dirclass 0.1.0

dirclass

Treat a directory like a dataclass for ease, recursively.

Main features

• Simple factory: dirclass(path, recursive=True, file_types="*") -> dataclass
• List files: d.all() returns a flattened list of file paths
• Read files: d.read_all() returns a flattened list of file contents
• Filtering: Accept glob patterns like "*.py" or ["*.py", "*.md"] in constructor or per-call
• Subfolder properties: Immediate subfolders become attributes (e.g., .src, .data)
  • Aliases: full sanitized name, tail after last dot (e.g., path.src → .path_src and .src), and case-insensitive access (e.g., .Gallery and .gallery)
• Robust reading: Read errors are logged and skipped (continues processing)

Install (PyPI)

pip install dirclass

Alternatively,

pip install -e .

Usage

from dirclass import dirclass

# Create a directory wrapper
d = dirclass("/path/to/folder", recursive=True, file_types=["*.py", "*.md"])

# List all matching files
paths = d.all()                 # list[str]

# Read contents of all matching files
contents = d.read_all()         # list[str]

# Override filters per call
py_only = d.all("*.py")
md_and_txt_contents = d.read_all(["*.md", "*.txt"])

# Subfolder properties (immediate children of the root)
# If your root has folders: src, data, and Gallery
src_files = d.src               # files under /path/to/folder/src
data_files = d.data             # files under /path/to/folder/data
gallery_files = d.Gallery       # case-insensitive access also works: d.gallery

# Dotted folder names (e.g., `path.src`) expose multiple aliases
# - full sanitized name: d.path_src
# - tail alias: d.src

Example

from dirclass import dirclass

d = dirclass("./project", recursive=True, file_types=["*.py", "*.md"]) 

print("All files:")
for p in d.all():
    print(p)

print("\nPython files only:")
for p in d.all("*.py"):
    print(p)

print("\nSubfolder: src")
for p in d.src:  # requires a ./project/src subfolder
    print(p)

print("\nReading markdown contents:")
for content in d.read_all("*.md"):
    print(content[:80], "...")

Codes are generated by GPT5, thus may not be reliable in the current version for test.