Myosotis-Researches 0.1.27

A repository for storing my progress of researches.

Myosotis-Researches

CcGAN (myosotis_researches.CcGAN)

visualize

The visualize module can display datasets as a webpage

Import with code

from myosotis_researches.CcGAN.visualize import *

Now we only have visualize_datasets function, defined as

visualize_datasets(
  indexes,
  datasets_path,
  list_path,
  template_path = resources.files("myosotis_researches").joinpath("CcGAN", "visualize", "src", "template,html"),
  host = "127.0.0.1",
  port = 8000,
  debug = True,
  img_size = 64
)

internal

The internal module is used for setting the local package itself, like installing datasets and so on.

Import with code

from myosotis_researches.CcGAN.internal import *

Function	Desctiption
install_datasets(datasets_name)	Install the datasets in datasets_name to the local python package.
uninstall_datasets()	Remove all the datasets installed to the local python package.
show_datasets()	Show all datasets installed.

Note:

1. The path of the installed datasets are

   resources.files("myosotis_researches").join("CcGAN", "<datasets_name>")

   To run this code, remember to add from importlib import resources at the beginning.

utils

The utils module contains some basic functions and classes which are frequently used during the CcGAN research.

Import with code

from myosotis_researches.CcGAN.utils import *

Function	Description
concat_image(img_list, gap=2, direction="vertical")	Concat images vertically or horizontally.
make_h5(old_datasets_name, size, new_datasets_path, image_indexes, train_indexes, val_indexes)	Get piece of original HDF5 datasets.
parse_opts()	Parse arguments.
print_hdf5(name, obj)	Print a basic structure of an HDF5 file.

Class	Description
IMGs_dataset	Images dataset.
SimpleProgressBar	Simple progress bars.

Note:

1. Function print_hdf5 should be used within a with block:

   import h5py
   
   with h5py.File(<HDF5_file_path>, "r") as f:
     f.visititems(print_hdf5)