Add extra linter checks and make all those checks pass.

Author: k-rl

examples/mrbles.ipynb

@@ -55,7 +55,7 @@
    "outputs": [],
    "source": [
     "# Setup the pipeline. You should rarely need to modify these values outside of `num_codes`\n",
-    "# unless you're switching between different image binning modes (since that changes the pixel sizes).\n",
+    "# unless you're switching between different image binning modes (since that changes pixel sizes).\n",
     "xp = mg.mrbles(  # The type of experiment we're processing, in this case mrbles.\n",
     "    \"data/xp1.ome.tif\",  # The path to the image we collected.\n",
     "    darkfield=\"data/darkfield.tiff\",  # The location of the darkfield correction image.\n",
@@ -103,7 +103,9 @@
     }
    ],
    "source": [
-    "# Now let's check the image corresponding to the 620 channel to make sure magnify correctly processed our experiment.\n",
+    "# Now let's check the image corresponding to the 620 channel to make sure magnify correctly\n",
+    "# processed our experiment.\n",
+    "\n",
     "# First display the image.\n",
     "plt.imshow(xp.image.sel(channel=\"620\"))\n",
     "\n",
@@ -163,7 +165,7 @@
     }
    ],
    "source": [
-    "# We can also visualize the lanthanide ratios for each bead. We see here that we get distinct clusters.\n",
+    "# We can also visualize the lanthanide ratios for each bead. We see that we get distinct clusters.\n",
     "plt.scatter(xp.ln_ratio.sel(ln=\"dy\"), xp.ln_ratio.sel(ln=\"sm\"), s=1.0, alpha=0.3)\n",
     "plt.xlabel(\"Dy Ratio\")\n",
     "plt.ylabel(\"Sm Ratio\")"
@@ -299,10 +301,13 @@
    },
    "outputs": [],
    "source": [
-    "# Setup the pipeline. Here we needed to change darkfield, flatfield, min_bead_radius, and max_bead_radius to account for the 2x2 binning.\n",
-    "# The filepath supports globs (https://en.wikipedia.org/wiki/Glob_(programming)) so * expands to match anything that matches the pattern.\n",
-    "# (row), (col) are equivalent to * but they also save the segment of the filename they match in the resulting experiment.\n",
-    "# In this case we're telling the pipeline that the filename specified the tile column and row.\n",
+    "# Setup the pipeline. Here we needed to change darkfield, flatfield, min_bead_radius,\n",
+    "# and max_bead_radius to account for the 2x2 binning.\n",
+    "# The filepath supports globs (https://en.wikipedia.org/wiki/Glob_(programming)) so * expands to\n",
+    "# match anything that matches the pattern.\n",
+    "# (row), (col) are equivalent to * but they also save the segment of the filename they match\n",
+    "# in the resulting experiment. In this case we're telling the pipeline that the filename specified\n",
+    "# the tile column and row.\n",
     "xp = mg.mrbles(\n",
     "    \"data/06.08.23 1-50 mix/mixed all tile 2b2_2/*Pos(col)_(row).ome.tif\",\n",
     "    darkfield=\"data/darkfield2x2.tiff\",\n",

pyproject.toml

@@ -53,5 +53,22 @@ build-backend = "uv_build"
 line-length = 100
 target-version = "py312"
 
+[tool.ruff.lint]
+select = [
+    # pycodestyle
+    "E",
+    "W",
+    # Pyflakes
+    "F",
+    # pyupgrade
+    "UP",
+    # flake8-bugbear
+    "B",
+    # flake8-simplify
+    "SIM",
+    # isort
+    "I",
+]
+
 [tool.ruff.format]
 docstring-code-format = true

src/magnify/filter.py

@@ -14,18 +14,14 @@ def filter_expression(
     search_channel: str | list[str] | None = None,
     min_contrast: int | None = None,
 ):
-    if search_channel is None:
-        search_channels = assay.channel
-    else:
-        search_channels = utils.to_list(search_channel)
-
+    search_channels = assay.channel if search_channel is None else utils.to_list(search_channel)
     valid = xr.zeros_like(assay.valid, dtype=bool)
     for channel in search_channels:
         subassay = assay.isel(time=0).sel(channel=channel)
         fg = subassay.roi.where(subassay.fg).median(dim=["roi_x", "roi_y"]).compute()
         bg = subassay.roi.where(subassay.bg).median(dim=["roi_x", "roi_y"]).compute()
         if min_contrast is None:
-            # Compute the intensity differences between every pair of backgrounds on the first timestep.
+            # Compute intensity differences between every pair of backgrounds on the first timestep.
             bg_n = bg.to_numpy().flatten()
             diffs = bg_n[:, np.newaxis] - bg_n[np.newaxis, :]
             offdiag = np.ones_like(diffs, dtype=bool) & (~np.eye(len(diffs), dtype=bool))
@@ -47,11 +43,7 @@ def filter_nonround(
     min_roundness: float = 0.75,
     search_channel: str | list[str] | None = None,
 ):
-    if search_channel is None:
-        search_channels = assay.channel
-    else:
-        search_channels = utils.to_list(search_channel)
-
+    search_channels = assay.channel if search_channel is None else utils.to_list(search_channel)
     valid = assay.valid.to_numpy()
     for channel in search_channels:
         subassay = assay.isel(time=0).sel(channel=channel)
@@ -72,11 +64,7 @@ def filter_nonround(
 
 @registry.component("filter_leaky")
 def filter_leaky_buttons(assay: xr.Dataset, search_channel: str | list[str] | None = None):
-    if search_channel is None:
-        search_channels = assay.channel
-    else:
-        search_channels = utils.to_list(search_channel)
-
+    search_channels = assay.channel if search_channel is None else utils.to_list(search_channel)
     tag = assay.tag.to_numpy()
     valid = assay.valid.to_numpy()
     rows = assay.mark_row.to_numpy()

src/magnify/find.py

@@ -145,13 +145,12 @@ def __call__(self, assay: xr.Dataset) -> xr.Dataset:
                 # Skip this timestep since we've already processed it.
                 continue
 
-            if t < self.search_timesteps[0]:
-                # Backfill timesteps that come before the first searched timestep.
-                copy_t = self.search_timesteps[0]
-            else:
-                # Re-use the button locations of the timestep just before this one since it's either
-                # a searched timestep or a timestep we just copied locations into.
-                copy_t = t - 1
+            # Either backfill timesteps that come before the first searched timestep or
+            # re-use the button locations of the timestep just before this one since it's either
+            # a searched timestep or a timestep we just copied locations into.
+            copy_t = self.search_timesteps[0] = (
+                self.search_timesteps[0] if t < self.search_timesteps[0] else t - 1
+            )
 
             # Preload all images for this timestep so we only read from disk once and
             # convert all relevant data to numpy arrays since iterating through xarrays is slow.
@@ -183,7 +182,8 @@ def __call__(self, assay: xr.Dataset) -> xr.Dataset:
             assay["fg"] = assay.fg.persist()
             assay["bg"] = assay.bg.persist()
         assay = assay.stack(mark=("mark_row", "mark_col"), create_index=True).transpose("mark", ...)
-        # Rechunk the array to chunk along markers since users will usually want to slice along that dimension.
+        # Rechunk the array to chunk along markers since users will usually want
+        # to slice along that dimension.
         mark_chunk_size = min(
             math.ceil(chunk_bytes / (roi_bytes * assay.sizes["time"] * assay.sizes["channel"])),
             num_rows,
@@ -739,7 +739,7 @@ def regress_clusters(
     )
     # Re-estimate intercepts using a weighted mean of global and local estimates.
     # This reduces outlier effects while still allowing uneven intercepts from image stitching.
-    for i, (x, y) in enumerate(cluster_points):
+    for i, (x, _y) in enumerate(cluster_points):
         if ideal_num_points[i] != 0 and not_nan[i]:
             weight = min(len(x), ideal_num_points[i]) / ideal_num_points[i]
             intercepts[i] = weight * intercepts[i] + (1 - weight) * (intercept_m * i + intercept_b)

src/magnify/identify.py

@@ -173,7 +173,8 @@ def fit_1d(points, codes, counts, N=100):
         if proportions[i] > 1:
             covs[i] += np.cov(X_r[tag_idxs == i], rowvar=False)
 
-    # Set all component variances to be the same since initializing individual covariances can lead to huge terms.
+    # Set all component variances to be identical since initializing individual
+    # covariances can lead to huge terms.
     covs[:] = np.median(covs, axis=0)
     # Initialize the uniform component.
     proportions[-1] = 1e-10
@@ -184,7 +185,7 @@ def fit_1d(points, codes, counts, N=100):
 
     tag_names = np.append(tag_names, "outlier")
     # Run the Expectation-Maximization algorithm.
-    for i in range(50):
+    for _ in range(50):
         # E-step: Compute the probability of each point belonging to each component.
         diff = X[:, np.newaxis, :] - means[np.newaxis, :, :]
         # Work in log space most of the time to avoid numerical issues.

src/magnify/pipeline.py

@@ -1,4 +1,4 @@
-from typing import Callable
+from collections.abc import Callable
 
 import xarray as xr
 from numpy.typing import ArrayLike
@@ -55,7 +55,7 @@ def func(xp):
             raise ValueError("Only one of after, before, first, and last can be set.")
 
         # Check that the new component's name is unique
-        if self.components and name in next(zip(*self.components)):
+        if self.components and name in next(zip(*self.components, strict=True)):
             raise ValueError(f"A component with the name '{name}' already exists in the pipeline.")
 
         # Find where to insert the component.
@@ -80,7 +80,7 @@ def func(xp):
     def remove_pipe(self, name: str) -> None:
         if not self.components:
             raise ValueError(f"Cannot remove pipe '{name}': pipeline has no components")
-        component_names = list(zip(*self.components))[0]
+        component_names = list(zip(*self.components, strict=True))[0]
         if name not in component_names:
             raise ValueError(f"Component '{name}' not found in pipeline")
         idx = component_names.index(name)

src/magnify/plot/__init__.py

@@ -1,15 +1,11 @@
 __all__ = [
     "imshow",
     "mrbles_clusters",
-    "ndplot",
-    "relplot",
     "roishow",
     "set_style",
 ]
 from magnify.plot.image import imshow, roishow
 from magnify.plot.mrbles import mrbles_clusters
-from magnify.plot.ndplot import ndplot
-from magnify.plot.relation import relplot
 from magnify.plot.style import set_style
 
 set_style()

src/magnify/plot/image.py

@@ -17,7 +17,7 @@ def roishow(xp: xr.Dataset):
     roi = np.zeros((counts.max(), len(tags)) + xp.roi.isel(mark=0).shape)
     fg = np.zeros((counts.max(), len(tags)) + xp.roi.isel(mark=0, channel=0).shape, dtype=bool)
     bg = np.zeros_like(fg)
-    for i, (tag, group) in enumerate(xp.roi.groupby("tag")):
+    for i, (_tag, group) in enumerate(xp.roi.groupby("tag")):
         roi[: group.sizes["mark"], i] = group
         fg[: group.sizes["mark"], i] = group.fg.isel(channel=0)
         bg[: group.sizes["mark"], i] = group.bg.isel(channel=0)