Fix/issue 131

.github/workflows/ci.yml

@@ -22,7 +22,7 @@ jobs:
  run: |
  python -m pip install --upgrade pip setuptools wheel
  python -m pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu
- python -m pip install --editable .[dev] --find-links https://girder.github.io/large_image_wheels
+ python -m pip install --editable .[dev]
  - name: Run tests
  run: python -m pytest --verbose tests/
  test-pytorch-nightly:
@@ -37,7 +37,7 @@ jobs:
  run: |
  python -m pip install --upgrade pip setuptools wheel
  python -m pip install --pre torch torchvision --extra-index-url https://download.pytorch.org/whl/nightly/cpu
- python -m pip install --editable .[dev] --find-links https://girder.github.io/large_image_wheels
+ python -m pip install --editable .[dev]
  - name: Check types
  run: python -m mypy --install-types --non-interactive wsinfer/
  - name: Run tests
@@ -77,7 +77,7 @@ jobs:
  run: |
  python -m pip install --upgrade pip setuptools wheel
  python -m pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu
- python -m pip install . --find-links https://girder.github.io/large_image_wheels
+ python -m pip install .
  - name: Run the wsinfer command in a new directory
  run: |
  mkdir newdir && cd newdir
@@ -101,7 +101,7 @@ jobs:
  run: |
  python -m pip install --upgrade pip setuptools wheel
  python -m pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu
- python -m pip install .[dev] --find-links https://girder.github.io/large_image_wheels
+ python -m pip install .[dev]
  - name: Check style (flake8)
  run: python -m flake8 wsinfer/
  - name: Check style (black)
@@ -120,7 +120,7 @@ jobs:
  run: |
  python -m pip install --upgrade pip setuptools wheel
  python -m pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu
- python -m pip install .[docs] --find-links https://girder.github.io/large_image_wheels
+ python -m pip install .[docs]
  - name: Build docs
  run: |
  cd docs

.readthedocs.yaml

@@ -10,7 +10,7 @@ build:
  jobs:
  post_create_environment:
  - python -m pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu
- - python -m pip install .[docs] --find-links https://girder.github.io/large_image_wheels
+ - python -m pip install .[docs]
  post_install:
  # Re-run the installation to ensure we have an appropriate version of sphinx.
  # We might not want to use the latest version.

Dockerfile

@@ -4,7 +4,6 @@ COPY . .
 RUN apt-get update \
  && apt-get install -y --no-install-recommends gcc g++ git libopenslide0 \
  && python -m pip install --no-cache-dir --editable . \
- --find-links https://girder.github.io/large_image_wheels \
  && rm -rf /var/lib/apt/lists/*
 # Use a writable directory for downloading model weights. Default is ~/.cache, which is
 # not guaranteed to be writable in a Docker container.

README.md

@@ -23,15 +23,13 @@ We do not install these dependencies automatically because their installation ca
 on a user's system. Then use the command below to install this package.
 
 ```
-python -m pip install --find-links https://girder.github.io/large_image_wheels wsinfer
+python -m pip install wsinfer
 ```
 
 To use the _bleeding edge_, use
 
 ```
-python -m pip install \
- --find-links https://girder.github.io/large_image_wheels \
- git+https:/SBU-BMI/wsinfer.git
+python -m pip install git+https:/SBU-BMI/wsinfer.git
 ```
 
 ## Developers
@@ -41,7 +39,7 @@ Clone this GitHub repository and install the package (in editable mode with the
 ```
 git clone https:/SBU-BMI/wsinfer.git
 cd wsinfer
-python -m pip install --editable .[dev] --find-links https://girder.github.io/large_image_wheels
+python -m pip install --editable .[dev]
 ```
 
 # Cutting a release

docs/installing.rst

@@ -21,11 +21,9 @@ the type of hardware a user has.
 Manual installation
 -------------------
 
-After having installed PyTorch, install releases of WSInfer from `PyPI <https://pypi.org/project/wsinfer/>`_.
-Be sure to include the line :code:`--find-links https://girder.github.io/large_image_wheels` to ensure
-dependencies are installed properly. ::
+After having installed PyTorch, install releases of WSInfer from `PyPI <https://pypi.org/project/wsinfer/>`_. ::
 
- pip install wsinfer --find-links https://girder.github.io/large_image_wheels
+ pip install wsinfer
 
 This installs the :code:`wsinfer` Python package and the :code:`wsinfer` command line program. ::
 

setup.cfg

@@ -34,7 +34,6 @@ install_requires =
  click>=8.0,<9
  h5py
  # OpenSlide and TIFF readers should handle all images we will encounter.
- large-image[openslide,tiff]>=1.8.0
  numpy
  opencv-python-headless>=4.0.0
  pandas
@@ -83,8 +82,6 @@ exclude = wsinfer/_version.py
 [mypy]
 [mypy-h5py]
 ignore_missing_imports = True
-[mypy-large_image]
-ignore_missing_imports = True
 [mypy-cv2]
 ignore_missing_imports = True
 [mypy-torchvision.*]
@@ -105,3 +102,8 @@ versionfile_source = wsinfer/_version.py
 versionfile_build = wsinfer/_version.py
 tag_prefix = v
 parentdir_prefix = wsinfer
+
+
+[isort]
+profile = black
+force_single_line = True

wsinfer/_version.py

@@ -15,7 +15,8 @@
 import re
 import subprocess
 import sys
-from typing import Callable, Dict
+from typing import Callable
+from typing import Dict
 
 
 def get_keywords():

wsinfer/cli/convert_csv_to_sbubmi.py

@@ -2,17 +2,17 @@
 
 Output directory tree for single class outputs:
 ├── heatmap_jsons
-│  ├── heatmap-SLIDEID.json
-│  └── meta-SLIDEID.json
+│  ├── heatmap-SLIDEID.json
+│  └── meta-SLIDEID.json
 └── heatmap_txt
  ├── color-SLIDEID
  └── prediction-SLIDEID
 
 Output directory tree for multi-class outputs:
 ├── heatmap_jsons
-│  └── CLASS_LABEL
-│  ├── heatmap-SLIDEID.json
-│  └── meta-SLIDEID.json
+│  └── CLASS_LABEL
+│  ├── heatmap-SLIDEID.json
+│  └── meta-SLIDEID.json
 └── heatmap_txt
  └── CLASS_LABEL
  ├── color-SLIDEID
@@ -29,8 +29,8 @@
 from pathlib import Path
 
 import click
-import large_image
 import numpy as np
+import openslide
 import pandas as pd
 import tqdm
 
@@ -184,7 +184,7 @@ def write_heatmap_txt(input: PathType, output: PathType, class_names: typing.Lis
 def write_color_txt(
  input: PathType,
  output: PathType,
- ts: large_image.tilesource.TileSource,
+ oslide: openslide.OpenSlide,
  num_processes: int = 6,
 ):
  def whiteness(arr):
@@ -204,15 +204,12 @@ def redness(arr):
  global get_color # Hack to please multiprocessing.
 
  def get_color(row: pd.Series):
- arr, _ = ts.getRegion(
- format=large_image.constants.TILE_FORMAT_NUMPY,
- region=dict(
- left=row["minx"],
- top=row["miny"],
- width=row["width"],
- height=row["height"],
- ),
+ patch_im = oslide.read_region(
+ location=(row["minx"], row["miny"]),
+ level=0,
+ size=(row["width"], row["height"]),
  )
+ arr = np.asarray(patch_im)
  white = whiteness(arr)
  black = blackness(arr)
  red = redness(arr)
@@ -356,11 +353,9 @@ def tosbu(
  click.secho(f"WSI file not found: {wsi_file}", bg="red")
  click.secho("Skipping...", bg="red")
  continue
- ts = large_image.getTileSource(wsi_file)
- if ts.sizeX is None or ts.sizeY is None:
- click.secho(f"Unknown size for WSI: {wsi_file}", bg="red")
- click.secho("Skipping...", bg="red")
- continue
+ oslide = openslide.OpenSlide(wsi_file)
+
+ slide_width, slide_height = oslide.level_dimensions[0]
 
  for class_name in class_names:
  if len(class_names) == 1:
@@ -382,8 +377,8 @@ def tosbu(
  input=input_csv,
  output_heatmap=output_heatmap,
  output_meta=output_meta,
- slide_width=ts.sizeX,
- slide_height=ts.sizeY,
+ slide_width=slide_width,
+ slide_height=slide_height,
  execution_id=execution_id,
  study_id=study_id,
  case_id=slide_id, # TODO: should case_id be different?
@@ -418,7 +413,7 @@ def tosbu(
  write_color_txt(
  input=input_csv,
  output=output_color,
- ts=ts,
+ oslide=oslide,
  num_processes=num_processes,
  )
  else:
@@ -429,7 +424,7 @@ def tosbu(
  write_color_txt(
  input=input_csv,
  output=output_color,
- ts=ts,
+ oslide=oslide,
  num_processes=num_processes,
  )
  # Copy this color file to all class-specific dirs.

wsinfer/cli/infer.py

@@ -10,13 +10,15 @@
 import sys
 from datetime import datetime
 from pathlib import Path
-from typing import Optional, Union
+from typing import Optional
+from typing import Union
 
 import click
 import wsinfer_zoo
 import wsinfer_zoo.client
 import yaml
-from wsinfer_zoo.client import HFModel, ModelConfiguration
+from wsinfer_zoo.client import HFModel
+from wsinfer_zoo.client import ModelConfiguration
 
 from ..modellib import models
 from ..modellib.run_inference import run_inference

wsinfer/errors.py

@@ -9,17 +9,21 @@ class UnknownArchitectureError(WsinferException):
  """Architecture is unknown and cannot be found."""
 
 
-class WholeSlideImageDirectoryNotFound(FileNotFoundError):
+class WholeSlideImageDirectoryNotFound(WsinferException, FileNotFoundError):
  ...
 
 
-class WholeSlideImagesNotFound(FileNotFoundError):
+class WholeSlideImagesNotFound(WsinferException, FileNotFoundError):
  ...
 
 
-class ResultsDirectoryNotFound(FileNotFoundError):
+class ResultsDirectoryNotFound(WsinferException, FileNotFoundError):
  ...
 
 
-class PatchDirectoryNotFound(FileNotFoundError):
+class PatchDirectoryNotFound(WsinferException, FileNotFoundError):
+ ...
+
+
+class CannotReadSpacing(WsinferException):
  ...

wsinfer/modellib/custom_models/inceptionv4_no_batchnorm.py

@@ -12,7 +12,8 @@
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-from timm.data import IMAGENET_INCEPTION_MEAN, IMAGENET_INCEPTION_STD
+from timm.data import IMAGENET_INCEPTION_MEAN
+from timm.data import IMAGENET_INCEPTION_STD
 from timm.models import register_model
 from timm.models.helpers import build_model_with_cfg
 from timm.models.layers import create_classifier

wsinfer/modellib/custom_models/resnet_preact.py

@@ -1,6 +1,6 @@
 """Pre-activation ResNet."""
 
-import torch
+import torch # noqa
 import torch.nn as nn
 import torch.nn.functional as F
 

wsinfer/modellib/data.py

@@ -1,15 +1,17 @@
 from pathlib import Path
-from typing import Callable, Optional, Sequence, Tuple, Union
+from typing import Callable
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Union
 
 import h5py
-import large_image
 import numpy as np
+import openslide
 import torch
 from PIL import Image
 
-# Set the maximum number of TileSource objects to cache. We use 1 to minimize how many
-# file handles we keep open.
-large_image.config.setConfig("cache_tilesource_maximum", 1)
+from ..slide_utils import get_avg_mpp
 
 PathType = Union[str, Path]
 
@@ -83,18 +85,13 @@ def __init__(
  assert Path(wsi_path).exists(), "wsi path not found"
  assert Path(patch_path).exists(), "patch path not found"
 
- self.tilesource: large_image.tilesource.TileSource = large_image.getTileSource(
- self.wsi_path
- )
- # Disable the tile cache. We wrap this in a try-except because we are accessing
- # a private attribute. It is possible that this attribute will change names
- # in the future, and if that happens, we do not want to raise errors.
- try:
- self.tilesource.cache._Cache__maxsize = 0
- except AttributeError:
- pass
-
+ self.oslide = openslide.OpenSlide(self.wsi_path)
+ self.slide_mpp = get_avg_mpp(self.wsi_path)
  self.patches = _read_patch_coords(self.patch_path)
+
+ # The factor by which to resize the patches.
+ self.size_factor = self.slide_mpp / self.um_px
+
  assert self.patches.ndim == 2, "expected 2D array of patch coordinates"
  # x, y, width, height
  assert self.patches.shape[1] == 4, "expected second dimension to have len 4"
@@ -108,17 +105,17 @@ def __getitem__(
  coords: Sequence[int] = self.patches[idx]
  assert len(coords) == 4, "expected 4 coords (minx, miny, width, height)"
  minx, miny, width, height = coords
- source_region = dict(
- left=minx, top=miny, width=width, height=height, units="base_pixels"
- )
- target_scale = dict(mm_x=self.um_px / 1000)
 
- patch_im, _ = self.tilesource.getRegionAtAnotherScale(
- sourceRegion=source_region,
- targetScale=target_scale,
- format=large_image.tilesource.TILE_FORMAT_PIL,
+ patch_im = self.oslide.read_region(
+ location=(minx, miny), level=0, size=(width, height)
  )
  patch_im = patch_im.convert("RGB")
+ # Resize to the expected spacing. We extract the patches at their highest
+ # resolution and resize to the prescribed MPP here.
+ patch_im = patch_im.resize(
+ (self.size_factor * width, self.size_factor * height)
+ )
+
  if self.transform is not None:
  patch_im = self.transform(patch_im)
  if not isinstance(patch_im, (Image.Image, torch.Tensor)):