Pathml

PathML

Scope and safety boundary

Use PathML for local computational pathology research. It is beta research software, not a validated medical device, diagnostic system, clinical decision support tool, or substitute for a pathologist. Do not use outputs to diagnose, grade, stage, or treat a patient.

Pathology files may contain faces, labels, accession numbers, patient identifiers, DICOM tags, filenames, or linked clinical data. Before processing:

1. Confirm authorization, consent/waiver, data-use terms, and institutional policy.
2. De-identify pixels and metadata; keep the re-identification key outside the analysis workspace.
3. Use pseudonymous patient_id, slide_id, and specimen_id values. Do not put direct identifiers in filenames, logs, .h5path labels, model cards, or reports.
4. Keep inputs, intermediates, and outputs on approved local encrypted storage.
5. Split by patient (then slide) before tiling or fitting any preprocessing step.

Version baseline, reviewed 2026-10-01

• Published stable release: PyPI pathml==3.0.8, published 2026-08-14, source tag v3.0.8 (fa49ffb66757ed8a8265756c92b6290834eddecb).
• The v3.0.5 release notes state Python 3.10-3.12 and sunset 3.9. PyPI does not declare Requires-Python and still has a stale 3.8 classifier, so use the release statement and test the exact environment.
• The current wheel incorporates the Torch/PyG dependency updates and explicit legacy ONNX export (dynamo=False). ReadTheDocs pages/search caches can show different versions; the 3.0.8 wheel and tagged source determine these APIs.
• Bundled CLIs are tested with synthetic data. PathML examples and installation commands are illustrative, source-checked, not a full native-stack or pathology-model execution claim. No slides, model weights, or datasets were downloaded for this review.
• This skill is MIT-licensed. PathML itself is GPL-2.0 with upstream commercial licensing options; review upstream terms before redistribution.

Reproducible installation

Use Python 3.11 unless the project has tested another supported interpreter:

uv venv --python 3.11
source .venv/bin/activate
uv pip install "pathml==3.0.8"
python -c "import importlib.metadata as m; print(m.version('pathml'))"

PathML 3.0.8 declares no package extras: do not use pathml[all]. Its base distribution pins a large scientific/ML stack, including Torch 2.12.0, ONNX 1.22.0, ONNX Runtime 1.17.x, OpenSlide Python 1.3.1, python-bioformats 4.1.0, and python-javabridge 4.0.4, torch-geometric 2.8.0, and onnxscript 0.7.1. It also pins NumPy below 2 and several older binary packages; do not upgrade them independently.

Install native prerequisites before the uv command:

# Debian/Ubuntu
sudo apt-get install openslide-tools gcc g++ libblas-dev liblapack-dev openjdk-17-jdk

# macOS
brew install openslide openjdk@17

# Windows OpenSlide option documented upstream
vcpkg install openslide

python-javabridge==4.0.4 is source-only on PyPI and needs a JDK (including javac/JNI headers), not just a JRE. Set JAVA_HOME to that JDK before building. PathML imports the Java bridge/Bio-Formats in its shared backend module even when the requested reader is OpenSlide; selecting OpenSlide does not avoid those installation dependencies. Old import-error advice mentioning Java 8 and javabridge==1.0.19 does not match the released wheel requirements.

Java/Bio-Formats provides the broad multidimensional format backend. OpenSlide handles common brightfield WSI formats more efficiently. CUDA is optional and must match the pinned PyTorch build; follow PyTorch's platform selector rather than guessing a CUDA wheel. See references/image_loading.md.

Stable minimal workflow

PathML 3.0.8 uses slide convenience classes and SlideData.run(). It does not provide SlideData.from_slide(), and Pipeline does not have run():

from pathml.core import HESlide
from pathml.preprocessing import BoxBlur, Pipeline, TissueDetectionHE

slide = HESlide("data/pseudonymous_slide.svs", backend="openslide")
pipeline = Pipeline(
    [
        BoxBlur(kernel_size=5),
        TissueDetectionHE(mask_name="tissue", min_region_size=5000),
    ]
)
slide.run(
    pipeline,
    distributed=False,
    tile_size=512,
    tile_stride=512,
    level=0,
    tile_pad=False,
)
slide.write("derived/pseudonymous_slide.h5path")

Start with a bounded manual sample before a full run:

from itertools import islice

for tile in islice(slide.generate_tiles(shape=512, stride=512, level=0), 8):
    pipeline.apply(tile)
    assert tile.masks["tissue"].shape[:2] == tile.image.shape[:2]

Tiles use (i, j) = (row, column) coordinates at the selected pyramid level. PathML 3.0.8 truncates OpenSlide's downsample to an integer when locating a region. Use level 0 or an exactly integral downsample; a fractional downsample can shift the sampled region. Record level, actual dimensions and MPP before converting to (x, y) or micrometres. See references/image_loading.md.

Before comparing tile features, cell distances, or areas across scanners, validate level-0 MPP separately for X and Y. Equal pixel tile sizes need not cover equal physical areas, and OpenSlide MPP may be absent or inaccurate. Keep results in pixel units when calibration is unknown, or document a validated calibration; do not infer it from objective magnification alone. See OpenSlide properties.

Research workflow

1. Inventory locally. Validate the manifest, reject URLs/symlinks, inspect only allowlisted technical metadata, and remove identifiers.
2. Freeze splits. Assign every patient and all their slides to one split before generating overlapping tiles, graphs, normalization references, or features.
3. Plan bounds. Estimate tile count, RAM, output size, and pipeline stages.
4. Pilot preprocessing. Inspect tissue masks, whitespace/artifact labels, stain behavior, edge padding, and empty-mask cases on representative training slides. Do not tune from test slides.
5. Run and preserve coordinates. Keep tile level, (i, j), downsample, MPP, mask names, QC decisions, and failed/skipped tiles.
6. Build spatial data deliberately. Validate channel order, physical units, instance labels, node-feature alignment, graph edges, and cell-to-tissue assignments.
7. Infer in bounded batches. Verify model provenance and checksum without loading unknown pickle checkpoints. Keep predictions linked to slide/tile coordinates and stitch overlaps with a documented rule.
8. Report provenance and limits. Include package lock, source hashes, scanner, stain, parameters, seeds, split manifest, model card, exclusions, and QC.

For multiplex measurement, pass normalize=False to Bio-Formats tile extraction to preserve intensities. .h5path still casts tile images and masks to float16: store quantitative images and integer instance maps separately when exact values matter (integers above 2048 are not all representable). See the data-management and multiplex references before quantification or storage.