asciify-them 1.0.0

No project description provided

This package allows you to turn images (only .jpg and .png tested) into ASCII art drawings.
Inspired by ascii-view.

Documentation

Full documentation is available at: asciify-them.readthedocs.io

Features

• CLI: This software can be accessed both as a Python library and as a CLI;
• Colored output: ANSI color codes allow to print colors (requires a modern terminal);
• Resizing flexibility: Images are scaled to keep aspect ratio and fit the image to the terminal, but both options can be disabled;
• Edge detection: Sobel and Canny algorithm are used to highlight edges;
• Output flexibility: Resulting images can be saved in a file (both using the terminal to determine optimal size or providing custom height and/or width).

Installation

The package can be installed through PyPi:

pip install asciify-them

But also from source:

git clone https://github.com/ndrscalia/asciify-them
cd <repo-dir>
python -m venv .venv
source .venv/bin/activate
pip install -e .

Usage

CLI

The only required argument is the path to the image:

asciify <path/to/image> [OPTIONS]

The following options are available:

• -bw, --black_white: Set the output to black&white.
• -e, --edges: Enable edge detection.
• -w, --width: Provide custom width. If not specified, terminal's size is going to determine this value. This value can be specified only when f_type='wide'.
• -he, --height: Provide custom height. If not specified, terminal's size is going to determine this value. This value can be specified only when f_type='tall'.
• -ar, --no_aspect_ratio: Disable original aspect ratio's protection.
• -f, --factor_type: Choose the downsampling factor type among the following values: in_terminal, wide, tall.
• -b, --blur: Provide a list with kernel size as a tuple, std for x axis, std for y axis. For more details refere to the docs for cv2.GaussianBlur. Changing the dafault values allow to tweak edge detection.
• ct, --canny_threshold: Provide edges detection threshold as a tuple. For more details refer to the docs for cv2.Canny.
• -at, --angles_threshold: Provide kernel size for angles calculation as an integer.
• -o, --output: Provide the output's path. If not specified, uses stdout (e.g.: terminal).

Details

The different factors available are meant for different scenarios:

• in_terminal allows to keep the output inside the terminal keeping aspect ratio;
• wide is better suited for images which are wider than taller but the output does not stay in the terminal. This option is also optimal for conversion to .png through ansee, regardless of the relation between height and width;
• tall is better suited for images which are taller than wider but the output does not stay in the terminal; If aspect ratio's protection is disabled, output will be squished by a factor to stay in the terminal.

Python library

This package can also be used as a python library. Most of the API is exposed to the user, but a convenient wrapper is also available for simpler use cases.

from asciify import asciify

# Minimal use
result = asciify("path/to/image")
print(result)

# More advanced use
result = asciify(
    "path/to/image",
    color_mode="bw",
    edges_detection=True,
    f_type="tall"
)

with open("output.txt", "w") as f:
    f.write(result)

The .txt output can be used with ansee to get a .png file out of it.
If needed, the core classes can be used as follows:

processor = ImgProcessor(image_path)

if not height and not width:
    term_height, term_width = processor.calculate_print_size()
else:
    term_height, term_width = height, width

ds_f = processor.calculate_downsample_factor(
    term_height=term_height,
    term_width=term_width,
    keep_aspect_ratio=keep_aspect_ratio,
    f_type=f_type
)

ds_img = processor.downsample_image(
    f=ds_f,
    keep_aspect_ratio=keep_aspect_ratio
)

img_hsv = processor.convert_to_hsv(image=ds_img)

angles = processor.calculate_angles(
    image=ds_img,
    k_size=angles_thresh
)

edges = processor.detect_edges(
    image=ds_img,
    blur=blur,
    canny_thresh=canny_thresh
)

renderer = Renderer(
    color_mode=color_mode,
    charset=DEFAULT_CHARSET
)

if edges_detection:
    return renderer.draw_in_ascii_with_edges(img_hsv=img_hsv, angles=angles, edges=edges)
else:
    return renderer.draw_in_ascii(img_hsv=img_hsv)

Examples

Testing

To test the codebase check tests/README.md.

Future updates and possible contributions

• Allow custom charset with different number of characters;
• Allow tuning brightness for better piping to ansee;
• Improve edges' detection.