secondo/bin/Scripts3/FixedGridPartitioningPD.sec

/*
//paragraph [10] title: [{\Large \bf ]  [}]
//[Contents] [\tableofcontents]
//[ue] [\"{u}]



[10] Fixed Grid Partitioning

[Contents] 

1 Overview

This script is for partitioning spatial data with a fixed grid.

It is a commented script suitable for viewing with ~pdview~. To be run with 

----    @%Scripts/FixedGridPartitioningPD.sec or
	@&Scripts/FixedGridPartitioningPD.sec
----

2 Preparations

Preparations: 

  * include the Distributed2Algebra in makefile.algebras and recompile secondo

  * get the desired shape-file from download.geofabrik.de

  * create and open a database

  * import relations ~Buildings~, ~Railways~ and ~Waterways~ using the script ~importGermanyOsm.psec~

  * restore Workers in relation ~Worker~

  * start the remoteMonitors for the workers

*/

#restore Worker from Workerfile

/*
Determine number of slots

*/

let SizeSlots = 64

/*
Determine number of cells in a row

*/

let SizeCells = 10


/*
3 Create a Box/Rectangle for all objects

Create a bounding box for all objects to be processed. 

HINT: In this example only relations ~Buildings~, ~Railways~ and ~Waterways~ are considered.

*/

let World = 
Buildings feed projectextend[; Box: bbox(.GeoData)]
Railways feed projectextend[; Box: bbox(.GeoData)] concat
Waterways feed projectextend[; Box: bbox(.GeoData)] concat
transformstream collect_box[TRUE]

/*
Extract the edges

*/

let Left = minD(World, 1)

let Right = maxD(World, 1)

#let Bottom = minD(World, 2)

#let Top = maxD(World, 2)

/*
4 Create grid

To determine the height and width of cells, just divide the difference of the left and right edges 
of the constructed rectangle by the desired number of cells in a row.

*/

let Dimension = (Right - Left) / SizeCells

/*
Create grid with the left bottom corner of the computed rectangle, the height and width of the cells, 
and the number of cells in a row.

*/

let grid = createCellGrid2D(Left, Right, Dimension, Dimension, SizeCells)

/*
5 Compute cells for the relations

*/

let BuildingsB4 = Buildings feed
extendstream[Cell: cellnumber(bbox(.GeoData), grid)] consume

let RailwaysB4 = Railways feed
extendstream[Cell: cellnumber(bbox(.GeoData), grid)] consume

let WaterwaysB4 = Waterways feed
extendstream[Cell: cellnumber(bbox(.GeoData), grid)] consume

/*
6 Distribute relations over workers

*/

let BuildingsB41 = BuildingsB4 feed 
dfdistribute2["BuildingsB41", Cell, SizeSlots, Worker]

let RailwaysB41 = RailwaysB4 feed 
dfdistribute2["RailwaysB41", Cell, SizeSlots, Worker]

let WaterwaysB41 = WaterwaysB4 feed 
dfdistribute2["WaterwaysB41", Cell, SizeSlots, Worker]

/*
7 Evaluation and histogram

This section is for evaluation of the grid.

*/

#let Eval = BuildingsB4 feed sortby[Cell] 
#groupby[Cell; Cnt:group count] addcounter[N,1] consume

#number of used cells
#query Eval feed max[N]

#number of duplicates
#let duplicates = BuildingsB4 count - Buildings count
#query duplicates

#smallest number of elements assigned to a cell
#query BuildingsB4 feed sortby[Cell] groupby[Cell; Cnt:group count] min[Cnt]
#query Eval feed min[Cnt]

#highest number of elements assigned to a cell
#query BuildingsB4 feed sortby[Cell] groupby[Cell; Cnt:group count] max[Cnt]
#query Eval feed max[Cnt]

#average number of elements assigned to a cell
#query BuildingsB4 feed sortby[Cell] groupby[Cell; Cnt:group count] avg[Cnt]
#query Eval feed avg[Cnt]

/*
Create a histogram over groupsizes

*/

#let myDia = realstream(0.0, 0.0 + Eval count, 1.0)
#set_histogram1d

#query Eval feed 
#projectextendstream[; P : fun( t : TUPLE) intstream(0,attr(t,Cnt)) 
#replaceElem[attr(t,N)  + 0.0] ] create_histogram1d[P, myDia]