HONG1992/geobases

Data services and visualization

GeoBases

Introduction

This project provides tools to play with geographical data. It also works with non-geographical data, except for map visualizations :).

There are embedded data sources in the project, but you can easily play with your own data in addition to the available ones. After data loading, you can:

• perform various types of queries (find this key, or find keys with this property)
• make fuzzy searches based on string distance (find things roughly named like this)
• make phonetic searches (find things sounding like this)
• make geographical searches (find things next to this place)
• get results on a map, or on a graph, or export it as csv data, or as a Python object

This is entirely written in Python. The core part is a Python package, but there is a command line tool as well! Get it with easy_install, then you can see where are airports with international in their name:

$ GeoBase --fuzzy international --map

You can perform all types of queries:

$ GeoBase --base cities --fuzzy "san francisko" # typo here :)

Of course, you can use your own data for map display:

$ cat coords.csv
p1,48.22,2.33
p2,49.33,2.24
$ cat coords.csv | GeoBase --map

And for every other thing as well:

$ cat edges.csv
A,B
A,C
D,A
$ cat edges.csv | GeoBase --graph

Administrate the data sources:

$ GeoBase --admin

We are currently gathering input from the community to define the next version features, so do not hesitate to open issues on the github page.

Documentation

Here are some useful links:

• the API documentation for the Python package
• the wiki pages for any question!
• the twitter account for the latest news

Installation

Prerequisites

These prerequisites are very standard packages which are often installed by default on Linux distributions. But make sure you have them anyway.

First you need to install setuptools (as root):

$ apt-get install python-setuptools    # for debian
$ yum install python-setuptools.noarch # for fedora

Then you need some basics compilation stuff to compile dependencies (also as root):

$ apt-get install python-dev g++    # for debian
$ yum install python-devel gcc-c++  # for fedora

From PyPI

You can install it from PyPI:

$ easy_install --user -U GeoBases

There is a development version also on PyPI (dev):

$ easy_install --user -U GeoBasesDev

From Github

You can clone the project from github:

$ git clone https://github.com/opentraveldata/geobases.git

Then install the package and its dependencies:

$ cd geobases
$ python setup.py install --user # for user space

Final steps

A script is put in ~/.local/bin, to be able to use it, put that in your ~/.bashrc or ~/.zshrc:

export PATH=$PATH:$HOME/.local/bin
export BACKGROUND_COLOR=black # or 'white', your call

If you use zsh and want to get awesome autocomplete for the main script, add this to your ~/.zshrc:

# Add custom completion scripts
fpath=(~/.zsh/completion $fpath)
autoload -U compinit
compinit

Python 3 and Pypy support

There is Python 3 and Pypy support, you can try it by changing branch before installation.

For Python 3, you have to install setuptools and python3-dev as prerequisites, then:

$ git checkout 3000
$ python3 setup.py install --user

You can also install the package for Python 3 from PyPI (3K):

$ easy_install-3.2 --user -U GeoBases3K

For Pypy, after pypy and pypy-dev installation:

$ git checkout pypy
$ sudo pypy setup.py install

You can also install the package for Pypy from PyPI (pypy):

$ easy_install --user -U GeoBasesPypy

Tests

Run the tests:

$ python test/test_GeoBases.py -v

Quickstart

>>> from GeoBases import GeoBase
>>> geo_o = GeoBase(data='ori_por', verbose=False)
>>> geo_a = GeoBase(data='airports', verbose=False)
>>> geo_t = GeoBase(data='stations', verbose=False)

You can provide other values for the data parameter. All data sources are documented in a single YAML file:

• data="ori_por" will load a local version of this file, this is the most complete source for airports, use it!
• data="airports" will use geonames as data source for airports
• data="stations" will use RFF data, from the open data website, as data source for french train stations
• data="stations_nls" will use NLS nomenclature as data source for french train stations
• data="stations_uic" will use UIC nomenclature as data source for french train stations
• data="countries" will load data on countries
• data="capitals" will load data on countries capitals
• data="continents" will load data on continents
• data="timezones" will load data on timezones
• data="languages" will load data on languages
• data="cities" will load data on cities, extracted from geonames
• data="currencies" will load data on currencies, extracted from wikipedia
• data="airlines" will load data on airlines, extracted from that file
• data="cabins" will load data on cabins
• data="locales" will load data on locales
• data="location_types" will load data on location types
• data="feature_classes" will load data on feature classes
• data="feature_codes" will load data on feature codes
• data="ori_por_non_iata" will load some non-iata data excluded from ori_por
• data="geonames_MC" will load MC data of geonames
• data="geonames_FR" will load FR data of geonames
• data="postal_codes_MC" will load MC postal codes data
• data="postal_codes_FR" will load FR postal codes data
• data="feed" will create an empty instance

All features are unaware of the underlying data, and are available as long as the headers are properly set in the configuration file, or from the Python API. For geographical features, you have to name the latitude field lat, and the longitude field lng.

Features

Information access

>>> geo_o.get('CDG', 'city_code')
'PAR'
>>> geo_o.get('BRU', 'name')
'Bruxelles National'
>>> geo_t.get('frnic', 'name')
'Nice-Ville'
>>> geo_t.get('fr_not_exist', 'name', default='NAME')
'NAME'

You can put your own data in a GeoBase class, either by loading your own file when creating the instance, or by creating an empty instance and using the set method.

Find things with properties

>>> conditions = [('city_code', 'PAR'), ('location_type', ('H',))]
>>> list(geo_o.findWith(conditions, mode='and'))
[(2, 'JDP'), (2, 'JPU')]
>>>
>>> conditions = [('city_code', 'PAR'), ('city_code', 'LON')]
>>> len(list(geo_o.findWith(conditions, mode='or')))
33

Distance computation

>>> geo_o.distance('CDG', 'NCE')
694.5162...

Find things near a geocode

>>> # Paris, airports <= 40km
>>> [k for _, k in sorted(geo_a.findNearPoint((48.84, 2.367), 40))]
['ORY', 'LBG', 'TNF', 'CDG']
>>>
>>> # Nice, stations <= 4km
>>> iterable = geo_t.findNearPoint((43.70, 7.26), 4)
>>> [geo_t.get(k, 'name') for _, k in iterable]
['Nice-Ville', 'Nice-St-Roch', 'Nice-Riquier']

Find things near another thing

>>> sorted(geo_a.findNearKey('ORY', 50)) # Orly, airports <= 50km
[(0.0, 'ORY'), (18.8..., 'TNF'), (27.8..., 'LBG'), (34.8..., 'CDG')]
>>>
>>> sorted(geo_t.findNearKey('frnic', 3)) # Nice station, <= 3km
[(0.0, 'frnic'), (2.2..., 'fr4342'), (2.3..., 'fr5737')]

Find closest things from a geocode

>>> list(geo_a.findClosestFromPoint((43.70, 7.26))) # Nice
[(5.82..., 'NCE')]
>>>
>>> list(geo_a.findClosestFromPoint((43.70, 7.26), N=3)) # Nice
[(5.82..., 'NCE'), (30.28..., 'CEQ'), (79.71..., 'ALL')]

Approximate name matching

>>> geo_t.fuzzyFind('Marseille Charles', 'name')[0]
(0.8..., 'frmsc')
>>> geo_a.fuzzyFind('paris de gaulle', 'name')[0]
(0.78..., 'CDG')

Map display

>>> geo_t.visualize()
* Added lines for duplicates linking, total 0
* Could not detect geocode support in join fields.
> Affecting category None     to color blue    | volume 3190
<BLANKLINE>
* Now you may use your browser to visualize:
./example_map.html ./example_table.html
<BLANKLINE>
* If you want to clean the temporary files:
rm ./example_map.json ...
<BLANKLINE>
(['map', 'table'], (['./example_map.html', './example_table.html'], ['./example_map.json', ...]))

Standalone script

Installation of the package will also deploy a standalone script named GeoBase:

$ GeoBase ORY CDG              # query on the keys ORY and CDG
$ GeoBase --closest CDG        # closest from CDG
$ GeoBase --near LIG           # near LIG
$ GeoBase --fuzzy marseille    # fuzzy search on 'marseille'
$ GeoBase --admin              # to administrate data sources
$ GeoBase --ask                # interactive learning mode
$ GeoBase --help               # your best friend

In the previous picture, you have an overview of the command line verbose display. Three displays are available for the command line tool:

• the verbose display
• the csv display with --quiet
• the map display with --map
• the graph display with --graph

With the verbose display, entries are displayed on each column, and the available fields on each line. Fields starting with __ like __field__ are special. This means they were added during data loading:

• __key__ is the field containing the id of the entry. Ids are defined with a list of fields in the configuration file.
• __dup__ is the field containing a list of duplicated keys. Indeed there is mechanism handling duplicated keys by default, which creates new keys if the key already exists in the GeoBase.
• __par__ is the field containing the parent key if the key is duplicated.
• __lno__ is the field containing the line number during loading.
• __gar__ is the field containing the data which was not loaded on the line (this can be because the line was not well formatted, or because there were missing headers).

More examples here, for example how to do a search on a field, like admin_code (B8 is french riviera):

$ GeoBase -E adm1_code -e B8

Same with csv output (customized with --show):

$ GeoBase -E adm1_code -e B8 --quiet --show __ref__ iata_code  name

Add a fuzzy search:

$ GeoBase -E adm1_code -e B8 --fuzzy sur mer

All heliports under 200 km from Paris:

$ GeoBase --near PAR -N 200 -E location_type@raw -e 'H'

50 train stations closest to a specific geocode:

$ GeoBase -E location_type@raw -e R --closest '48.853, 2.348' -C 50

Countries with non-empty postal code regex:

$ GeoBase -b countries -E postal_code_regex -e '' --reverse --quiet

Reading data input on stdin:

$ echo -e 'ORY^Orly\nCDG^Charles' | GeoBase

Display on a map:

$ GeoBase -b stations --map

Marker-less map for a specific GMT offset:

$ GeoBase -E gmt_offset -e 1.0 --map -M _ _ country_code  __none__

Display your data on a map:

$ cat coords.csv
p1,48.22,2.33
p2,49.33,2.24
$ cat coords.csv | GeoBase --map

Display your data on a graph:

$ cat edges.csv
A,B
A,C
D,A
$ cat edges.csv | GeoBase --graph

Packaging

The MANIFEST.in file is used to determine which files will be included in a source distribution. package_data directive in setup.py file is about which file will be exported in site-package after installation. So you really need both if you want to produce installable packages like rpms or zip which can be installed afterwards.

You will also find a Rakefile at the root of the project. This can be used to build and deploy the packages. Deployment can be done using webdav, and the Rakefile expects nd to be installed (this is a webdav client). To install nd, fetch the sources and compile them.

Virtualenv still has some bugs on 64 bits systems, if you are using such a system, you absolutely need to upgrade to the very last unreleased version of virtualenv, before executing rake:

$ pip uninstall virtualenv
$ pip install https://github.com/pypa/virtualenv/tarball/develop