Optimize file I/O by caching pyproject data during first pass

[Copilot]

Addresses review feedback about inefficient file I/O in generate_dependency_diagram.py where each pyproject.toml was read twice—once for package names, once for dependencies.

Changes

• Added pyproject_data_cache dictionary to store parsed pyproject data during the first pass
• Second pass now uses cached data instead of re-reading files

# Before: reading file twice
for path_str, pkg_type in package_configs:
    pyproject_data = read_pyproject_toml(pyproject_path)  # first read
    pkg_name = pyproject_data['project']['name']
    ...

for pkg in packages:
    pyproject_data = read_pyproject_toml(pkg.path)  # second read
    pkg.dependencies = extract_dependencies(pyproject_data, all_package_names)

# After: cache and reuse
pyproject_data_cache: Dict[str, Dict] = {}

for path_str, pkg_type in package_configs:
    pyproject_data = read_pyproject_toml(pyproject_path)
    pyproject_data_cache[pkg_name] = pyproject_data  # cache it
    ...

for pkg in packages:
    pyproject_data = pyproject_data_cache[pkg.name]  # use cache
    pkg.dependencies = extract_dependencies(pyproject_data, all_package_names)

Two-pass approach retained since all package names must be known before filtering internal dependencies.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.