pyterse 0.1.4

A Python interface for the TRPX compression algorithm

Pyterse

Python package of next generation TERSE/PROLIX diffraction data compression algorithm

The pyterse python package uses the c++ TERSE/PROLIX(TRPX) compression algorithm scheme (https://github.com/senikm/trpx) and adds python binders to it's main class.

---

Before you consider pyterse

• Your data is signed or unsigned integral type.
• Your data is grayscale.
• Preferably has high dynamic range.

How to install the package

Create a virtual environment

conda create -n pyterse python pip numpy pillow 

Install the package

pip install pyterse 

---

Testing the functionality of the library

Basic commands

• Create terse object:

import pyterse

#Allow for multithreading
pyterse.set_parallelism(1.0) #0 means no parallelism, while 1.0 uses all available cores

# Constructor 1
terse = pyterse.Terse() 
pyterse.compress(data, terse)

# Constructor 2
 terse = pyterse.Terse(data) #nD NumPy array or slice from an array

 # Constructor 3
 terse = Terse(data, block_size) #Provide the costum block size: default is 12
 - Example:
terse = Terse(data, 12)

• Add additional entry to Terse object:

terse.push_back(data) #The NumPy array or slice from a nD array should correspond to existing set shape (terse.shape).

• Save compressed data to a file:

terse.save('filename.trpx')

• Load compressed data from a file:

loaded_terse = Terse.load('filename.trpx')

• Decompress the data:

decompressed_frame = pyterse.decompress(loaded_terse)

• Decompress a specific entry:

decompressed_frame = pyterse.decompress(loaded_terse.at(0))
# You can check the number of entries by terse.number_of_entries()

• Add and retrieve metadata to the terse object:

terse.set_metadata("Some metadata", frame=0)
metadata = terse.get_metadata(frame=0)
metadata

• Additional functionality:

terse.size() #unique elements in the data
terse.shape #dimension of one frame
terse.number_of_frames()
terse.number_of_bytes()
terse.erase(0)
terse.insert(0, data)