segfast is a library for interacting with SEG-Y seismic data. Main features are:

• Faster access to read data: both traces headers and values
• Optional bufferization, where user can provide a preallocated memory to load the data into
• Convenient API that relies on numpy.memmap for most operations, while providing segyio as a fallback engine

# pip / pip3
pip3 install segfast
# developer version (add `--depth 1` if needed)
git clone https://github.com/analysiscenter/segfast.git

Timings for reading data along various projections:

After installation just import segfast into your code. A quick demo of our primitives and methods:

import segfast
# Open file and read some meta info. Engine can be `segyio` or `memmap`
segfast_file = segfast.open('/path/to/cube.sgy', engine='memmap')
# Load requested headers as dataframe
segfast_file.load_headers(['INLINE_3D', 'CROSSLINE_3D', ...])
# Data access. All methods support optional buffer as target memory
segfast_file.load_traces([123, 333, 777], buffer=None)
segfast_file.load_depth_slices([5, 10, 15], buffer=None)
# Convert data format to IEEE float32: speeds up operations by a lot
segfast_file.convert(format=5)

You can get more familiar with the library, its functional and timings by reading examples.