The QGIS Python Console window can be accessed by using the Python Console item in the Plugins menu. When you select this command, the Python Console will appear in the lower-right corner of the QGIS window. Here's what the Python Console looks like when you first open it:

While the Python Console is an excellent tool for interacting with an existing QGIS project, we are going to use it to create a new project from scratch. Before we can do this, though, we'll need to download some geospatial data sources for our QGIS project.

We are going to need a suitable base map for our project, as well as some river and city information to display on top of this base map. Let's use the Natural Earth website to obtain the information we need. Go to http://naturalearthdata.com and click on the Downloads tab.

Firstly, we'll want to download a nice-looking base map for our project. To do this, select the Raster link under the Medium scale data, 1:50m section, choose the Natural Earth 1 dataset, and click on the Download small size link under the Natural Earth I with Shaded Relief and Water heading.

Next, we need an overlay, which will show lakes and rivers on top of our base map. To get this information, go back to the Downloads tab and select the Physical link under the Medium scale data, 1:50m section. The dataset you want is called Rivers, Lake Centerlines, so click on the Download rivers and lake centerlines link to obtain this file.

Finally, we'll want to highlight the cities on top of our base map. Go back to the Downloads page and select the Cultural link under the Medium scale data, 1:50m heading. At the bottom is a section labelled Urban Areas. Click on the Download urban areas link to download this file.

Once you've done all this, you should have the following three files:

Since these are ZIP archives, you will need to unzip these files and store them somewhere at a convenient location on your hard disk.

Now that we have our data, let's use the QGIS Python Console to import this data into a project. If you've already loaded some data into QGIS (for example, by following the tutorial in the QGIS User Guide), choose the New option from the Project menu to start again with a blank project. Then, type the following into the QGIS Python Console:

layer1 = iface.addRasterLayer("/path/to/NE1_50M_SR_W/NE1_50M_SR_W.tif", "basemap")

Make sure you replace /path/to/ with the full path to the NE1_50M_SR_W directory you downloaded. Assuming you typed the path correctly, the Natural Earth 1 base map should appear in the QGIS window:

As you can see, our base map is a bit small right now. You can use the various panning and zooming commands in the toolbar at the top of the window to make it bigger, but let's use Python to do the same thing:

iface.zoomFull()

This will expand the base map to fill the entire window.

Now that we have a base map, let's add our two vector layers to the project. To do this, type the following:

layer2 = iface.addVectorLayer("/path/to/ne_50m_urban_areas/ne_50m_urban_areas.shp", "urban", "ogr")

Once again, make sure you replace /path/to/ with the full path to the ne_50m_urban_areas directory you downloaded earlier. The urban areas shapefile will be loaded into the QGIS project and will appear as a series of colored areas on top of the base map. Let's zoom in to an area of California so that we can see what this looks like more clearly. To do this, type the following commands into the Python Console window:

iface.mapCanvas().setExtent(QgsRectangle(-125, 31, -113, 38))
iface.mapCanvas().refresh()

This will zoom in on the map in so that an area of California, including Los Angeles and the southern part of San Francisco, is now shown on the map:

Finally, let's add our river and lake data to our project. To do this, enter the following into the Python Console:

layer3 = iface.addVectorLayer("/path/to/ne_50m_rivers_lake_centerlines/ne_50m_rivers_lake_centerlines.shp", "water", "ogr")

If you look at the map, you'll see that the rivers and lakes are now visible. However, they are drawn in a default green color. Let's change this so that the water is now blue:

from PyQt4.QtGui import QColor
layer3.rendererV2().symbols()[0].setColor(QColor("#4040FF"))
iface.mapCanvas().refresh()

This code might be a bit confusing, but don't worry—we'll learn about renderers and symbols in Chapter 3, Learning the QGIS Python API.

Now that we are finished, you can save your project using the Save As... item in the Project menu. As you can see, it's quite possible to set up and customize your QGIS project using Python.

While the Python Console is a fantastic tool for interactive coding, it isn't all that useful if you want to use Python to extend the functionality of QGIS. This is where QGIS plugins come in; you can create (or download) a plugin that adds new features or changes the way QGIS works.

Because QGIS is written using the Qt framework, QGIS plugins make use of the Python bindings in Qt, which are called PyQt. We will download and install PyQt and the related tools when we start to build our own plugins in Chapter 4, Creating QGIS Plugins.

To get an idea of how a Python plugin works, let's take a look at the Zoom to Point plugin. As the name suggests, this plugin lets you zoom to display a given coordinate on the map. It's also written in Python, and is a convenient example for learning about plugins in general.

Before we can use it, we have to install this plugin. Choose the Manage and Install Plugins... item from the Plugins menu, and click on the Not Installed tab. You should see Zoom to Point listed near the bottom of the list of available plugins; click on this plugin, and then click on the Install Plugin button to download and install it.

Let's run this plugin to see how it works; with the project you created earlier still loaded, click on the Zoom to Point plugin's icon in the toolbar, which looks like this:

Try entering the longitude/latitude of your current location (if you don't know it, you might find http://itouchmap.com/latlong.html helpful). You should see the base map, urban areas, and waterways for your current location.

Now that we know what the plugin does, let's see how it works. The downloaded plugins are stored in a hidden directory named .qgis2 in your user or home directory. Go to this hidden directory using your favorite file manager (for Mac OS X, you can use the Go to Folder... item in the Finder's Go menu), and find the python/plugins subdirectory. This is where the Python plugins are stored.

You should see a directory named zoomtopoint (the full path to this directory will be ~/.qgis2/python/plugins/zoomtopoint). Inside this directory, you will find the various files that make up the Zoom to Point plugin:

Let's see what these various files do:

Open the zoomtopoint.py module in your favorite text editor. As you can see, this contains the main Python code for the plugin, in the form of a ZoomToPoint class. This class has the following basic structure:

class ZoomToPoint:
    def __init__(self, iface):
        self.iface = iface

    def initGui(self):
        ...

    def unload(self):
        ...

    def run(self):
        ...

If you open the __init__.py module, you'll see how this class is used to define the plugin's behavior:

def classFactory(iface): 
      from zoomtopoint import ZoomToPoint 
      return ZoomToPoint(iface) 

When the plugin is loaded, a parameter named iface is passed to the ClassFactory function. This parameter is an instance of QgsInterface, and provides access to the various parts of the running QGIS application. As you can see, the class factory creates a ZoomToPoint object, and passes the iface parameter to the initializer so that ZoomToPoint can make use of it.

Notice how ZoomToPoint.__init__(), in the Zoomtopoint.py module, stores a reference to the iface parameter in an instance variable, so that the other methods can refer to the QGIS interface using self.iface. For example:

def __init__(self, iface):
    self.iface = iface

def initGui(self):
    ...
    self.iface.addPluginToMenu("&Zoom to point...", self.action)

This allows the plugin to interact with and manipulate the QGIS user interface.

The four methods defined by the ZoomToPoint class are all quite straightforward:

Don't worry too much about all the details here; we'll look at the process of initializing and unloading a plugin in a later chapter. For now, take a closer look at the run() method. This method essentially looks like the following:

def run(self):
    dlg = ZoomToPointDialog()
    ...
    dlg.show()
    result = dlg.exec_()
    if result == 1:
        x = dlg.ui.xCoord.text()
        y = dlg.ui.yCoord.text()
        scale = dlg.ui.spinBoxScale.value()
        rect = QgsRectangle(float(x) – scale,
                            float(y) - scale,
                            float(x) + scale,
                            float(y) + scale)
        mc=self.iface.mapCanvas() 
        mc.setExtent(rect)
        mc.refresh()
        ...

We've excluded the code that remembers the values the user entered previously, and copies those values back into the dialog when the plugin is run. Looking at the previous code, the logic seems to be fairly straightforward and is explained as follows:

While this plugin is quite straightforward and the actual code doesn't do all that much, it is a useful example of what a Python plugin should look like, as well as the various files that are needed by a Python plugin. In particular, you should note that:

Plugins are a useful way of extending and customizing QGIS. We will return to the topic of QGIS plugins in Chapter 4, Creating QGIS Plugins, where we will create our own plugin from scratch.

The final way to work with Python and QGIS is to write a completely standalone Python program that imports the QGIS libraries and works with them directly. In many ways, this is an ideal way of writing your own custom mapping applications, because your program doesn't have to run within the existing QGIS user interface. There are, however, a few things you need to be aware of when you attempt to use Python and QGIS in this way:

None of this is too onerous, though it can trip you up the first time you attempt to access PyQGIS from your external Python code. Let's take a look at how we can avoid these traps when writing your own Python programs.

Firstly, to allow your program to access the PyQGIS libraries, you need to modify your Python path (and possibly some other environment variables) before you can import any of the QGIS packages. For MS Windows, you can do this by running the following in the command line:

SET OSGEO4W_ROOT=C:\OSGeo4W
SET QGIS_PREFIX=%OSGEO4W_ROOT%\apps\qgis
SET PATH=%PATH%;%QGIS_PREFIX%\bin
SET PYTHONPATH=%QGIS_PREFIX%\python;%PYTHONPATH%

If you are running Mac OS X, the following commands will set up the Python path for you:

export PYTHONPATH="$PYTHONPATH:/Applications/QGIS.app/Contents/Resources/python"
export DYLD_FRAMEWORK_PATH="/Applications/QGIS.app/Contents/Frameworks"
export QGIS_PREFIX="/Applications/QGIS.app/Contents/Resources"

For computers that run a version of Linux, you can use the following:

export PYTHONPATH="/path/to/qgis/build/output/python/"
export LD_LIBRARY_PATH="/path/to/qgis/build/output/lib/"
export QGIS_PREFIX="/path/to/qgis/build/output/"

If you have QGIS installed in a nonstandard location, you might need to modify these commands before they will work. To check if they have worked, start up the Python interpreter and enter the following command:

>>> import qgis

If everything goes well, you'll simply see the Python prompt:

>>> 

On the other hand, you might see the following error:

ImportError: No module named qgis

In this case, the PYTHONPATH variable has not been set up correctly, and you will have to check the commands you entered earlier to set this environment variable, and possibly modify it to allow for a nonstandard location of the QGIS libraries.

import qgis.core

ImportError: libqgis_core.so.1.5.0: cannot open shared object file: No such file or directory

With the path set, you can now import the various parts of the PyQGIS library that you want to use, for example:

from qgis.core import *

Now that we have access to the PyQGIS libraries, our next task is to initialize these libraries. As mentioned earlier, we have to tell PyQGIS where to find the various QGIS resources. We do this using the QgsApplication.setPrefixPath() function, like this:

import os
QgsApplication.setPrefixPath(os.environ['QGIS_PREFIX'], True)

This uses the QGIS_PREFIX environment variable we set earlier to tell QGIS where to find its resources. With this done, you can then initialize the PyQGIS library by making the following call:

QgsApplication.initQgis()

We can now use PyQGIS to do whatever we want in our application. When our program exits, we also need to inform the PyQGIS library that we are exiting:

QgsApplication.exitQgis()

Putting all this together, our minimal Python application looks like this:

import os
from qgis.core import *

QgsApplication.setPrefixPath(os.environ['QGIS_PREFIX'], True)
QgsApplication.initQgis()

# ...

QgsApplication.exitQgis()

Of course, this application doesn't do anything useful yet—it simply starts up and shuts down the PyQGIS libraries. So let's replace the "..." line with some useful code that displays a basic map widget. To do this, we need to define a QMainWindow subclass, which displays the map widget, and then create and use a QApplication object to display this window and handle the various user-interface events while the application is running.

Let's start by replacing the "..." line with the following code, which displays a map viewer and then runs the application's main event loop:

app = QApplication(sys.argv)

viewer = MapViewer("/path/to/shapefile.shp")
viewer.show()

app.exec_()

As you can see, a MapViewer instance (which we will define shortly) is created and displayed, and the QApplication object is run by calling the exec_() method. For simplicity, we pass the name of a shapefile to display within the map viewer.

Running this code will cause the map viewer to be displayed, and the application will run until the user closes the window or chooses the Quit command from the menu.

Now, let's define the MapViewer class. Here is what the class definition looks like:

class MapViewer(QMainWindow):
    def __init__(self, shapefile):
        QMainWindow.__init__(self)
        self.setWindowTitle("Map Viewer")

        canvas = QgsMapCanvas()
        canvas.useImageToRender(False)
        canvas.setCanvasColor(Qt.white)
        canvas.show()

        layer = QgsVectorLayer(shapefile, "layer1", "ogr")
        if not layer.isValid():
            raise IOError("Invalid shapefile")

        QgsMapLayerRegistry.instance().addMapLayer(layer)
        canvas.setExtent(layer.extent())
        canvas.setLayerSet([QgsMapCanvasLayer(layer)])

        layout = QVBoxLayout()
        layout.addWidget(canvas)

        contents = QWidget()
        contents.setLayout(layout)
        self.setCentralWidget(contents)

Don't worry too much about the details of this class; we basically just create a window and place a QgsMapCanvas object within it. We then create a map layer (an instance of QgsVectorLayer) and add it to the map canvas. Finally, we add the canvas to the window's contents.

Notice that QgsMapCanvas and QgsVectorLayer are both part of PyQGIS, while QMainWindow, QVBoxLayout, and QWidget are all PyQt classes. This application uses the PyQGIS classes within a PyQt application, mixing the classes from both sources. This is possible because QGIS is built using Qt, and the various PyQGIS classes are based on PyQt.

To turn the preceding code into a working application, all we need to do is add some more import statements to the top of the module:

import sys
from PyQt4.QtGui import *
from PyQt4.QtCore import Qt

If you run this application, the map viewer will be displayed, showing the contents of the shapefile referred to by the code. For example:

This application is still a bit ugly—you can see white space at the top and bottom this map because it doesn't take into account the aspect ratio of the map data. There's also no feature of zooming in or scrolling around the map. However, these can be added quite easily, and as you can see, it's not very difficult to create your own standalone mapping applications built on top of QGIS.