SegyioLoader

class segfast.segyio_loader.SegyioLoader(path, endian='big', strict=False, ignore_geometry=True)

A thin wrapper around segyio library for convenient loading of headers and traces.

Most of the methods directly call the public API of segyio.

For trace loading, we use private methods and attributes of segyio.SegyFile, which allow:

• reading data into the pre-defined buffer

• read only parts of the trace

This gives up to 50% speed-up over public API for the scenario of loading sequence of traces, and up to 15% over public API in case of loading full lines (inlines or crosslines).

property sample_interval

Sample interval of seismic traces.

property delay

Delay recording time of seismic traces.

load_headers(headers, indices=None, reconstruct_tsf=True, sort_columns=True, return_specs=False, tracewise=True, pbar=False, **kwargs)

Load requested trace headers from a SEG-Y file for each trace into a dataframe. If needed, we reconstruct the 'TRACE_SEQUENCE_FILE' manually be re-indexing traces.

Parameters:

• headers (sequence) –

  An array-like where each element can be:

  • str – header name,

  • int – header starting byte,

  • TraceHeaderSpec – used as is,

  • tuple – args to init TraceHeaderSpec,

  • dict – kwargs to init TraceHeaderSpec.

  Note that for SegyioLoader all nonstandard headers byte positions and dtypes will be ignored.

• indices (sequence or None) – Indices of traces to load trace headers for. If not given, trace headers are loaded for all traces.

• reconstruct_tsf (bool) – Whether to reconstruct TRACE_SEQUENCE_FILE manually.

• sort_columns (bool) – Whether to sort columns in the resulting dataframe by their starting bytes.

• return_specs (bool) – Whether to return header specs used to load trace headers.

• tracewise (bool) – Whether to iterate over the file in a trace-wise manner, instead of header-wise.

• pbar (bool, str) – If bool, then whether to display the progress bar over the file sweep. If str, then type of progress bar to display: 't' for textual, 'n' for widget.

Return type:

pandas.DataFrame

load_header(header, indices=None, **kwargs)

Load exactly one header.

make_headers_specs(headers)

Transform headers list to list of TraceHeaderSpec instances.

load_traces(indices, limits=None, buffer=None)

Load traces by their indices. By pre-allocating memory for all of the requested traces, we significantly speed up the process.

Parameters:

• indices (sequence) – Indices (TRACE_SEQUENCE_FILE) of the traces to read.

• limits (sequence of ints, slice, optional) – Slice of the data along the depth axis.

• buffer (numpy.ndarray, optional) – Buffer to read the data into. If possible, avoids copies.

process_limits(limits)

Convert given limits to a slice instance.

load_trace(index, buffer, limits)

Load one trace into the buffer.

load_depth_slices(indices, buffer=None)

Load horizontal (depth) slices of the data. Requires an almost full sweep through SEG-Y, therefore is slow.

Parameters:

• indices (sequence) – Indices (ordinals) of the depth slices to read.

• buffer (numpy.ndarray, optional) – Buffer to read the data into. If possible, avoids copies.

Return type:

numpy.ndarray

load_depth_slice(index, buffer)

Load one depth slice into buffer.

make_chunk_iterator(chunk_size=None, n_chunks=None, limits=None, buffer=None)

Create an iterator over the entire file traces in chunks.

Each chunk contains no more than chunk_size traces. If chunk_size is not provided and n_chunks is given instead, there are no more than n_chunks chunks. One and only one of chunk_size and n_chunks should be provided.

Each element in the iterator is a dictionary with 'data', 'start' and 'end' keys.

Parameters:

• chunk_size (int, optional) – Maximum size of the chunk.

• n_chunks (int, optional) – Maximum number of chunks.

• limits (sequence of ints, slice, optional) – Slice of the data along the depth (last) axis. Passed directly to load_traces().

• buffer (numpy.ndarray, optional) – Buffer to read the data into. If possible, avoids copies. Passed directly to load_traces().

Returns:

• iterator, info (tuple with two elements)

• iterator (iterable) – An iterator over the entire SEG-Y traces. Each element in the iterator is a dictionary with 'data', 'start' and 'end' keys.

• info (dict) – Description of the iterator with 'chunk_size', 'n_chunks', 'chunk_starts' and 'chunk_ends' keys.

chunk_iterator(chunk_size=None, n_chunks=None, limits=None, buffer=None)

A shorthand for make_chunk_iterator() with no info returned.

class segfast.segyio_loader.SafeSegyioLoader(path, endian='big', strict=False, ignore_geometry=True)

A thin wrapper around segyio library for convenient loading of headers and traces.

Unlike SegyioLoader, uses only public APIs to load traces.

Used mainly for performance measurements.

load_trace(index, buffer, limits)

Load one trace into buffer.

load_depth_slice(index, buffer)

Load one depth slice into buffer.