Related to this work - I've been wondering whether we should drop all the doxygen "Not available in Python bindings" notes from the c++ docs and headers. To me this information seems redundant now - either it's available in the bindings and included in the Python docs, or it's not and it won't be in the Python docs. This information was really just a hack while we only had c++ docs available for Python devs.

see QgsCoordinateReferenceSystem::CrsType

See https://qgis.org/pyqgis/master/core/Layout/QgsLayoutManager.html (I'm not sure Methods and signals should be embedded in the note frame)

I know there's a feature in QGIS to eliminate sliver polygons and it can also be implemented in standalone script. It seems like a blackbox to get the results directly. Is there a way in standalone script to identify (extract) sliver polygons? Also, can I use other parameters (e.g. the neighboring polygon that shares the same attribute value such as block ID) not the ones given in QGIS to assign slivers?

Thanks!

Since this commit, qgis/QGIS@64387a2 it's better to use from qgis.processing import run. But these functions are not visible in the pyqgis documentation.

Not sure whether I should open it here or in the source repository

There should probably be code sample converted here

as we can see in the source .h file (missing in the sip.in one)

https://qgis.org/pyqgis/master/core/QgsProviderMetadata.html incorrectly only contains:

All the real api docs for this class are missing

For example in https://qgis.org/pyqgis/master/core/QgsLogger.html

The asterisks are not present in the API documentation https://qgis.org/api/classQgsLogger.html

The clone-Method of QgsLayerTreeGroup doesn't seem to work as expected. According to the documentation for QGIS 3.10 (and also current master) the children are cloned as well. However, when cloning a group with of map layers, the layers of the clone just seems to be references to the originals. Any changes applied to the supposedly cloned layer are also applied to the original.

From the documentation:

clone(self) → QgsLayerTreeGroup
Returns a clone of the group. The children are cloned too.
Return type
QgsLayerTreeGroup

https://qgis.org/pyqgis/master/core/QgsLayerTreeGroup.html?#qgis.core.QgsLayerTreeGroup.clone

Code Sample:

root = iface.layerTreeCanvasBridge().rootGroup()
group1 = QgsLayerTreeGroup('group 1')
layer1 = QgsVectorLayer("path" , "layer name 1")
group1.addLayer(layer1)
root.addChildNode(group1)
group2 = group1.clone()
group2.setName("group 2")
layer2 = group2.children()[0]
layer2.setName("layer name 2")
root.addChildNode(group2)

This code will result into the following layer-structure. In this example, layer2 is renamed, but renaming (as way as any other changes) also affects layer1

• group 1
  • layer name 2
• group 2
  • layer name 2

QGIS-Version | 3.10.0-A Coruña
BS-Version | Ubuntu 19.10

When you are browsing https://qgis.org/pyqgis/master and you find an issue, it might be really really convenient if there was a "fix me" link in the footer that would bring you to the class page (.h?) in qgis/QGIS repository. That would make contribution easier to proceed (we have this in qgis/QGIS-Documentation and it greatly saves time, avoiding waste of time with search in the repository)

I wonder if the rendering of the brewer_string is as expected

Paragraphs are wrongly formed (split, indentation) in the upper part of the QgsLayoutItemMapOverview class
Also, the "see also" frame is merged with the Methods one.

https://qgis.org/pyqgis/master/core/Layout/QgsLayoutItemMapOverview.html

Hi, i was reading the README.md file, and in Prerequisites tells to install requirements.txt but this file is missing.

See eg qgis.core.QgsAbstractCacheIndex.getCacheIterator - the featureIterator Param docs get incorrectly split and half is formatted as a normal text block

Hi, i have been studing PyQgis docs and i found something.
Some descriptions are bigger than cell, that needs a scroll to see all description.

Maybe, can it set at CSS or html from the theme to make a line break.

I have tested in chrome and firefox.

Dear developers,

I am developing some code in the QGIS Python Console. As I cannot set any breakpoints and print cannot be used due to multithreading I tried to use the QgsLogger class as described in the documentation. Unfortunately I couldn't get any output. Also there are no examples or comments regarding this topic in any public GIS related user group.
I even tried to set some environment variables in order to force the logging which didn't work out.

This is the code I use:

from qgis.core import QgsLogger

logname= "c:/temp/test.log"

Set environment variables

os.environ['QGIS_LOG_FILE'] = logname
os.environ['QGIS_DEBUG'] = '3'

print ("Logfile: ",QgsLogger.logFile())
QgsLogger.critical("QgsLoggerTest1")
QgsLogger.debug("QgsLoggerTest2")

QgsLogger.logMessageToFile("Mymessage")
print ("Debuglevel: ",QgsLogger.debugLevel())

The output in the console is as follows:

Logfile:
Debuglevel: 0

No file was created in c:/temp/test.log nor was the output written to any tab in the log messages window.

In the regarding QGIS documentation https://docs.qgis.org/3.10/en/docs/pyqgis_developer_cookbook/communicating.html#logging there is a link to https://qgis.org/pyqgis/3.10/core/QgsLogger.html#qgis.core.QgsLogger where the first sentence clearly states "QgsLogger is a class to print debug/warning/error messages to the console" which does not happen.

I'm running the following system:
QGIS version | 3.14.1-Pi | QGIS code revision | de08d6b
Compiled against Qt | 5.11.2 | Running against Qt | 5.11.2
Compiled against GDAL/OGR | 3.0.4 | Running against GDAL/OGR | 3.0.4
Compiled against GEOS | 3.8.1-CAPI-1.13.3 | Running against GEOS | 3.8.1-CAPI-1.13.3
Compiled against SQLite | 3.29.0 | Running against SQLite | 3.29.0
PostgreSQL Client Version | 11.5 | SpatiaLite Version | 4.3.0
QWT Version | 6.1.3 | QScintilla2 Version | 2.10.8
Compiled against PROJ | 6.3.2 | Running against PROJ | Rel. 6.3.2, May 1st, 2020
OS Version | Windows 10 (10.0)
Active python plugins | GroupStats; nominatim; QuickOSM; db_manager; MetaSearch; processing

Please note that I had opened an issue at the QGIS bug tracking site as well as I didn't know pyqgis has it's own site:
qgis/QGIS#38188

I had also set up a bug report for documentation at qgis/QGIS-Documentation#5995

As soon as we're sure what's the reason for the failure and that this site here is the correct one where to handle the bug I suggest to close it where obsolete.

Currently, all tables displaying Methods and Signals have a max width of 800px which often forces the user to scroll horizontally to read all the content of a cell. I think it would be more user friendly (and elegant) if the cells had instead an overflow wrap. The max width could stay at 800px, but could also very well be unrestricted, the width being responsive anyway. I tried to find myself in the repo where the change could be made but I'm at a complete loss! I suppose the template with all the CSS/HTML is elsewhere...

Go to https://qgis.org/pyqgis/master/core/Layout/QgsLayoutExporter.html#qgis.core.QgsLayoutExporter.exportToPdf
The exportToPdf has more than one way to call it but only the one in the grayed area (next to the name) is obvious, and when looking for a function you may miss the other alternative. This is also true for exportToImage, exportToSVG, print... (at least in this class).
Moreover, the paragraph showing the second alternative does not look sexy to read and you want to skip it (I unfortunately did it (!) for hours until I find in the github repo what I was looking for, which is not the way we want people to proceed.).

There are some \section and \subsection tags(?) used in the code to structure the doc but they are not rendered in the PyQGIS doc, making things less readable. See eg https://qgis.org/pyqgis/master/core/QgsVectorLayer.html (having sections in this long description would certainly help to quickly identify the provider you are interested in).

Reading the docs at https://docs.qgis.org/3.22/en/docs/pyqgis_developer_cookbook/geometry.html#access-to-geometry, trying to identify the QgsWkbTypes.GeometryType and QgsWkbTypes.Type classes (actually identify the enumeration of each) seems impossible (at least for the beginner I'm).

such as https://github.com/qgis/QGIS/blob/master/python/core/auto_generated/qgshstoreutils.sip.in

Hi all,
I'm trying to convert a 2D+Z vector layer to a 3D objet (using VTK) ...
So far I managed to get the geometry right :

but i'm looking to color the 3D object using the iface.mapCanvas() rendering ...

all I could find is to request the QgsSymbol for each feature and collect its color/label/whatever.
maybe there is a better way, but i could not find it

so when i try to call layer.renderer().symbolForFeature(feature, context) I got an unexpected crash ... linked to SIP and python binded methods

# -*- coding: utf-8 -*-
from qgis.core import QgsRenderContext
from qgis.core import QgsVectorLayer
from qgis.core import QgsWkbTypes
from qgis.utils import iface
from vtk import vtkCellArray
from vtk import vtkIntArray
from vtk import vtkPoints
from vtk import vtkPolyData
from vtk.util.numpy_support import numpy_to_vtk

__all__ = ["QgsVectorLayerToPolyData"]


def set_cells(poly: vtkPolyData, cells: vtkCellArray, dim: int):
    if dim == 0:
        poly.SetVerts(cells)
    elif dim == 1:
        poly.SetLines(cells)
    elif dim == 2:
        poly.SetPolys(cells)
    else:
        raise ValueError(f"Invalid data dimension: {dim}")


def QgsVectorLayerToPolyData(layer: QgsVectorLayer, build_lut=False):
    _type = layer.wkbType()
    _dim = QgsWkbTypes.wkbDimensions(_type)

    assert layer.isSpatial()  # i.e. have some geometries (might be empty)
    assert QgsWkbTypes.hasZ(_type)  # i.e. geometries are defined in R^3

    painter = layer.renderer().clone()
    context = QgsRenderContext.fromMapSettings(iface.mapCanvas().mapSettings())

    # Python side mesh primitives
    n, vertices = 0, []
    offsets, connectivity = [0], []
    colors = []
    for f in layer.getFeatures():
        symbol = painter.symbolForFeature(f, context)  # HERE IS THE PROBLEM
        colors.append(int(symbol.color().name().lstrip("#"), base = 16))
        for p in f.geometry().constParts():
            for v in p.vertices():
                vertices.append((v.x(), v.y(), v.z()))
                connectivity.append(n)
                n += 1
            if _dim == 2:  # OGR/WKT polygons duplicates first/last vertex
                del vertices[-1]
                del connectivity[-1]
                n -= 1
            offsets.append(n)

    # VTK/C++ side mesh
    polydata = vtkPolyData()
    points = vtkPoints()
    points.SetData(numpy_to_vtk(vertices, 1))
    cells = vtkCellArray()
    cells.SetData(numpy_to_vtk(offsets, 1), numpy_to_vtk(connectivity, 1))
    scalars = vtkIntArray()
    scalars.SetNumberOfComponents(1)
    scalars.SetName("Legend")
    scalars.SetData(numpy_to_vtk(colors, 1))
    # vtkPolyData generation
    polydata.SetPoints(points)
    set_cells(polydata, cells, _dim)
    polydata.GetCellData().AddArray(scalars)
    return polydata

the full crash repport:

Crash ID: cb1c4f2b11cbd392a38c522a5ed1e194e060e4a0


Stack Trace


QgsExpression::evaluate :
QgsCategorizedSymbolRenderer::valueForFeature :
QgsCategorizedSymbolRenderer::originalSymbolForFeature :
PyInit__core :
PyMethodDef_RawFastCallKeywords :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyEval_EvalCodeWithName :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyEval_EvalCodeWithName :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyEval_EvalCodeWithName :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyEval_EvalCodeWithName :
PyFunction_FastCallDict :
PyTuple_New :
PyObject_FastCallKeywords :
PyObject_FastCallKeywords :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyEval_EvalCodeWithName :
PyFunction_FastCallDict :
PyTuple_New :
PyObject_FastCallKeywords :
PyObject_FastCallKeywords :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyMethodDef_RawFastCallKeywords :
PyEval_EvalFrameDefault :
PyFunction_FastCallDict :
PyMethodDef_RawFastCallDict :
PyObject_Call :
PyInit_QtCore :
PyInit_QtCore :
PyInit_QtCore :
PyInit_QtCore :
PyInit_QtCore :
QMetaObject::activate :
QAction::activate :
QAbstractButton::click :
QAbstractButton::mouseReleaseEvent :
QToolButton::mouseReleaseEvent :
QWidget::event :
QApplicationPrivate::notify_helper :
QApplication::notify :
QgsApplication::notify :
QCoreApplication::notifyInternal2 :
QApplicationPrivate::sendMouseEvent :
QSizePolicy::QSizePolicy :
QSizePolicy::QSizePolicy :
QApplicationPrivate::notify_helper :
QApplication::notify :
QgsApplication::notify :
QCoreApplication::notifyInternal2 :
QGuiApplicationPrivate::processMouseEvent :
QWindowSystemInterface::sendWindowSystemEvents :
QEventDispatcherWin32::processEvents :
UserCallWinProcCheckWow :
DispatchMessageWorker :
QEventDispatcherWin32::processEvents :
qt_plugin_query_metadata :
QEventLoop::exec :
QCoreApplication::exec :
main :
BaseThreadInitThunk :
RtlUserThreadStart :




QGIS Info
QGIS Version: 3.16.2-Hannover
QGIS code revision: f1660f9da5
Compiled against Qt: 5.11.2
Running against Qt: 5.11.2
Compiled against GDAL: 3.1.4
Running against GDAL: 3.1.4



System Info
CPU Type: x86_64
Kernel Type: winnt
Kernel Version: 10.0.17763

Environment:

• Ubuntu 18.04.4
• QGIS 3.12.1-București
• python 3.6.9

No matter which provider:algorithm algorithmHelp is trying to display details for it seems to encounter the same issue when running from terminal outside of QGIS desktop.

Create grid (native:creategrid)
...
CRS: Grid CRS

        Parameter type: QgsProcessingParameterCrs

        Accepted data types:
Traceback (most recent call last):
    ...
    processing.algorithmHelp("native:creategrid")
  File "/usr/share/qgis/python/plugins/processing/tools/general.py", line 75, in algorithmHelp
    print('\n'.join(opts))
UnicodeEncodeError: 'ascii' codec can't encode character '\u2026' in position 94: ordinal not in range(128)

https://qgis.org/pyqgis/master/core/QgsCoordinateReferenceSystem.html

There are several problems with the way how the docs are generated now:

• constructors are listed before the class description, making the text slightly confusing
• the first line of text "Constructs an invalid CRS object" belongs to the default constructor which is not listed there
• constructor description should have the same style as ordinary methods: the constructor definition should be highlighted and the description should be indented.

Description

In QGIS Python API documentation pages there are same inconsistencies among various versions:

for master, 3.10 and 3.4:

for 3.8, 3.6, 3.2, 3.0:

The title says all. This is the same issue as #31 but I fail to find my way in the repo structure to propose a PR so...

See e.g http://python.qgis.org/master/core/other/QgsXmlUtils.html

the documentation for qgis.core.QgsWkbTypes shows different IDs depending on the wkbType. Link: Documention
Unfortunately the id for polygons is the same as for unknownGeometry which is 3:

Looking into why https://qgis.org/pyqgis/3.4/core/Raster/QgsRasterHistogram.html has such poor docs, it's because the member variable documentation isn't transferred to sip docstrings.

This isn't currently supported by sip, but I think it should be possible. See https://stackoverflow.com/a/40488936/1861260

leads to overwhelmed doc
see QgsAdvancedDigitizingCanvasItem for instance

3.6 docs are available, but on index page (https://qgis.github.io/pyqgis/ or https://qgis.org/pyqgis) the link to it is not yet added

I think I should be able to fix it myself, or do a PR but fail to find the page...

Maybe also add a link to https://github.com/qgis/pyqgis on the bottom of the index page?

E.g. see QgsFeature::attribute( int fieldIdx ). In this case the MethodCode requires the sip definition to use SIP_PYOBJECT as the return type, but this prevents the docs from showing the actual return type and only "object" is used.

I developing QGIS module and I have a question, how to capture QgsMapCanvas window with OpenCV for shoving in different frame or different application?
OpenCV will have data from map view window, in my case raster data for interactive processing and show processed picture in other frame.

See eg http://python.qgis.org/master/core/Layout/QgsLayoutItemLabel.html

The class docs start with:

"A layout item subclass for text labels.

New in version 3.0: Methods"

This makes it seem like the following methods were added in version 3 - when it's actually the class which was new in 3.0, and the ":" should be removed and replaced with a new paragraph instead.

Seems to happen with all classes marked with a since annotation

I'm not sure if this is a right place to ask the question. Since I asked my question in QT forum, but no one could answer it.

Is there a way to send signals among different plugins directly to make them communicate?

Thanks in advance.

and param documentation is discarded

E.g in http://python.qgis.org/master/core/Layout/QgsLayoutItemLabel.html#qgis.core.QgsLayoutItemLabel.adjustPointForReferencePosition, there's the line

Bases: qgis._core.QgsLayoutItem

But because this isn't a link, you can't directly go to the base class docs without searching for the class manually.

• QgsExpression
  https://qgis.org/pyqgis/master/core/QgsExpression.html#module-QgsExpression
  

• QgsFeature
  https://qgis.org/pyqgis/master/core/QgsFeature.html#qgis.core.QgsFeature.setGeometry
  

• QgsProjectDirtyBlocker
  https://qgis.org/pyqgis/3.18/core/QgsProjectDirtyBlocker.html#module-QgsProjectDirtyBlocker
  

See e.g. https://qgis.org/pyqgis/master/core/QgsTriangle.html

The "Example" headings from methods like isDegenerate() do not show

https://lists.osgeo.org/pipermail/qgis-developer/2018-February/052148.html

The conversion to python code seems to not render correctly (https://qgis.org/pyqgis/master/core/QgsActionScopeRegistry.html)

Silly issue maybe, but I was looking for:

https://qgis.org/pyqgis/master/core/QgsNetworkAccessManager.html?highlight=qgsnetwork

And then you see:

And I was thinking... uh. spaces between Class names? I tried to copy the class name by double clicking, and then also you only copy QgsNetwork???
Then I removed the ?highlight=qgsnetwork and then double clicking the classname rightfully selects the whole classname.

So in theme.css there is:

.rst-content .highlighted {
background: #F1C40F;
display: inline-block;
font-weight: bold;
padding: 0 6px;
}
removing the padding: 0 6px; line it's (I think better).

I cloned the repo for a PR, but cannot inf the theme.css? It is probably in the theme used in the build?

Even non - QObject derived classes are showing an empty "signals" section - but this has no meaning and is misleading

one improvement which would be really nice would be to make the return types hyperlinks
i think to do this we need this plugin
https://github.com/agronholm/sphinx-autodoc-typehints

at least, that's my understanding based on https://stackoverflow.com/questions/13920165/sphinx-method-return-types-with-links
although possible this could work which would just be sipify magic: https://stackoverflow.com/a/19400374/1861260

The QgsVectorLayer description has sections and subsections but the current rendering does not make this structure obvious nor ease the description reading.