10 releases

0.1.4 Apr 14, 2024
0.1.3 Mar 15, 2024
0.1.1 Feb 27, 2024
0.1.0 Jul 21, 2023
0.0.2-alpha Jun 18, 2023

#33 in Geospatial

Download history 178/week @ 2024-02-19 287/week @ 2024-02-26 46/week @ 2024-03-04 209/week @ 2024-03-11 29/week @ 2024-03-18 49/week @ 2024-04-01 118/week @ 2024-04-08 49/week @ 2024-04-15

217 downloads per month

MIT/Apache

8.5MB
8K SLoC

Rust 6.5K SLoC // 0.0% comments SQL 1K SLoC // 0.5% comments

osm-io

This crate provides tools used to manipulate the Open Street Map data.

The basic OSM data model is broadly defined by OSM XML with Node and Way objects defining geometry and Relation and Tag objects providing unstructured extension mechanisms. Due to different uses of the OSM data, variety of data formats emerged to store and transmit it. Our main focus here are the following formats:

The goal at this stage is to be able to load large *.osm.pbf files, such as planet.osm.pbf into a Postgresql OSM database (apidb schema) and to dump an entire Postgres OSM database into a planet.osm.pbf file. See osm-admin

Because the data sets are very large, a special attention is given to maintaining control over memory size and utilizing multiple CPU cores whenever is possible.

Roadmap

  • implement *.osm.pbf reader and writer - Done
  • implement *.osm.pbf parallel reader and parallel writer - Done
  • implement apidb reader and writer - Done
  • provide basic filtering (see example below) - Done
  • convert between *.osm.pbf and apidb and vice versa - Done see examples.
  • S2 indexing - index the entire OSM dataset by S2 cells for farther processing
  • context indexing - index the entire OSM dataset by relations between its objects. So, for example, it would be possible to efficiently discard all Nodes that belong to a deleted Way.

Issues

Issues are welcome and appreciated. Please submit to https://github.com/navigatorsguild/osm-io/issues

Examples

Example for filtering out nodes from *.osm.pbf extract

use std::path::PathBuf;

use anyhow;
use benchmark_rs::stopwatch::StopWatch;
use simple_logger::SimpleLogger;

use osm_io::osm::model::element::Element;
use osm_io::osm::pbf;
use osm_io::osm::pbf::compression_type::CompressionType;
use osm_io::osm::pbf::file_info::FileInfo;

pub fn main() -> Result<(), anyhow::Error> {
    SimpleLogger::new().init()?;
    log::info!("Started pbf io pipeline");
    let mut stopwatch = StopWatch::new();
    stopwatch.start();
    let input_path = PathBuf::from("./tests/fixtures/niue-230109.osm.pbf");
    let output_path = PathBuf::from("./target/results/niue-230109.osm.pbf");
    let reader = pbf::reader::Reader::new(&input_path)?;
    let mut file_info = FileInfo::default();
    file_info.with_writingprogram_str("pbf-io-example");
    let mut writer = pbf::writer::Writer::from_file_info(
        output_path,
        file_info,
        CompressionType::Zlib,
    )?;

    writer.write_header()?;

    for element in reader.elements()? {
        let mut filter_out = false;
        match &element {
            Element::Node { node } => {
                for tag in node.tags() {
                    if tag.k() == "natural" && tag.v() == "tree" {
                        filter_out = true;
                        break;
                    }
                }
            }
            Element::Way { way: _ } => {}
            Element::Relation { relation: _ } => {}
            Element::Sentinel => {
                filter_out = true;
            }
        }
        if !filter_out {
            writer.write_element(element)?;
        }
    }

    writer.close()?;

    log::info!("Finished pbf io pipeline, time: {}", stopwatch);
    Ok(())
}

Similar Software

  • libosmium - very fast and very mature with a Python wrapper.
  • osmosis - reference implementation for most if not all features.
  • osmpbf - very efficient *.osm.pbf reader written in Rust

License: MIT OR Apache-2.0

Dependencies

~15–31MB
~529K SLoC