wip: add spell checking to notebooks CI

Author: yaugenst-flex

.codespellrc

@@ -0,0 +1,7 @@
+[codespell]
+skip = .git,*.pdf,*.png,*.jpg,*.jpeg,*.gif,*.svg,*.bmp,*.tiff,*.pyc,venv,.venv,.ipynb_checkpoints
+check-filenames = true
+quiet-level = 2
+ignore-words-list = flexcompute,tidy3d,TE,TM,te,tm,FOM,fom,Commun,Thru
+ignore-regex = [a-f0-9]{40}
+builtin = clear,rare,informal 

spellcheck.py

@@ -0,0 +1,80 @@
+#!/usr/bin/env python3
+
+import argparse
+import os
+import re
+import subprocess
+import sys
+
+
+def get_relative_path(notebook):
+    """Get the relative path of the notebook from the current directory."""
+    return os.path.relpath(notebook, os.getcwd())
+
+
+def check_spelling(notebook):
+    """Check spelling in a notebook and return any errors."""
+    rel_path = get_relative_path(notebook)
+
+    with open(notebook, encoding="utf-8") as f:
+        content = f.read()
+
+    # nbstripout to remove outputs
+    nbstripout_proc = subprocess.run(
+        ["uvx", "nbstripout"], input=content, capture_output=True, text=True
+    )
+
+    # remove image tags with base64 data
+    stripped_content = re.sub(
+        r'<img\s+src="data:image/[^"]+;base64,[^"]+"[^>]*>|<img\s+src="data:image/[^"]+;base64,[^"]+"[^/>]*/>',
+        "",
+        nbstripout_proc.stdout,
+        flags=re.DOTALL,
+    )
+
+    # remove any remaining base64 strings that might appear without proper HTML tags
+    stripped_content = re.sub(
+        r"data:image/[^;]+;base64,[A-Za-z0-9+/=]+",
+        "",
+        stripped_content,
+        flags=re.DOTALL,
+    )
+
+    codespell_proc = subprocess.run(
+        ["uvx", "codespell", "-"], input=stripped_content, capture_output=True, text=True
+    )
+
+    # sadly we can't get rid of the "Used config files: ..." so we filter it here
+    output_lines = []
+    for line in codespell_proc.stdout.splitlines():
+        if "Used config files:" in line or "    1: .codespellrc" in line:
+            continue
+        output_lines.append(line)
+
+    output = "\n".join(output_lines)
+
+    if output:
+        print(f"{rel_path}:")
+        print(output.replace("-", ""))
+        print("-------------------------------------------")
+
+    return bool(output)
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Check spelling in Jupyter notebooks")
+    parser.add_argument("notebooks", nargs="+", help="List of notebook files to check")
+    args = parser.parse_args()
+
+    has_errors = False
+
+    for notebook in args.notebooks:
+        if check_spelling(notebook):
+            has_errors = True
+
+    if has_errors:
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()