Use ruff-check, ruff-format instead of black, flake8 (#1500)

Issue: rapidsai/build-planning#130
Ops-Bot-Merge-Barrier: true

Authors:
  - Kyle Edwards (https://github.com/KyleFromNVIDIA)

Approvers:
  - Gil Forsyth (https://github.com/gforsyth)
  - Bradley Dice (https://github.com/bdice)
  - Ben Frederickson (https://github.com/benfred)

URL: #1500

Author: KyleFromNVIDIA

.flake8

@@ -17,23 +17,13 @@ repos:
                 # project can specify its own first/third-party packages.
                 args: ["--config-root=python/", "--resolve-all-configs"]
                 files: python/.*
-                types_or: [python, cython, pyi]
-      - repo: https://github.com/psf/black
-        rev: 22.3.0
+                types: [cython]
+      - repo: https://github.com/astral-sh/ruff-pre-commit
+        rev: v0.14.3
         hooks:
-              - id: black
-                files: python/.*
-                # Explicitly specify the pyproject.toml at the repo root, not per-project.
-                args: ["--config", "pyproject.toml"]
-      - repo: https://github.com/PyCQA/flake8
-        rev: 7.1.1
-        hooks:
-              - id: flake8
-                args: ["--config=.flake8"]
-                files: python/.*$
-                types: [file]
-                types_or: [python, cython]
-                additional_dependencies: ["flake8-force"]
+              - id: ruff-check
+                args: [--fix]
+              - id: ruff-format
       - repo: https://github.com/pre-commit/mirrors-mypy
         rev: 'v0.971'
         hooks:
@@ -110,7 +100,7 @@ repos:
                     ^CHANGELOG[.]md$|
                     ^cpp/cmake/patches/cutlass/build-export[.]patch$
       - repo: https://github.com/rapidsai/pre-commit-hooks
-        rev: v1.2.0
+        rev: v1.2.1
         hooks:
           - id: verify-copyright
             name: verify-copyright-cuvs

.pre-commit-config.yaml

@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
-# SPDX-FileCopyrightText: Copyright (c) 2023, NVIDIA CORPORATION.
+# SPDX-FileCopyrightText: Copyright (c) 2023-2025, NVIDIA CORPORATION.
 # SPDX-License-Identifier: Apache-2.0
 
 import sys
 import pandas as pd
-import numpy as np
 import matplotlib.pyplot as plt
 import seaborn as sns
 from pathlib import Path
 from matplotlib import colors
 
+
 def main(input_path):
     input_path = Path(input_path)
     print("-- loading data")
@@ -22,40 +22,56 @@ def main(input_path):
     df["file"] = df["source file name"]
     df["phase"] = df["phase name"].str.strip()
 
-    dfp = (df
-           # Remove nvcc driver entries. They don't contain a source file name
-           .query("phase!='nvcc (driver)'")
-           # Make a pivot table containing files as row, phase (preprocessing,
-           # cicc, etc.) as column and the total times as table entries. NOTE:
-           # if compiled for multiple archs, the archs will be summed.
-           .pivot_table(index="file", values="seconds", columns="phase", aggfunc='sum'))
+    dfp = (
+        df
+        # Remove nvcc driver entries. They don't contain a source file name
+        .query("phase!='nvcc (driver)'")
+        # Make a pivot table containing files as row, phase (preprocessing,
+        # cicc, etc.) as column and the total times as table entries. NOTE:
+        # if compiled for multiple archs, the archs will be summed.
+        .pivot_table(
+            index="file", values="seconds", columns="phase", aggfunc="sum"
+        )
+    )
 
     dfp_sum = dfp.sum(axis="columns")
 
     df_fraction = dfp.divide(dfp_sum, axis="index")
     df_fraction["total time"] = dfp_sum
-    df_fraction = df_fraction.melt(ignore_index=False, id_vars="total time", var_name="phase", value_name="fraction")
+    df_fraction = df_fraction.melt(
+        ignore_index=False,
+        id_vars="total time",
+        var_name="phase",
+        value_name="fraction",
+    )
 
     dfp["total time"] = dfp_sum
-    df_absolute = dfp.melt(ignore_index=False, id_vars="total time", var_name="phase", value_name="seconds")
+    df_absolute = dfp.melt(
+        ignore_index=False,
+        id_vars="total time",
+        var_name="phase",
+        value_name="seconds",
+    )
 
     # host: light red to dark red (preprocessing, cudafe, gcc (compiling))
     # device: ligt green to dark green (preprocessing, cicc, ptxas)
     palette = {
         "gcc (preprocessing 4)": colors.hsv_to_rgb((0, 1, 1)),
-        'cudafe++': colors.hsv_to_rgb((0, 1, .75)),
-        'gcc (compiling)': colors.hsv_to_rgb((0, 1, .4)),
-        "gcc (preprocessing 1)": colors.hsv_to_rgb((.33, 1, 1)),
-        'cicc': colors.hsv_to_rgb((.33, 1, 0.75)),
-        'ptxas': colors.hsv_to_rgb((.33, 1, 0.4)),
-        'fatbinary': "grey",
+        "cudafe++": colors.hsv_to_rgb((0, 1, 0.75)),
+        "gcc (compiling)": colors.hsv_to_rgb((0, 1, 0.4)),
+        "gcc (preprocessing 1)": colors.hsv_to_rgb((0.33, 1, 1)),
+        "cicc": colors.hsv_to_rgb((0.33, 1, 0.75)),
+        "ptxas": colors.hsv_to_rgb((0.33, 1, 0.4)),
+        "fatbinary": "grey",
     }
 
     print("-- Ten longest translation units:")
-    colwidth = pd.get_option('display.max_colwidth') - 1
+    colwidth = pd.get_option("display.max_colwidth") - 1
     dfp = dfp.reset_index()
     dfp["file"] = dfp["file"].apply(lambda s: s[-colwidth:])
-    print(dfp.sort_values("total time", ascending=False).reset_index().loc[:10])
+    print(
+        dfp.sort_values("total time", ascending=False).reset_index().loc[:10]
+    )
 
     print("-- Plotting absolute compile times")
     abs_out_path = f"{input_path}.absolute.compile_times.png"
@@ -64,43 +80,57 @@ def main(input_path):
         y="file",
         hue="phase",
         hue_order=reversed(
-            ["gcc (preprocessing 4)", 'cudafe++', 'gcc (compiling)',
-             "gcc (preprocessing 1)", 'cicc', 'ptxas',
-             'fatbinary',
-        ]),
+            [
+                "gcc (preprocessing 4)",
+                "cudafe++",
+                "gcc (compiling)",
+                "gcc (preprocessing 1)",
+                "cicc",
+                "ptxas",
+                "fatbinary",
+            ]
+        ),
         palette=palette,
         weights="seconds",
         multiple="stack",
         kind="hist",
         height=20,
     )
-    plt.xlabel("seconds");
+    plt.xlabel("seconds")
     plt.savefig(abs_out_path)
     print(f"-- Wrote absolute compile time plot to {abs_out_path}")
 
     print("-- Plotting relative compile times")
     rel_out_path = f"{input_path}.relative.compile_times.png"
     sns.displot(
-        df_fraction.sort_values('total time').reset_index(),
+        df_fraction.sort_values("total time").reset_index(),
         y="file",
         hue="phase",
-        hue_order=reversed(["gcc (preprocessing 4)", 'cudafe++', 'gcc (compiling)',
-                   "gcc (preprocessing 1)", 'cicc', 'ptxas',
-                   'fatbinary',
-                  ]),
+        hue_order=reversed(
+            [
+                "gcc (preprocessing 4)",
+                "cudafe++",
+                "gcc (compiling)",
+                "gcc (preprocessing 1)",
+                "cicc",
+                "ptxas",
+                "fatbinary",
+            ]
+        ),
         palette=palette,
         weights="fraction",
         multiple="stack",
         kind="hist",
         height=15,
     )
-    plt.xlabel("fraction");
+    plt.xlabel("fraction")
     plt.savefig(rel_out_path)
     print(f"-- Wrote relative compile time plot to {rel_out_path}")
 
+
 if __name__ == "__main__":
     if len(sys.argv) != 2:
-        printf("""NVCC log analyzer
+        print("""NVCC log analyzer
 
         Analyzes nvcc logs and outputs a figure with highest ranking translation
         units.

cpp/scripts/analyze_nvcc_log.py

@@ -55,17 +55,20 @@ def repo_version_major_minor():
 
     full_repo_version = repo_version()
 
-    match = re.match(r"^v?(?P<major>[0-9]+)(?:\.(?P<minor>[0-9]+))?",
-                     full_repo_version)
-
-    if (match is None):
-        print("   [DEBUG] Could not determine repo major minor version. "
-              f"Full repo version: {full_repo_version}.")
+    match = re.match(
+        r"^v?(?P<major>[0-9]+)(?:\.(?P<minor>[0-9]+))?", full_repo_version
+    )
+
+    if match is None:
+        print(
+            "   [DEBUG] Could not determine repo major minor version. "
+            f"Full repo version: {full_repo_version}."
+        )
         return None
 
     out_version = match.group("major")
 
-    if (match.group("minor")):
+    if match.group("minor"):
         out_version += "." + match.group("minor")
 
     return out_version
@@ -91,44 +94,50 @@ def determine_merge_commit(current_branch="HEAD"):
 
     try:
         # Try to determine the target branch from the most recent tag
-        head_branch = __git("describe",
-                            "--all",
-                            "--tags",
-                            "--match='branch-*'",
-                            "--abbrev=0")
+        head_branch = __git(
+            "describe", "--all", "--tags", "--match='branch-*'", "--abbrev=0"
+        )
     except subprocess.CalledProcessError:
-        print("   [DEBUG] Could not determine target branch from most recent "
-              "tag. Falling back to 'branch-{major}.{minor}.")
+        print(
+            "   [DEBUG] Could not determine target branch from most recent "
+            "tag. Falling back to 'branch-{major}.{minor}."
+        )
         head_branch = None
 
-    if (head_branch is not None):
+    if head_branch is not None:
         # Convert from head to branch name
         head_branch = __git("name-rev", "--name-only", head_branch)
     else:
         # Try and guess the target branch as "branch-<major>.<minor>"
         version = repo_version_major_minor()
 
-        if (version is None):
+        if version is None:
             return None
 
         head_branch = "branch-{}".format(version)
 
     try:
         # Now get the remote tracking branch
-        remote_branch = __git("rev-parse",
-                              "--abbrev-ref",
-                              "--symbolic-full-name",
-                              head_branch + "@{upstream}")
+        remote_branch = __git(
+            "rev-parse",
+            "--abbrev-ref",
+            "--symbolic-full-name",
+            head_branch + "@{upstream}",
+        )
     except subprocess.CalledProcessError:
-        print("   [DEBUG] Could not remote tracking reference for "
-              f"branch {head_branch}.")
+        print(
+            "   [DEBUG] Could not remote tracking reference for "
+            f"branch {head_branch}."
+        )
         remote_branch = None
 
-    if (remote_branch is None):
+    if remote_branch is None:
         return None
 
-    print(f"   [DEBUG] Determined TARGET_BRANCH as: '{remote_branch}'. "
-          "Finding common ancestor.")
+    print(
+        f"   [DEBUG] Determined TARGET_BRANCH as: '{remote_branch}'. "
+        "Finding common ancestor."
+    )
 
     common_commit = __git("merge-base", remote_branch, current_branch)
 
@@ -166,9 +175,9 @@ def changedFilesBetween(baseName, branchName, commitHash):
     # checkout latest commit from branch
     __git("checkout", "-fq", commitHash)
 
-    files = __gitdiff("--name-only",
-                      "--ignore-submodules",
-                      f"{baseName}..{branchName}")
+    files = __gitdiff(
+        "--name-only", "--ignore-submodules", f"{baseName}..{branchName}"
+    )
 
     # restore the original branch
     __git("checkout", "--force", current)
@@ -180,13 +189,15 @@ def changesInFileBetween(file, b1, b2, filter=None):
     current = branch()
     __git("checkout", "--quiet", b1)
     __git("checkout", "--quiet", b2)
-    diffs = __gitdiff("--ignore-submodules",
-                      "-w",
-                      "--minimal",
-                      "-U0",
-                      "%s...%s" % (b1, b2),
-                      "--",
-                      file)
+    diffs = __gitdiff(
+        "--ignore-submodules",
+        "-w",
+        "--minimal",
+        "-U0",
+        "%s...%s" % (b1, b2),
+        "--",
+        file,
+    )
     __git("checkout", "--quiet", current)
     lines = []
     for line in diffs.splitlines():
@@ -215,25 +226,29 @@ def modifiedFiles(pathFilter=None):
     currentBranch = branch()
     print(
         f"   [DEBUG] TARGET_BRANCH={targetBranch}, COMMIT_HASH={commitHash}, "
-        f"currentBranch={currentBranch}")
+        f"currentBranch={currentBranch}"
+    )
 
     if targetBranch and commitHash and (currentBranch == "current-pr-branch"):
         print("   [DEBUG] Assuming a CI environment.")
         allFiles = changedFilesBetween(targetBranch, currentBranch, commitHash)
     else:
-        print("   [DEBUG] Did not detect CI environment. "
-              "Determining TARGET_BRANCH locally.")
+        print(
+            "   [DEBUG] Did not detect CI environment. "
+            "Determining TARGET_BRANCH locally."
+        )
 
         common_commit = determine_merge_commit(currentBranch)
 
-        if (common_commit is not None):
-
+        if common_commit is not None:
             # Now get the diff. Use --staged to get both diff between
             # common_commit..HEAD and any locally staged files
-            allFiles = __gitdiff("--name-only",
-                                 "--ignore-submodules",
-                                 "--staged",
-                                 f"{common_commit}").splitlines()
+            allFiles = __gitdiff(
+                "--name-only",
+                "--ignore-submodules",
+                "--staged",
+                f"{common_commit}",
+            ).splitlines()
         else:
             # Fallback to just uncommitted files
             allFiles = uncommittedFiles()