openlr-webtool-python

Overview

Implements the OpenLR python decoder protocol for a PostgreSQL/PostGIS DB containing the schema used by the TomTom OpenLR WebTool. That is, lines (roads) can be two-way, as opposed to one-way only, and a meta column links the map entries to map elements in a separate datastore. This module dynamically duplicates and/or reverses roads so that the decoder sees only one-way roads. It uses the flowdir column in the lines table to determine whether this is necessary.

This code is a wrapper around the TomTom Python OpenLR decoder implementation. For more information about OpenLR, please see the OpenLR Whitepaper.

DB schema

The PostgreSQL / PostGIS DB must contain tables representing lines (road segments) and nodes (intersections). The schema for these tables is the same as is recognized by the OpenLR WebTool Docker utility, and is described below.

Lines table schema

Column	Type	Comments
id	bigint	Unique primary key
meta	text	Arbitrary string
fow	smallint	Form of Way
frc	smallint	Functional Road Class
flowdir	smallint	Traffic flow direction
from_int	bigint	Start node ID
to_int	bigint	End node ID
len	double precision	Line length in meters
geom	geometry(LineString,4326)	PostGIS LineString

Nodes table schema

Column	Type	Comments
id	bigint	Unique primary key
geom	geometry(Point,4326)	PostGIS Point geometry

Notes

• The line table meta field can be an arbitrary string, and is used to correlate roads elements in the lines table to the corresponding elements in the target map
• flowdir column definition:
  • 1: traffic flow is bidirectional
  • 2: traffic flow is one way, from digitization end to digitization start ( S <- E ) )
  • 3: traffic flow is one way, from digitization start to digitization end ( S -> E ) )
• from_int and to_int are foreign key references to the id column of the nodes table

Sample usage:

from openlr import FRC, FOW
from webtool.map_databases import WebToolMapReader
from typing import Dict
from openlr_dereferencer import Config

my_tolerated_lfrc: Dict[FRC, FRC] = {
  FRC.FRC0: FRC.FRC1,
  FRC.FRC1: FRC.FRC2,
  FRC.FRC2: FRC.FRC3,
  FRC.FRC3: FRC.FRC4,
  FRC.FRC4: FRC.FRC5,
  FRC.FRC5: FRC.FRC6,
  FRC.FRC6: FRC.FRC7,
  FRC.FRC7: FRC.FRC7,
}

my_config = Config(
  tolerated_lfrc=my_tolerated_lfrc,
  max_bear_deviation=30,
  search_radius=30,
  geo_weight=0.66,
  frc_weight=0.17,
  fow_weight=0.17,
  bear_weight=0.0
)

rdr = WebToolMapReader(
  host="",
  port=5432,
  user="",
  password="",
  dbname="test_db",
  schema="sample",
  lines_table="roads",
  nodes_table="intersections",
  config=my_config
)

res = rdr.match("C7yP6xTT6QEWE/t6/+0BCA==")
# Out[11]: <openlr_dereferencer.decoding.line_location.LineLocation at 0x7fae582721c0>
res.lines
# Out[14]:
# [Line with id=3736240 of length 14.658514187337138,
# Line with id=2487940 of length 271.09294979343383,
# Line with id=1531179 of length 114.98001126157855,
# Line with id=2564465 of length 762.4772421777637]

len(res.coordinates())
# Out[16]: 22

res.n_off
# Out[20]: 45.791227508416654

res.p_off
# Out[21]: 0.0

Batch decoder utility

The batch_match.py utility can bulk-match a file containing binary openlr codes and write the results to a tab separated CSV file. See comments in the file for more details.

Dependencies:

- python >= 3.7
- openlr >= 1.0.1
- openlr-dereferencer >= 1.2.0
- pyproj >= 2.6.1
- psycopg2 >= 2.8.6
- param >= 1.12.0

Optional dependencies

- pytest >= 3.5.0 (to run unit tests)
- folium >= 0.12.1
- panel >= 0.12.6
- ipywidgets >= 7.7.0

Mac silicon notes

// libspatialite v5.1.0 is here: $ export DYLD_LIBRARY_PATH=/opt/homebrew/lib $ ls -la /opt/homebrew/lib/mod_spatialite.dylib lrwxr-xr-x 1 dave admin 54 30 Sep 13:28 /opt/homebrew/lib/mod_spatialite.dylib -> ../Cellar/libspatialite/5.1.0/lib/mod_spatialite.dylib

// and also here, built from source: $ ls -la /usr/local/lib/mod_spatialite.dylib lrwxr-xr-x 1 root wheel 22 1 Oct 12:58 /usr/local/lib/mod_spatialite.dylib -> mod_spatialite.8.dylib

zsh ❯ /usr/local/bin/sqlite3 /Users/dave/projects/python/openlr/data/france.sqlite SQLite version 3.40.0 2022-08-12 18:46:01 Enter ".help" for usage hints. sqlite> .headers on sqlite> .mode column sqlite> SELECT load_extension('/usr/local/lib/mod_spatialite');