python-FloatImg

A pythonesque binding to FloatImg library available at https://git.tetalab.org/tTh/FloatImg

python-FloatImg requires a Python 3 implementation (tested with 3.6.9).

Most function are wrapped into a similar api, sometime with a more consistent naming and an object-oriented approach.

Color values are specified using triplets of float values, like (128.5, 34.234, 242.23).

Installation

python-FloatImg requires a shared object file (.so) that is not yet officialy available from FloatImg.

The file floatimg/settings.py contains the path to the .so file and the encoding of path characters.

Usage

import floatimg

The FloatImg class encapsulate core and file functionnalities of FloatImg library. Instances are created with functions from the module.

Colors are tuple of three elements (r, g, b).

Image creation

FloatImg has two class methods to create an image.

img = floatimg.create(640, 480, floatimg.RGB)

Or use the create_rgb shortcut:

img = floatimg.create_rgb(640, 480)

Note that calling the destroy method of the instance is not necessary as it will be called automatically at instance destruction.

Or completely duplicate an existing image:

# without copying data
new_img = img.clone()

# with copying data
new_img = img.clone(True)

Basic pixel manipulation

# Get pixel value
color = (r, g, b) = img.get(x, y)

# Set pixel value
img.put(x, y, (r, g, b))

# Reset all pixels to zero.
img.clear()

# Copy pixels to another image of the same format
img.copy_data(another_img)

File manipulation

Raw dump files

# Dump image to file
img.dump(path)

# Restore image data from a dump file
img.load(path)

# Create a new image from a dump file
img = floatimg.create_from_dump(path)

# Get dump metadata (TODO create a FileInfo named tuple)
witdh, height, img_type = floatimg.fileinfos(path)

External formats

# export supports the following extentions : .fimg, .png, .tiff, .pnm, .fits
# save as png
img.export("test.png")

# create an image instance from png file
img = floatimg.create_from_png("test.png")

# load pixel data from png file in an existing instance. size and type have to be compatible
img.load_png("test.png")

Operators

# addition
res_img = img + img2

# substraction
res_img = img - img2

# multiplication
res_img = img * img2

# min value per color channel
res_img = img.min(img2)

# max value per color channel
res_img = img.max(img2)

Development

Run tests

Install development requirements (you may add --user if not in a virtual environment or not in system-wide install):

pip install -r requirements-devel.txt

Ensure floatimg containing path is in the PYTHONPATH, this will run the tests and show a coverage report:

pytest --cov=floatimg tests/*