Programming in GIS 10'11'2005 Using geospatial data

1
Programming in GIS 10.11.2005Using geospatial
data

• Object-oriented programming
• Reading/writing rasters / vector data formats
• Manipulating geospatial data

2
Object-oriented programming

• Object
• Usually a thing created by a program during its
  execution
• An instance of a class, I.e. an object has an
  identity and a state, and its behaviour is
  defined by its class
• Class
• a declaration of data storage space
• a declaration of a set of methods

3
Jargon

• polymorphism
• same definitions (interfaces) are used for
  different types of data (objects)
• inheritance
• forming new classes using classes that have
  already been defined
• information hiding
• access to object data is only through methods,
  not all methods can be called from everywhere
• abstract classes
• used to define generic interfaces

4
Associations between classes

• inheritance
• subclass (child c., derived c.) superclass
  (parent c., base c.)
• multiple inheritance (e.g. not in Java but Java
  has interfaces)
• hierarchy, inheritance structure
• other associations
• containment

5
Notes

• Object-oriented analysis is an attempt to bring
  ideas from OOP to analysis
• OOP does not require an OO programming language
• for example GTK, an OO GUI library is written in
  C (not C)
• Pure OO languages and hybrid languages

6
OOP in Perl

• Pure Perl
• Object is typically a hash blessed to a class
  (package)
• Perl interface to a library
• Object is typically an opaque scalar, which can
  be used to invoke the methods

7
Geospatial data and GDAL
8
Geospatial data

• geospatial data is georeferenced points (places),
  paths (line strings), and areas (polygons) with
  associated attribute data
• in practice we have raster data and vector data
• we will use GDAL and OGR (which is a part of GDAL
  library) to access geospatial data

9
GDAL

• GDAL is a translator library for raster
  geospatial data formats, and OGR provides similar
  capability for simple features vector data
• GDALs homepage is www.gdal.org
• current GDAL version is 1.3.1 and it is the first
  version which has the NG scripting language
  support
• GDAL is the data access engine embedded into most
  of the free GIS programs

10
GDAL data model http//www.gdal.org/gdal_datamode
l.html
11
GDAL-Perl

• http//map.hut.fi/gdal-perl/

12
XML and beyond

• XML
• Parsing XML
• Schemas
• GML

13
XML

• structured document
• markup (metadata)
• content (data)
• programmers point of view
• nodes
• node has a name and type
• node may have attributes
• node may have a parent, siblings (previous and
  next) and children
• node may have a value

14
Parsing XML

• Two main methods
• parsing into a DOM (Document Object Model,
  w3.org/DOM)
• which is a tree of nodes
• event-based parsing (SAX, Simple API for XML,
  sax.sourceforge.net)
• which reports events to the application through
  callback functions

15
Validity of XML 1

• Validating input is often a pain
• is this a number?
• what about security?
• what is univeristy?
• well-formed XML
• markup is ok
• characters, tag names, attributes, elements
  close, etc
• contents is ok
• character set

16
Validity of XML 2

• The benefit of a known structure
• application knows what to look for
• DTD, XML Schema, RELAX NG, SOX, Schematron
• structure, context, data types, ...
• application (its programmer) knows what to look
  for and what to expect

17
Shared ontologies
copy is sent
XML
XML
creates
interprets
XSD
receiver
uses
sender
uses
18
Namespaces 1

• attribute xmlnssome_prefixsome_URI in the
  document element
• shows the document creators intent that elements
  whose names are of form some_prefixSome_name are
  created according to certain rules
• what the rules are, are assumed to be clear from
  the URI
• further hints may be given with an attribute
  xsischemaLocationsome_URI name_of_the_schema
  (for example)
• BTW Compare to namespaces in Perl

19
Namespaces 2

• if everything works out ok,
• the creator can use a validating parser to make
  sure she is sending a valid document and
• the receiver can check the received document is
  ok for further processing
• Remember the schema is in any case hard-coded
  into the applications using the schema

20
GML

• A namespace with and associated XML schema
  document
• the schema is written in XML Schema
• a draft ISO standard
• an OpenGIS Recommendation Paper
• now in version 3.1
• Note the usual xmlnsgmlhttp//www.opengis.net/
  gml does not tell which version is used
• May be used
• as one namespace in an XML document
• as a basis for higher level schemas

21
DB Access

• 1. Basic SQL and accessing RDBMSs
• 2. Geographic extensions to SQL

22
Architecture
Server program
Client program
DB library
Data
files in /var/lib/pgsql/...
postmaster
psql
DBI, DBDPg
myprog.pl
www.postgresql.org, man DBI, man DBDPg
23
SQL (1)

• Creating a table (CREATE TABLE...)
• Adding a column to a table (ALTER TABLE)
• Adding a row to a table (INSERT INTO...)
• Deleting a row from a table (DELETE FROM...)
• Updating the data in a table (UPDATE...)
• Querying a table (SELECT...)

24
SQL (2)

• Joins between tables
• Aggregate functions
• Views
• Keys
• Transactions
• Inheritance

25
DB access from Perl

• use DBI
• dbh DBI?connect(dbiPgdbnamepingis)
• die unless dbh
• sql select from students
• sth dbh?prepare(sql) or die
• sth?execute or die
• for (1..sth?rows)
• _at_d sth?fetchrow_array

26
Design issues

• Data model
• No redundancy
• Standards
• Performance
• Data interface
• Stability
• Performance
• Program code
• Division of labor with SQL
• Error checking

27
ODBC

• Open DataBase Connectivity
• Standard database access method developed by the
  SQL Access group in 1992
• Goal access any data from any application
• Requires a driver for the database system and a
  driver for the application (programming language)
• ODBC drivers exist for PostgreSQL and for DBI
  (DBDODBC)

28
Extensions to SQL for Geodata

• Simple Features Specification for SQL
• Open Geospatial Consortium 1999
• New column types
• Point, linestring, polygon, multipoint,
• New functions
• For spatial relations, spatial analysis

29
Basic usage

• Spatial columns are added to tables using OGC
  "AddGeometryColumn" function
• AddGeometryColumn(lttable_namegt, ltcolumn_namegt,
  ltsridgt, lttypegt, ltdimensiongt)
• srid the id of the spatial reference system
  used for this coordinate geometry, or -1
• type the geometry type
• dimension 1,2,3, or 4

postgis.org
30
Two metadata tables

• geometry_columns
• Metadata about spatial columns
• spatial_ref_sys
• Spatial reference systems

31
Inserting spatial data

• In INSERT Using function GeometryFromText(wkt,s
  rid)
• wkt Well Known Text

32
Retrieving spatial data

• in SELECT Using function AsText(geometry_column)

33
Design issues

• Indexes