After the initial ChatGPT hype in 2023 (when we saw the first LLM-backed QGIS plugins, e.g. QChatGPT and QGPT Agent), there has been a notable slump in new development. As far as I can tell, none of the early plugins are actively maintained anymore. They were nice tech demos but with limited utility.

However, in the last month, I saw two new approaches for combining LLMs with QGIS that I want to share in this post:

IntelliGeo plugin: generating PyQGIS scripts or graphical models

At the QGIS User Conference in Bratislava, I had the pleasure to attend the “Large Language Models and GIS” workshop presented by Gustavo Garcia and Zehao Lu from the the University of Twente. There, they presented the IntelliGeo Plugin which enables the automatic generation of PyQGIS scripts and graphical models.

The workshop was packed. After we installed all dependencies and the plugin, it was exciting to test the graphical model generation capabilities. During the workshop, we used OpenAI’s API but the readme also mentions support for Cohere.

I was surprised to learn that even simple graphical models are actually pretty large files. This makes it very challenging to generate and/or modify models because they take up a big part of the LLM’s context window. Therefore, I expect that the PyQGIS script generation will be easier to achieve. But, of course, model generation would be even more impressive and useful since models are easier to edit for most users than code.

Image source: https://github.com/MahdiFarnaghi/intelli_geo

ChatGeoAI: chat with PyQGIS

ChatGeoAI is an approach presented in Mansourian, A.; Oucheikh, R. (2024). ChatGeoAI: Enabling Geospatial Analysis for Public through Natural Language, with Large Language Models. ISPRS Int. J. Geo-Inf., 13, 348.

It uses a fine-tuned Llama 2 model in combination with spaCy for entity recognition and WorldKG ontology to write PyQGIS code that can perform a variety of different geospatial analysis tasks on OpenStreetMap data.

The paper is very interesting, describing the LLM fine-tuning, integration with QGIS, and evaluation of the generated code using different metrics. However, as far as I can tell, the tool is not publicly available and, therefore, cannot be tested.

Image source: https://www.mdpi.com/2220-9964/13/10/348

Are you aware of more examples that integrate QGIS with LLMs? Please share them in the comments below. I’d love to hear about them.

Adding basemaps to PyQGIS maps in Jupyter notebooks

By underdark

2023-11-17

QGIS

In the previous post, we investigated how to bring QGIS maps into Jupyter notebooks.

Today, we’ll take the next step and add basemaps to our maps. This is trickier than I would have expected. In particular, I was fighting with “invalid” OSM tile layers until I realized that my QGIS application instance somehow lacked the “WMS” provider.

In addition, getting basemaps to work also means that we have to take care of layer and project CRSes and on-the-fly reprojections. So let’s get to work:

from IPython.display import Image
from PyQt5.QtGui import QColor
from PyQt5.QtWidgets import QApplication
from qgis.core import QgsApplication, QgsVectorLayer, QgsProject, QgsRasterLayer, \
    QgsCoordinateReferenceSystem, QgsProviderRegistry, QgsSimpleMarkerSymbolLayerBase
from qgis.gui import QgsMapCanvas
app = QApplication([])
qgs = QgsApplication([], False)
qgs.setPrefixPath(r"C:\temp", True)  # setting a prefix path should enable the WMS provider
qgs.initQgis()
canvas = QgsMapCanvas()
project = QgsProject.instance()
map_crs = QgsCoordinateReferenceSystem('EPSG:3857')
canvas.setDestinationCrs(map_crs)
print("providers: ", QgsProviderRegistry.instance().providerList())

To add an OSM basemap, we use the xyz tiles option of the WMS provider:

urlWithParams = 'type=xyz&url=https://tile.openstreetmap.org/{z}/{x}/{y}.png&zmax=19&zmin=0&crs=EPSG3857'
rlayer = QgsRasterLayer(urlWithParams, 'OpenStreetMap', 'wms')  
print(rlayer.crs())
if rlayer.isValid():
    project.addMapLayer(rlayer)
else:
    print('invalid layer')
    print(rlayer.error().summary())

If there are issues with the WMS provider, rlayer.error().summary() should point them out.

With both the vector layer and the basemap ready, we can finally plot the map:

canvas.setExtent(rlayer.extent())
plot_layers([vlayer,rlayer])

Of course, we can get more creative and style our vector layers:

vlayer.renderer().symbol().setColor(QColor("yellow"))
vlayer.renderer().symbol().symbolLayer(0).setShape(QgsSimpleMarkerSymbolLayerBase.Star)
vlayer.renderer().symbol().symbolLayer(0).setSize(10)
plot_layers([vlayer,rlayer])

And to switch to other basemaps, we just need to update the URL accordingly, for example, to load Carto tiles instead:

urlWithParams = 'type=xyz&url=http://basemaps.cartocdn.com/dark_all/{z}/{x}/{y}.png&zmax=19&zmin=0&crs=EPSG3857'
rlayer2 = QgsRasterLayer(urlWithParams, 'Carto', 'wms')  
print(rlayer2.crs())
if rlayer2.isValid():
    project.addMapLayer(rlayer2)
else:
    print('invalid layer')
    print(rlayer2.error().summary()) 
    
plot_layers([vlayer,rlayer2])

You can find the whole notebook at: https://github.com/anitagraser/QGIS-resources/blob/master/qgis3/notebooks/basemaps.ipynb

Plotting large point CSV files quickly & interactively

By underdark

2020-12-06

Big Data, Visualization

Even with all their downsides, CSV files are still a common data exchange format – particularly between disciplines with different tech stacks. Indeed, “How to Specify Data Types of CSV Columns for Use in QGIS” (originally written in 2011) is still one of the most popular posts on this blog. QGIS continues to be quite handy for visualizing CSV file contents. However, there are times when it’s just not enough, particularly when the number of rows in the CSV is in the range of multiple million. The following example uses a 12 million point CSV:

To give you an idea of the waiting times in QGIS, I’ve run the following script which loads and renders the CSV:

from datetime import datetime

def get_time():
    t2 = datetime.now()
    print(t2)
    print(t2-t1)
    print('Done :)')

canvas = iface.mapCanvas()
canvas.mapCanvasRefreshed.connect(get_time)

print('Starting ...')

t0 = datetime.now()
print(t0)

print('Loading CSV ...')

uri = "file:///E:/Geodata/AISDK/raw_ais/aisdk_20170701.csv?type=csv&amp;xField=Longitude&amp;yField=Latitude&amp;crs=EPSG:4326&amp;"
vlayer = QgsVectorLayer(uri, "layer name you like", "delimitedtext")

t1 = datetime.now()
print(t1)
print(t1 - t0)

print('Rendering ...')

QgsProject.instance().addMapLayer(vlayer)

The script output shows that creating the vector layer takes 02:39 minutes and rendering it takes over 05:10 minutes:

Starting ...
2020-12-06 12:35:56.266002
Loading CSV ...
2020-12-06 12:38:35.565332
0:02:39.299330
Rendering ...
2020-12-06 12:43:45.637504
0:05:10.072172
Done :)

Rendered CSV file in QGIS

Panning and zooming around are no fun either since rendering takes so long. Changing from a single symbol renderer to, for example, a heatmap renderer does not improve the rendering times. So we need a different solutions when we want to efficiently explore large point CSV files.

The Pandas data analysis library is well-know for being a convenient tool for handling CSVs. However, it’s less clear how to use it as a replacement for desktop GIS for exploring large CSVs with point coordinates. My favorite solution so far uses hvPlot + HoloViews + Datashader to provide interactive Bokeh plots in Jupyter notebooks.

hvPlot provides a high-level plotting API built on HoloViews that provides a general and consistent API for plotting data in (Geo)Pandas, xarray, NetworkX, dask, and others. (Image source: https://hvplot.holoviz.org)

But first things first! Loading the CSV as a Pandas Dataframe takes 10.7 seconds. Pandas’ default plotting function (based on Matplotlib), however, takes around 13 seconds and only produces a static scatter plot.

Loading and plotting the CSV with Pandas

hvPlot to the rescue!

We only need two more steps to get faster and interactive map plots (plus background maps!): First, we need to reproject the lat/lon values. (There’s a warning here, most likely since some of the input lat/lon values are invalid.) Then, we replace plot() with hvplot() and voilà:

Plotting the CSV with Datashader

As you can see from the above GIF, the whole process barely takes 2 seconds and the resulting map plot is interactive and very responsive.

12 million points are far from the limit. As long as the Pandas DataFrame fits into memory, we are good and when the datasets get bigger than that, there are Dask DataFrames. But that’s a story for another day.

Movement data in GIS #30: synchronized trajectory animations with QGIS temporal controller

By underdark

2020-05-24

Movement data in GIS, QGIS

4 Comments

QGIS Temporal Controller is a powerful successor of TimeManager. Temporal Controller is a new core feature of the current development version and will be shipped with the 3.14 release. This post demonstrates two key advantages of this new temporal support:

Expression support for defining start and end timestamps
Integration into the PyQGIS API

These features come in very handy in many use cases. For example, they make it much easier to create animations from folders full of GPS tracks since the files can now be loaded and configured automatically:

Script & Temporal Controller in action (click for full resolution)

All tracks start at the same location but at different times. (Kudos for Andrew Fletcher for recordings these tracks and sharing them with me!) To create an animation that shows all tracks start simultaneously, we need to synchronize them. This synchronization can be achieved on-the-fly by subtracting the start time from all track timestamps using an expression:

directory = "E:/Google Drive/QGIS_Course/05_TimeManager/Example_Dayrides/"

def load_and_configure(filename):
    path = os.path.join(directory, filename)
    uri = 'file:///' + path + "?type=csv&amp;escape=&amp;useHeader=No&amp;detectTypes=yes"
    uri = uri + "&amp;crs=EPSG:4326&amp;xField=field_3&amp;yField=field_2"
    vlayer = QgsVectorLayer(uri, filename, "delimitedtext")
    QgsProject.instance().addMapLayer(vlayer)

    mode = QgsVectorLayerTemporalProperties.ModeFeatureDateTimeStartAndEndFromExpressions
    expression = """to_datetime(field_1) -
    make_interval(seconds:=minimum(epoch(to_datetime("field_1")))/1000)
    """

    tprops = vlayer.temporalProperties()
    tprops.setStartExpression(expression)
    tprops.setEndExpression(expression) # optional
    tprops.setMode(mode)
    tprops.setIsActive(True)

for filename in os.listdir(directory):
    if filename.endswith(".csv"):
        load_and_configure(filename)

The above script loads all CSV files from the given directory (field_1 is the timestamp, field_2 is y, and field_3 is x), enables sets the start and end expression as well as the corresponding temporal control mode and finally activates temporal rendering. The resulting config can be verified in the layer properties dialog:

To adapt this script to other datasets, it’s sufficient to change the file directory and revisit the layer uri definition as well as the field names referenced in the expression.

This post is part of a series. Read more about movement data in GIS.

Stand-alone PyQGIS scripts with OSGeo4W

By underdark

2019-03-03

GIS

12 Comments

PyQGIS scripts are great to automate spatial processing workflows. It’s easy to run these scripts inside QGIS but it can be even more convenient to run PyQGIS scripts without even having to launch QGIS. To create a so-called “stand-alone” PyQGIS script, there are a few things that need to be taken care of. The following steps show how to set up PyCharm for stand-alone PyQGIS development on Windows10 with OSGeo4W.

An essential first step is to ensure that all environment variables are set correctly. The most reliable approach is to go to C:\OSGeo4W64\bin (or wherever OSGeo4W is installed on your machine), make a copy of qgis-dev-g7.bat (or any other QGIS version that you have installed) and rename it to pycharm.bat:

Instead of launching QGIS, we want that pycharm.bat launches PyCharm. Therefore, we edit the final line in the .bat file to start pycharm64.exe:

In PyCharm itself, the main task to finish our setup is configuring the project interpreter:

First, we add a new “system interpreter” for Python 3.7 using the corresponding OSGeo4W Python installation.

To finish the interpreter config, we need to add two additional paths pointing to QGIS\python and QGIS\python\plugins:

That’s it! Now we can start developing our stand-alone PyQGIS script.

The following example shows the necessary steps, particularly:

Initializing QGIS
Initializing Processing
Running a Processing algorithm

import sys

from qgis.core import QgsApplication, QgsProcessingFeedback
from qgis.analysis import QgsNativeAlgorithms

QgsApplication.setPrefixPath(r'C:\OSGeo4W64\apps\qgis-dev', True)
qgs = QgsApplication([], False)
qgs.initQgis()

# Add the path to processing so we can import it next
sys.path.append(r'C:\OSGeo4W64\apps\qgis-dev\python\plugins')
# Imports usually should be at the top of a script but this unconventional 
# order is necessary here because QGIS has to be initialized first
import processing
from processing.core.Processing import Processing

Processing.initialize()
QgsApplication.processingRegistry().addProvider(QgsNativeAlgorithms())
feedback = QgsProcessingFeedback()

rivers = r'D:\Documents\Geodata\NaturalEarthData\Natural_Earth_quick_start\10m_physical\ne_10m_rivers_lake_centerlines.shp'
output = r'D:\Documents\Geodata\temp\danube3.shp'
expression = "name LIKE '%Danube%'"

danube = processing.run(
    'native:extractbyexpression',
    {'INPUT': rivers, 'EXPRESSION': expression, 'OUTPUT': output},
    feedback=feedback
    )['OUTPUT']

print(danube)

PyQGIS101 part 10 published!

By underdark

2019-01-05

GIS

PyQGIS 101: Introduction to QGIS Python programming for non-programmers has now reached the part 10 milestone!

Beyond the obligatory Hello world! example, the contents so far include:

If you’ve been thinking about learning Python programming, but never got around to actually start doing it, give PyQGIS101 a try.

I’d like to thank everyone who has already provided feedback to the exercises. Every comment is important to help me understand the pain points of learning Python for QGIS.

I recently read an article – unfortunately I forgot to bookmark it and cannot locate it anymore – that described the problems with learning to program very well: in the beginning, it’s rather slow going, you don’t know the right terminology and therefore don’t know what to google for when you run into issues. But there comes this point, when you finally get it, when the terminology becomes clearer, when you start thinking “that might work” and it actually does! I hope that PyQGIS101 will be a help along the way.

PyQGIS for non-programmers

By underdark

2018-06-06

QGIS

If you’re are following me on Twitter, you’ve certainly already read that I’m working on PyQGIS 101 a tutorial to help GIS users to get started with Python programming for QGIS.

https://twitter.com/underdarkGIS/status/1002287546464919552

I’ve often been asked to recommend Python tutorials for beginners and I’ve been surprised how difficult it can be to find an engaging tutorial for Python 3 that does not assume that the reader already knows all kinds of programming concepts.

It’s been a while since I started programming, but I do teach QGIS and Python programming for QGIS to university students and therefore have some ideas of which concepts are challenging. Nonetheless, it’s well possible that I overlook something that is not self explanatory. If you’re using PyQGIS 101 and find that some points could use further explanations, please leave a comment on the corresponding page.

PyQGIS 101 is a work in progress. I’d appreciate any feedback, particularly from beginners!

Porting Processing scripts to QGIS3

By underdark

2018-01-28

QGIS

8 Comments

I’ll start with some tech talk first. Feel free to jump to the usage example further down if you are here for the edge bundling plugin.

As you certainly know, QGIS 3 brings a lot of improvements and under-the-hood changes. One of those changes affects all Python scripts. They need to be updated to Python 3 and the new PyQGIS API. (See the official migration guide for details.)

To get ready for the big 3.0 release, I’ve started porting my Processing tools. The edge bundling script is my first candidate for porting to QGIS 3. I also wanted to use this opportunity to “upgrade” from a simple script to a plugin that integrates into Processing.

I used Alexander Bruy’s “prepair for Processing” plugin as a template but you can also find an example template in your Processing folder. (On my system, it is located in C:\OSGeo4W64\apps\qgis-dev\python\plugins\processing\algs\exampleprovider.)

Since I didn’t want to miss the advantages of a good IDE, I set up PyCharm as described by Heikki Vesanto. This will give you code completion for Python 3 and PyQGIS which is very helpful for refactoring and porting. (I also tried Eclipse with PyDev but if you don’t have a favorite IDE yet, I find PyCharm easier to install and configure.)

My PyCharm startup script qgis3_pycharm.bat is a copy of C:\OSGeo4W64\bin\python-qgis-dev.bat with the last line altered to start PyCharm:

@echo off
call "%~dp0\o4w_env.bat"
call qt5_env.bat
call py3_env.bat
@echo off<span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start"></span>
path %OSGEO4W_ROOT%\apps\qgis-dev\bin;%PATH%
set QGIS_PREFIX_PATH=%OSGEO4W_ROOT:\=/%/apps/qgis-dev
set GDAL_FILENAME_IS_UTF8=YES
rem Set VSI cache to be used as buffer, see #6448
set VSI_CACHE=TRUE
set VSI_CACHE_SIZE=1000000
set QT_PLUGIN_PATH=%OSGEO4W_ROOT%\apps\qgis-dev\qtplugins;%OSGEO4W_ROOT%\apps\qt5\plugins
set PYTHONPATH=%OSGEO4W_ROOT%\apps\qgis-dev\python;%PYTHONPATH%
start /d "C:\Program Files\JetBrains\PyCharm\bin\" pycharm64.exe

In PyCharm File | Settings, I configured the OSGeo4W Python 3.6 interpreter and added qgis-dev and the plugin folder to its path:

With this setup done, we can go back to the code.

I first resolved all occurrences of import * in my script to follow good coding practices. For example:

from qgis.core import *

became

from qgis.core import QgsFeature, QgsPoint, QgsVector, QgsGeometry, QgsField, QGis<span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start"></span>

in this PR.

I didn’t even run the 2to3 script that is provided to make porting from Python 2 to Python 3 easier. Since the edge bundling code is mostly Numpy, there were almost no changes necessary. The only head scratching moment was when Numpy refused to add a map() return value to an array. So (with the help of Stackoverflow of course) I added a work around to convert the map() return value to an array as well:

flocal_x = map(forcecalcx, subtr_x, subtr_y, distance)
electrostaticforces_x[e_idx, :] += np.array(list(flocal_x))

The biggest change related to Processing is that the VectorWriter has been replaced by a QgsFeatureSink. It’s defined as a parameter of the edgebundling QgsProcessingAlgorithm:

self.addParameter(QgsProcessingParameterFeatureSink(
   self.OUTPUT,
   self.tr("Bundled edges"),
   QgsProcessing.TypeVectorLine)
)

And when the algorithm is run, the sink is filled with the output features:

(sink, dest_id) = self.parameterAsSink(
   parameters, self.OUTPUT, context,
   source.fields(), source.wkbType(), source.sourceCrs()
)

# code that creates features

sink.addFeature(feat, QgsFeatureSink.FastInsert)

The ported plugin is available on Github.

The edge bundling plugin in action

I haven’t uploaded the plugin to the official plugin repository yet, but you can already download if from Github and give it a try:

For this example, I’m using taxi pick-up and drop-off data provided by the NYC Taxi & Limousine Commission. I downloaded the January 2017 green taxi data and extracted all trips for the 1st of January. Then I created origin-destination (OD) lines using the QGIS virtual layer feature:

To get an interesting subset of the data, I extracted only those OD flows that cross the East River and have a count of at least 5 taxis:

Now the data is ready for bundling.

If you have installed the edge bundling plugin, the force-directed edge bundling algorithm should be available in the Processing toolbox. The UI of the edge bundling algorithm looks pretty much the same as it did for the QGIS 2 Processing script:

Since this is a small dataset with only 148 OD flows, the edge bundling processes is pretty quick and we can explore the results:

Beyond this core edge bundling algorithm, the repository also contains two more scripts that still need to be ported. They include dependencies on sklearn, so it will be interesting to see how straightforward it is to convert them.