4 releases

0.4.0 Oct 6, 2023
0.3.2 Feb 28, 2023
0.3.1 Jan 12, 2023
0.3.0 Jan 10, 2023

#802 in Parser implementations

Download history 128/week @ 2023-12-14 40/week @ 2023-12-21 19/week @ 2023-12-28 272/week @ 2024-01-04 147/week @ 2024-01-11 69/week @ 2024-01-18 110/week @ 2024-01-25 150/week @ 2024-02-01 147/week @ 2024-02-08 157/week @ 2024-02-15 106/week @ 2024-02-22 113/week @ 2024-02-29 150/week @ 2024-03-07 84/week @ 2024-03-14 207/week @ 2024-03-21 73/week @ 2024-03-28

548 downloads per month
Used in giga-segy-out

MIT/Apache

195KB
3.5K SLoC

giga-segy-in

A set of tools for reading and writing SEGY files conforming to the SEG Technical Standards Committee's SEG-Y_r2.0 standard, written in the Rust programming language.

giga-segy-in is part of the giga-segy library workspace, which is a tool for working with data in the SEG-Y format. The giga-segy-in library provides functionality for parsing SEG-Y files of arbitrary size with a variety of options.

The library is quite lightweight, but provides options (feature flags) for allowing serialization/deserialization via serde/serde_json. NB: Functionality for the production of C bindings for header structures requires the direct use of giga-segy-core.


Getting started

Using the basic functionality of giga-segy is as simple as adding the dependencies to the [dependencies] section of the Cargo.toml of your project. Usually you only need giga-segy-in or giga-segy-out as they re-export all the necessities. However, for the generation of C bindings, you will need giga-segy-core.

[dependencies]
# I am using `giga-segy-in` for my parser.
giga-segy-in = "0.3.1"
# I only need core as a dependency because I want C bindings for the headers.
giga-segy-core = { version = "0.3.1", features = ["gen_cbindings"]}

Here is an example of a super simple SEG-Y parser that uses giga-segy.

use std::path::PathBuf;
use giga_segy_in::SegyFile;

let dir = PathBuf::from("/my/data/lives/here");
let full_path = dir.join("MyFavouriteSEGYDataset.sgy");

let file = SegyFile::open(name.to_str().unwrap(), Default::default()).unwrap();

// I want to get the text header and dump it to the terminal.
let text_header: &str = file.get_text_header();
println!("Text header: {:?}", text_header);

// Oops. SEG-Y headers look messy if we don't go line by line...
for line in file.get_text_header_lines() {
    println!("{}", line);
}

// Now to have a look at the binary header.
let bin_header = file.get_bin_header();
println!("Bin header: {}", bin_header);

// Get the data in the order of appearance of traces in the file.
// Of course there are more organised ways of doing this,
// but I just want to see the data...
for trace in file.traces_iter() {
    // First a quick peek at the trace header.
    println!("Trace header: {}", trace.get_header());
    // ..And then the data.
    // NB: trace data is not loaded to RAM until this is called.
    let data:Vec<f32> = file.get_trace_data_as_f32_from_trace(trace).unwrap();
    println!("Data: {:?}", data);
}

Flavour

The library was designed to work foremost for the GiGa infosystems codebase and thus has something of a "GiGa flavour" to it.


License

Dependencies

~0.9–1.6MB
~34K SLoC