pySISF Documentation

Module contents

pySISF.

pySISF.sisf.create_metadata(version, dtype, channel_count, mchunk, res, size)

pySISF.sisf.create_shard(fname_data: str, fname_meta: str, data, chunk_size, compression, compression_opts=None, thread_count=8, chunk_batch=1024, crop=None, progress=True) → None

Function to create a SISF shard.

Parameters:

• fname_data (str) – Name of the data file to create.

• fname_meta (str) – Name of the metadata file to create.

• data (3D numpy array-like) – Raw data for the chunk.

• chunk_size (3-tuple of int) – Size of each chunk

• compression (int) – compression codec to use

• thread_count (int, default 8) – number of threads to use for data packing

• chunk_batch (int, default 1024) – number of jobs to allocate per thread for load balancing

• crop (3-tuple of int, default None) – if set, encodes a crop factor into the shard

• progress (bool, default True) – prints a loading bar using tqdm

pySISF.sisf.create_shard_worker(data, coords, compression, compression_opts=None)

pySISF.sisf.create_sisf(fname: str, data, mchunk_size, chunk_size, res, enable_status=True, downsampling=None, compression=1, thread_count=8, compression_opts=None) → None

Function to create a SISF archive.

Parameters:

• fname (string) – Name of the folder to place the SISF archive into, created if does not exist.

• data (numpy array-like) – Represents the data to be converted in CXYZ format.

• mchunk_size (3-tuple) – size of metachunks to create, e.g. (2000,2000,2000).

• res (3-tuple) – resolution of the dataset in nanometers (nm), e.g. (100,100,100).

• enable_status (bool, default True) – If true, print out a loading bar for creation using tqdm.

• downsampling (int, default None) – How many downsample tiers to generate.

• compression (int, default 1->ZSTD) – What compression codec to use.

• thread_count (int, default 8) – How many threads to use for data packing.

pySISF.sisf.get_dtype_code(i)

pySISF.sisf.iterate_bounded(max_val, step_size)

class pySISF.sisf.sisf(fname)

Bases: object

get_chunk(x, y, z, c, s)

parse_metadata()

property shape

class pySISF.sisf.sisf_chunk(fname_data, fname_meta, parent=None)

Bases: object

find_index(x, y, z)

get_chunk(idx)

get_chunk_coords(idx)

get_chunk_numpy(idx)

get_chunk_size(idx)

get_metadata(idx)

static iterate_chunks(rstart, rstop, cs)

parse_metadata()

read_pixel(x, y, z)

property shape

Module contents

pySISF.sndif_utils.downsample(in_array, out_array)

Utility function to downsample a 3D image by a factor of 2X in all dimensions.

Parameters:

• in_array (numpy) – 3D array containing the input image

• out_array (numpy) – identical-sized 3D array to write the output to

Returns:

Numpy array containing the downsampled image

pySISF.sndif_utils.load_from_zip(file_name, stack_size=2000, stack_select=None, thread_count=1)

Load image frames from a SNDIF ZIP archive.

Parameters:

• file_name (str) – Name of the input ZIP file.

• stack_size (int) – Number of frames expected.

• stack_select (slice) – A parameter passed to the file list to select inputs.

• thread_count (int) – Number of threads to launch on the ThreadPoolExecutor.

Returns:

Numpy array containing the loaded frames

Module contents

class pySISF.vidlib.EncoderType(value)

Bases: Enum

An enumeration.

AV1_AOM = 3

AV1_SVT = 4

X264 = 1

X265 = 2

pySISF.vidlib.decode_stack(input_blob, dims=(128, 128), method='libx264', debug=False, fps='24/1')

pySISF.vidlib.encode_stack(input_stack, method=EncoderType.X264, debug=False, fps=24, compression_opts=None)