From c3b506fca6cd03a434e0974c66d0bf4b345dd039 Mon Sep 17 00:00:00 2001
From: konsti <konstin@mailbox.org>
Date: Fri, 21 Jul 2023 11:32:35 +0200
Subject: [PATCH] Add script to shrink all formatter errors (#5943)

**Summary** Add script to shrink all formatter errors: This started as a
fun idea and turned out really useful: This script gives us a single
Python file with all formatter stability errors. I want to keep it
around to occasionally update #5828 so I added it to the git.

**Test Plan** None, this is a helper script
---
 .../shrink_formatter_errors.py                | 75 +++++++++++++++++++
 1 file changed, 75 insertions(+)
 create mode 100644 crates/ruff_python_formatter/shrink_formatter_errors.py

diff --git a/crates/ruff_python_formatter/shrink_formatter_errors.py b/crates/ruff_python_formatter/shrink_formatter_errors.py
new file mode 100644
index 0000000000000..3d104b7c0317e
--- /dev/null
+++ b/crates/ruff_python_formatter/shrink_formatter_errors.py
@@ -0,0 +1,75 @@
+"""Take `format-dev --stability-check` output and shrink all stability errors into a
+single Python file. Used to update https://github.com/astral-sh/ruff/issues/5828 ."""
+
+from __future__ import annotations
+
+import json
+import os
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from pathlib import Path
+from subprocess import check_output
+from tempfile import NamedTemporaryFile
+
+from tqdm import tqdm
+
+root = Path(
+    check_output(["git", "rev-parse", "--show-toplevel"], text=True).strip(),
+)
+target = root.joinpath("target")
+
+error_report = target.joinpath("formatter-ecosystem-errors.txt")
+error_lines_prefix = "Unstable formatting "
+
+
+def get_filenames() -> list[str]:
+    files = []
+    for line in error_report.read_text().splitlines():
+        if not line.startswith(error_lines_prefix):
+            continue
+        files.append(line.removeprefix(error_lines_prefix))
+    return files
+
+
+def shrink_file(file: str) -> tuple[str, str]:
+    """Returns filename and minimization"""
+    with NamedTemporaryFile(suffix=".py") as temp_file:
+        print(f"Starting {file}")
+        ruff_dev = target.joinpath("release").joinpath("ruff_dev")
+        check_output(
+            [
+                target.joinpath("release").joinpath("ruff_shrinking"),
+                file,
+                temp_file.name,
+                "Unstable formatting",
+                f"{ruff_dev} format-dev --stability-check {temp_file.name}",
+            ],
+        )
+        print(f"Finished {file}")
+        return file, Path(temp_file.name).read_text()
+
+
+def main():
+    storage = target.joinpath("minimizations.json")
+    output_file = target.joinpath("minimizations.py")
+    if storage.is_file():
+        outputs = json.loads(storage.read_text())
+    else:
+        outputs = {}
+    files = sorted(set(get_filenames()) - set(outputs))
+    # Each process will saturate one core
+    with ThreadPoolExecutor(max_workers=os.cpu_count()) as executor:
+        tasks = [executor.submit(shrink_file, file) for file in files]
+        for future in tqdm(as_completed(tasks), total=len(files)):
+            file, output = future.result()
+            outputs[file] = output
+            storage.write_text(json.dumps(outputs, indent=4))
+
+    # Write to one shareable python file
+    with output_file.open("w") as formatted:
+        for file, code in sorted(json.loads(storage.read_text()).items()):
+            file = file.split("/target/checkouts/")[1]
+            formatted.write(f"# {file}\n{code}\n")
+
+
+if __name__ == "__main__":
+    main()