[GIS] Python library or algorithm to generate arc geometry from three coordinate pairs

djangogeodjangogeometrypython

I'm trying to write code to generate fixtures for a django/geodjango project. I need to generate geometry for arcs (line string) given three lat/long pairs; start of arc, end of arc and centre of arc. I read through the geodjango documentation but this functionality doesn't seem to be available. Does anyone know if there is a Python library that provides this functionality? Or does anyone have an algorithm I can port to python? Thanks

Best Answer

Shapely PyQGIS, and GeoDjango use the same API based on the GEOS library:

With Shapely:

with a list of points:

from shapely.geometry import Point, LineString, mapping
pt1 = Point(0,0)
pt2 = Point(20,20)
pt3 = Point (50,50)
line = LineString([pt1,pt2,pt3])
#GeoJSON format
print mapping(line)
{'type': 'LineString', 'coordinates': ((0.0, 0.0), (20.0, 20.0), (50.0, 50.0))}

or with a list of coordinates:

line = LineString([(0, 0), (20,20),(50,50)])
#GeoJSON format
print mapping(line)
{'type': 'LineString', 'coordinates': ((0.0, 0.0), (20.0, 20.0), (50.0, 50.0))}

with PyQGIS:

with a list of points:

line = QgsGeometry.fromPolyline([QgsPoint(0,0),QgsPoint(20,20),QgsPoint(50,50)])
#GeoJSON format
print line.exportToGeoJSON()
{ "type": "LineString", "coordinates": [ [0, 0], [20, 20], [50, 50] ] }

with GeoDjango (look at Django: GEOS API)

with a list of coordinates:

from django.contrib.gis.geos import LineString
line = LineString((0, 0), (20, 20), (50, 50))
# GeoJSON format
print line.json
{ "type": "LineString", "coordinates": [ [ 0.0, 0.0 ], [ 20.0, 20.0 ], [ 50.0, 50.0 ] ] }

The resulting line is made up of two segments: (0,0) to (20,20) and (20,20) to (50,50)

enter image description here

Related Solutions

Python Geometry Operations – Best Pure Python Libraries

I don't know about GAE compatibility, but you migh try SymPy for pure math & geometry functions. Otherwise, the only other one I know about is Shapely but it has a libgeos_c dependency which might disqualify it based on your requirements.

UPDATE: Also SymPy is BSD licensed, which might be an important benefit to some devs.

[GIS] Updating Fusion Table layers from csv files using python

In order to upload csv to FusionTables, you'll need to make sure you authenticate first. I see you've included the dependencies for OAuth, but if you're just trying to have a script do a simple upload, it'd be easiest to use ClientLogin (and you don't need the dependencies). A skeleton of the code should look something like this (adapted from the sample code in the python library you referenced):

from authorization.clientlogin import ClientLogin
from sql.sqlbuilder import SQL
import ftclient
from fileimport.fileimporter import CSVImporter

token = ClientLogin().authorize(username, password) #Replace with your username and password (or prompt for them would be more secure)
ft_client = ftclient.ClientLoginFTClient(token) #ft_client needs to be used from here on out to make any calls to the FusionTables SQL API

#If you want to CREATE a table from the csv...
tableid = int(CSVImporter(ft_client).importFile("data.csv")) #Replace with your csv data source (or prompt for the location would be less hard-coded)
print tableid #This will print out the ID of your new table, but the table should also show up in Google Docs for the username you used to authenticate.

If you want to add rows to an existing FusionTable, that'll be trickier. It would probably take parsing out the CSV with csv.DictReader or something and then iterating over that and uploading those rows using something like...

#update row using dictionary of values
print ft_client.query(SQL().update(tableid, cols={'strings': 'mystring3'}, row_id=rowid))
print ft_client.query(SQL().select(tableid))

...for each row.