ Files ignored due to path filters (1)
docs/_static/nvidia-logo.svgis excluded by!**/*.svg
📒 Files selected for processing (24)
.github/workflows/README.mdREADME.mddocs/API_MAP.mddocs/_static/custom.cssdocs/conf.pydocs/faq.rstdocs/index.rstdocs/installation.rstdocs/quickstart.rstdocs/troubleshooting.rstdocs/tutorials.rstpyproject.tomlsrc/physiomotion4d/test_tools.pysrc/physiomotion4d/vtk_to_usd/usd_mesh_converter.pysrc/physiomotion4d/workflow_convert_heart_gated_ct_to_usd.pytests/test_tutorials.pytutorials/README.mdtutorials/__init__.pytutorials/tutorial_01_heart_gated_ct_to_usd.pytutorials/tutorial_02_ct_to_vtk.pytutorials/tutorial_03_fit_statistical_model_to_patient.pytutorials/tutorial_04_create_statistical_model.pytutorials/tutorial_05_vtk_to_usd.pytutorials/tutorial_06_reconstruct_highres_4d_ct.py
Comment on lines
+54
to
+57
| 1. **Tutorial 1** and **Tutorial 2** use Slicer-Heart-CT - prepare it per `data/README.md`, then start here. | ||
| 2. **Tutorial 5** uses the VTK surfaces produced by Tutorial 2 - run Tutorial 2 first. | ||
| 3. **Tutorials 3 and 4** require the KCL-Heart-Model - download it per `data/README.md`. | ||
| 4. **Tutorial 6** requires DirLab-4DCT - download it per `data/README.md`. |
There was a problem hiding this comment.
Tutorial 3 depends on Tutorial 4 — recommended order is misleading.
Tutorial 3 (WorkflowFitStatisticalModelToPatient) requires a statistical model as input, which is produced by Tutorial 4 (WorkflowCreateStatisticalModel). The recommended order section groups them as "Tutorials 3 and 4" without clarifying that Tutorial 4 must run first.
📝 Suggested fix
-3. **Tutorials 3 and 4** require the KCL-Heart-Model - download it per `data/README.md`.
+3. **Tutorial 4** (`WorkflowCreateStatisticalModel`) requires the KCL-Heart-Model - download it per `data/README.md`, then run Tutorial 4 before Tutorial 3.
+4. **Tutorial 3** (`WorkflowFitStatisticalModelToPatient`) requires the output of Tutorial 4.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@tutorials/README.md` around lines 54 - 57, The "Tutorials 3 and 4" line is
misleading because Tutorial 3 (WorkflowFitStatisticalModelToPatient) requires
the model produced by Tutorial 4 (WorkflowCreateStatisticalModel); update the
README to explicitly state that Tutorial 4 must be run before Tutorial 3 (either
swap their mention or change the text to "Run Tutorial 4
(WorkflowCreateStatisticalModel) before Tutorial 3
(WorkflowFitStatisticalModelToPatient)"), so readers know the required execution
order.
- Make the docs index the NVIDIA-styled tutorial landing page - Add tutorial cards, doc topic cards, NVIDIA logo, and accessible card focus states - Move detailed tutorial content back to tutorials.rst and link index cards there - Configure uv-managed PyTorch packages to use the cu130 index by default - Remove cuda12/cu128 install paths and update README/docs/workflow guidance - Align GitHub workflow runner docs with CUDA 13.0 and cuda-toolkit-13 - Harden tutorial screenshot handling, sample discovery, and empty output checks - Guard empty heart-gated CT time-series conversion before reference-frame indexing - Validate with Ruff, fast pytest, and Sphinx HTML build
|
|
||
| python tutorials/tutorial_05_vtk_to_usd.py \ | ||
| --data-dir ./data --output-dir ./output/tutorial_05 \ | ||
| --input output/tutorial_02/patient_surfaces.vtp |
Comment on lines
34
to
39
| class TestTools(PhysioMotion4DBase): | ||
| """ | ||
| Utilities for pytest image comparison: baseline directory, results directory, | ||
| Utilities for pytest image comparison: baseline directory, result directory, | ||
| and comparison with configurable tolerances. Inherits from PhysioMotion4DBase | ||
| for logging. All image I/O uses ITK .mha with compression. | ||
| for logging. All image I/O uses ITK with compression where supported. | ||
| """ |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit
New Features
Documentation
Tests
Chores