Welcome to Stetl

Stetl, Streaming ETL, is an open source (GNU GPL) toolkit for the transformation (ETL) of geospatial data. Stetl is based on existing ETL tools like GDAL/OGR and XSLT. Stetl processing is driven from a configuration (.ini) file. Stetl is written in Python and in particular suited for processing GML.

This is the documentation of the Stetl toolkit. The code is on GitHub: https://github.com/geopython/stetl. Since July 2016 the project is a proud member of the GeoPython GitHub organization.

See an introductory Stetl presentation on Slideshare.

This is document version 2.1 generated on Jan 09, 2023.

Contents:

Intro

Stetl, streaming ETL, pronounced “staedl”, is a lightweight ETL-framework for the conversion of rich (such as GML) geospatial data. Stetl is Open Source (GNU GPL v3).

Read a 5-minute introduction here: http://www.slideshare.net/justb4/5-minute-intro-to-setl and a longer presentation here: http://www.slideshare.net/justb4/geospatial-etl-with-stetl-geopython-2016. Plus a presentation of Stetl for use in INSPIRE transformation: http://www.slideshare.net/justb4/2-stetlinspiretransformv1 with even a video recording: https://www.youtube.com/watch?v=vjdpYBm4AaM

Stetl originated in the INSPIRE-FOSS project and was originally created by Just van den Broecke. Subsequently, Stetl evolved into a wider use transforming Dutch GML-based datasets such as Top10NL, IMGEO/BGT (Large Scale Topography) and IMKAD/BRK (Kadastral Data). Therefore Stetl now has a repository of its own at GitHub.

Stetl basically glues together existing parsing and transformation tools like GDAL/OGR (ogr2ogr) and XSLT. By using native tools like libxml2 and libxslt (via Python lxml) Stetl is speed-optimized.

Stetl has (currently) no GUI. There are powerful Open Source ETL tools like GeoKettle and Talend Geospatial with a GUI. Check these out. But some of us would like to stay close to the commandline, be Pythonic and reuse existing tools ‘close to the iron’.

So why and when to use Stetl:

  • when ogr2ogr or XSLT alone cannot do the job
  • when having to deal with complex GML as source or destination
  • when you want to use simple command-line tooling or (Python) program integrations
  • when you need speed
  • when you are a Pythonista

Stetl is in particularly useful for INSPIRE-related transformations and other complex GML-related ETL.

Stetl was presented at FOSS4G 2013 in Nottingham, see http://2013.foss4g.org/conf/programme/presentations/156 and the slides: http://www.slideshare.net/justb4/stetl-foss4g20131024v1

Installation

Stetl up to and including version 1.3 only runs with Python 2 (2.7+). Starting with Stetl v2.0 only Python 3 (3.6+) will be supported. You may want to read Upgrade to Python 3 when upgrading from a Stetl pre-v2 version.

Easiest is to first install the Stetl-dependencies (see below) and then install and maintain Stetl on your system as a Python package (pip is preferred).

(sudo) pip install stetl
or
easy_install stetl

Alternatively you can download Stetl from Github: by cloning (preferred) or downloading: https://github.com/geopython/stetl/archive/master.zip and then install locally

(sudo) python setup.py install

Try the examples first. This should work on Linuxes and Mac OSX.

Windows installation may be more involved depending on your local Python setup. Platform-specific installations below.

You may also want to download the complete .tar.gz distro from PyPi: https://pypi.python.org/pypi/Stetl . This includes the examples and tests.

Docker

Since version 1.0.9 Stetl also can be installed and run via Docker. See Install with Docker below.

Debian/Ubuntu

Thanks to Bas Couwenberg, work is performed to provide Stetl as Debian packages on both Debian and Ubuntu, see details: https://packages.debian.org/search?keywords=stetl (Debian) and https://launchpad.net/ubuntu/+source/python-stetl (Ubuntu, Xenial and later). Stetl is split into 2 packages python-stetl, the Python framework and stetl the command line utility. NB the versions of these packages may be older than when installing Stetl via pip from PyPi or directly from GitHub. Always check this first.

Dependencies

Stetl depends on the following Python packages:

  • GDAL (>=v2.4) bindings for Python
  • psycopg2 (PostgreSQL client)
  • lxml >=4.4.2
  • Jinja2 templating
  • Deprecated

GDAL Python binding requires the native GDAL/OGR libs and tools to be installed.

lxml http://lxml.de/installation.html requires the native (C) libraries:

  • libxslt (required by lxml)
  • libxml2 with Python bindings (required by lxml)

When using the Jinja2 templating filter, Jinja2TemplatingFilter, see http://jinja.pocoo.org:

  • Python Jinja2 package

Deprecated is used to indicated deprecated functions and classes.

Platform-specific guidelines for dependencies follow next.

Linux

For Debian-based distro’s like Ubuntu and Debian itself, most packages should be able to be installed via apt-get.

Tip: to get latest versions of GDAL and other Open Source geospatial software, best is to add the UbuntuGIS Repository. Below a setup that works in Ubuntu 16.04 Xenial using Debian/Ubuntu packages. In some cases you may choose to install the same packages via pip3 to have more recent versions like for lxml.

More Linux Tips. See also:

  • Python dependencies:

    apt-get install python3-setuptools
apt-get install python3-dev
apt-get install python3-pip
pip3 install --upgrade pip

  • libxml2/libxslt libs are usually already installed. Together with Python lxml, the total install for lxml is:

    apt-get install python3-libxml2
apt-get install python-libxslt1
apt-get install libxml2-dev libxslt1-dev lib32z1-dev
apt-get install python3-lxml

  • GDAL (http://gdal.org) version 2+ with Python bindings:

    # Add UbuntuGIS repo to get latest GDAL, at least v2 on Ubuntu 16.04, Xenial.
add-apt-repository ppa:ubuntugis/ubuntugis-unstable
apt-get update
apt-get install gdal-bin
gdalinfo --version
# should show something like: GDAL 2.4.0, released 2019/03/04

apt-get install python3-gdal

  • the PostgreSQL client library for Python psycopg2:

    apt-get install python3-psycopg2

  • for Jinja2:

    apt-get install python3-jinja2

Mac OSX

Dependencies can best be installed via Homebrew.

Tip: sometimes installing GDAL Python bindings can be tricky as the installed GDAL binaries must be compatible. To install the right version you may use:

pip install GDAL==`gdalinfo --version | cut -d' ' -f2 | cut -d',' -f1`

Windows

Best is to install GDAL and python using the OSGeo4W Installer from http://trac.osgeo.org/osgeo4w.

  • Download and run the OSGeo4W Installer
  • Choose Advanced Install
  • On the Select Packages page expand Commandline_Utilities and Select from the list gdal and python
  • (psycopg2??)
  • Install easy_install to allow you to install lxml
  • Download the ez_setup.py script
  • Open the OSGeo4W Shell (Start > Programs > OSGeo4W > OSGeo4W > OSGeo4W Shell)
  • Change to the folder that you downloaded ez_setup.py to (if you downloaded to C:Temp then run cd C:Temp)
  • Install easy_install by running python ez_setup.py
  • To install lxml with easy_install run easy_install lxml

Only Psycopg2 needs explicit installation. Many install via: http://www.stickpeople.com/projects/python/win-psycopg. Once the above has been installed you should have everything required to run Stetl.

Alternatively you may use Portable GIS. Still you will need to manually install psycopg2. See http://www.archaeogeek.com/portable-gis.html for details.

Test Installation

If you installed via Python ‘pip’ you can check if you run the latest version

stetl -h

You should get meaningful output like

2013-09-16 18:25:12,093 util INFO running with lxml.etree, good!
2013-09-16 18:25:12,100 util INFO running with cStringIO, fabulous!
2013-09-16 18:25:12,122 main INFO Stetl version = 1.0.3
usage: stetl [-h] -c  CONFIG_FILE [-s  CONFIG_SECTION] [-a  CONFIG_ARGS]

Especially check the Stetl version number. You can also use the -v or –version option for stetl.

Try running the examples when running with a downloaded distro.

cd examples/basics
./runall.sh

Look for any error messages in your output.

Run Unit Tests

You can run unit tests to completely verify your installation. First install some extra packages:

pip install -r requirements-dev.txt

Then run the tests using nose2.

nose2

Install with Docker

The fastest way to use Stetl is via Docker. The Stetl Docker Image is lightweight, compressed just over 100MB, based on a Debian “slim” Docker Image.

Your environment needs to be setup to use Docker and probably you want to use some tooling like Vagrant. The author uses a combination of VirtualBox with Ubuntu and Vagrant on Mac OSX to run Docker, but this is a bit out of scope here.

Assuming you have a working Docker environment, there are two ways to install Stetl with Docker:

When rebuilding you can add build arguments for your environment, defaults:

ARG TIMEZONE="Europe/Amsterdam"
ARG LOCALE="en_US.UTF-8"
ARG ADD_PYTHON_DEB_PACKAGES=""
ARG ADD_PYTHON_PIP_PACKAGES=""

For example building with extra Python packages, building your local Docker Image:

docker build --build-arg ADD_PYTHON_DEB_PACKAGES="python-requests python-tz" -t geopython/stetl:local .
docker build --build-arg ADD_PYTHON_PIP_PACKAGES="scikit-learn==0.18 influxdb" -t geopython/stetl:local .

Or you may extend the Stetl Dockerfile with your own Dockerfile.

For running Stetl using Docker see Using Docker.

Upgrade to Python 3

Stetl development started in Python 2. With PEP 373 the EOL of python 2.7 was announced and python 2 will not be officialy supported after 2020. Stetl was therefore upgraded to Python 3.

Python 3

Work started early 2019 to upgrade Stetl from Python 2 to Python 3. The last version of Stetl that supports Python 2 is version 1.3. This version might receive quick fixes and updates, but users are encouraged to upgrade to Stetl version 2 or higher and thus use Python 3.

For the full discussion on the Python 2 to Python 3 migration: see the conversation in pull request #81 within the GitHub repository.

Important changes for developers

Python 2 and 3 are very similar, but there are a couple of important changes that developers need to keep in mind and are worth mentioning:

  • Stetl 2 supports Python 3.6 (3.4 and 3.5 were dropped) and higher (so with support of f strings)
  • Python 3 uses Unicode strings, meaning encoding/decoding is a bit different
  • stringIO and cstringIO were moved around
  • slight syntax change on calling next() for iterators
  • update on import statements
  • differences in urllib to make http-calls (although issue 80 might change it to the requests library).

Important changes for users

The specification of the Stetl tool chain uses a configuration file. You can use the Inputs, Filters, and Outputs that are provided by Stetl, or write your own. If you use Stetl Components in your configuration, you must specify the stetl. package prefix in the class specification. For example before Stetl version 2 the input XML file was specified as

[input_xml_file]
class = inputs.fileinput.XmlFileInput
file_path = input/cities.xml

for Stetl version 2 this is changed to

[input_xml_file]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input/cities.xml

Note the extra stetl. part in the class specification.

Background

The text below gives some introduction to ETL, the rationale why Stetl was developed and where and how it attempts to fit in.

Problem

Data conversion combined with model and coordinate transformation from a source to a target datastore (files, databases) is a recurring task in almost every geospatial project. This process is often referred to as ETL (Extract Transform Load). Source and/or target geo-data formats are increasingly encoded as GML (Geography Markup Language), either as flat records, so called Simple Features, but more and more using domain-specific, object oriented OGC/ISO GML Application Schema’s.

GML Application Schema’s are for example heavily used within the INSPIRE Data Harmonization effort in Europe. Many National Mapping and Cadastral Agencies (NMCAs) use GML-encoded datasets as their bulk format for download and exchange and via Web Feature Services (WFSs). As geospatial professionals we are often confronted with ETL-tasks involving (complex) GML or worse: “GML-lookalikes”, which are often XML Schemas embedded with GML-namespaced elements.

Luckily, in many cases GDAL/OGR, the Swiss Army Knife for geo-data conversion, can do the job. If “ogr2ogr” sounds like gibberish to you, check out http://gdal.org ! But when complex, some say rich, GML Application Schemas are involved, data conversion can be a daunting task when GDAL/OGR alone is not sufficient. Firstly, often complex data model transformations have to be applied.

In addition we may be confronted with the bulkiness of GML:

  • Megabyte/Gigabyte-files.
  • Deeply nested elements where the nuggets, the actual attribute values, reside.
  • Trees of .zip files and possibly more nasty surprises once we have unboxed a GML-delivery.
  • High resource consumption in memory and CPU and long processing hours, up to complete machine-lockup, can be the the side-effects of naive GML-processing.

Existing (partial) solutions

Within the FOSS4G world we can resort to high level, GUI-based, ETL-tools such as GeoKettle, Humboldt tools and Talend GeoSpatial. These are very powerful tools by themselves, check them out as well. Some of us, like the author, like to stay closer to GDAL/OGR and XSLT for model transforms, some command line tools and a bit of Python scripting, but without having to write a complete, ad-hoc ETL-program each time. This is the space where Stetl tries to fit in, so read on.

We already have great FOSS tools for XML/GML parsing, data-conversion and model-transformation like GDAL/OGR (ogr2ogr!), XSLT (Extensible Stylesheet Language Transformations, for transforming XML) and native XML-parsing libraries like libxml2. Each individual tool/library is extremely powerful and performant by itself. But we would like to combine of these tools. Take for example flat, national adres data in a PostGIS database that we need to transform to multiple INSPIRE Application Schema GML files. Each individual FOSS tool can handle part of the ETL: ogr2ogr for converting from PostGIS (including coordinate tranformation) into to simple feature GML, XSLT (xsltproc/libxslt) to transform the resulting flat GML to rich INSPIRE GML. But with millions of addresses we cannot simply use a single GML memory datastructure (DOM) or single intermediate GML-file.

Stetl: Python, streaming and configuration

Add Python and a configuration convention to this equation and we have Stetl: Streaming ETL. Stetl is a lightweight, geospatial ETL (Extract Transform Load) framework written in Python. ETL-processing with Stetl is driven from a configuration file. Within a Stetl configuration file a chain of ETL-processing modules is declared through which the data flows (“streams”). A module may be an input, filter or output module. Modules have input and output data types declared such that only compatible modules can be connected. However, Stetl does not define a grand internal data structure to which all data is mapped as many ETL-tools do. Data formats are kept close to the external tools that Stetl uses.

Stetl comes with pre-defined modules for:

  • GML-parsing
  • XSLT processing
  • XSD Validation
  • PostGIS/OGR input and output
  • GML-splitting
  • … and many more.

Stetl calls on the above tools like OGR, libxslt and libxml2 via their native interfaces. Stetl is even more speed-optimized as no intermediate file-storage is used: we use other means such as native string buffers. For example large XML/GML-files can be split into manageable documents and streamed into an XSLT filter module. Stetl-modules are off course extensible and can be user-defined. Reusable ETL-configurations invoked through parameterized commandline scripts can be defined without programming.

Stetl evolved from and is used within the INSPIRE-FOSS project (http://inspire-foss.org). Here for example, Dutch national addresses (BAG) were transformed into INSPIRE Addresses GML (files and database). Special Stetl integration modules are available to extract and publish data from/to a deegree WFS and deegree “Blobstore-database”. The combination Stetl/deegree is an ideal setup for INSPIRE deployments.

Other Dutch national datasets like Top10NL and BGT (Dutch topo vector datasets) have been completely and successfully transformed. Work is in progress to use Stetl as the basis for NLExtract (http://nlextract.nl), a project that provides ETL tools for Dutch open geo-datasets. Stetl development is now (april 2013) in an initial phase and takes place in GitHub. The current version is workable but we hope to present a v1.0 at FOSS4G with more documentation and as a standard Python Package via PyPi. The main link is: http://stetl.org (now links to GitHub). To get started find some basic examples here: https://github.com/geopython/stetl/tree/master/examples/basics.

Using Stetl

This section explains how to use Stetl for your ETL. It assumes Stetl is installed and you are able to run the examples. It may be useful to study some of the examples, especially the core ones found in the examples/basics directory. These examples start numbering from 1, building up more complex ETL cases like (INSPIRE) transformation using Jinja2 Templating.

In addition there are example cases like the Dutch Topo map (Top10NL) ETL in the examples/top10nl directory .

The core concepts of Stetl remain pretty simple: an input resource like a file or a database table is mapped to an output resource (also a file, a database, a remote HTTP server etc) via one or more filters. The input, filters and output are connected in a pipeline called a processing chain or Chain. This is a bit similar to a current in electrical engineering: an input flows through several filters, that each modify the current. In our case the current is (geospatial) data. Stetl design follows the so-called Pipes and Filters Architectural Pattern.

Stetl Config

Stetl components (Inputs, Filters, Outputs) and their interconnection (the Pipeline/Chain) are specified in a Stetl config file. The file format follows the Python .ini file-format.

To illustrate, let’s look at the example 2_xslt. This example takes the input file input/cities.xml and transforms this file to a valid GML file called output/gmlcities.gml. The Stetl config file looks as follows.

[etl]
chains = input_xml_file|transformer_xslt|output_file

[input_xml_file]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input/cities.xml

[transformer_xslt]
class = stetl.filters.xsltfilter.XsltFilter
script = cities2gml.xsl

[output_file]
class = stetl.outputs.fileoutput.FileOutput
file_path = output/gmlcities.gml

Most of the sections in this ini-file specify a Stetl component: an Input, Filter or Output component. Each component is specified by its (Python) class and per-component specific parameters. For example [input_xml_file] uses the class stetl.inputs.fileinput.XmlFileInput reading and parsing the file input/cities.xml specified by the file_path property. [transformer_xslt] is a Filter that applies XSLT with the script file cities2gml.xsl that is in the same directory. The [output_file] component specifies the output, in this case a file.

These components are coupled in a Stetl Chain using the special .ini section [etl]. That section specifies one or more processing chains. Each Chain is specified by the names of the component sections, their interconnection using a the Unix pipe symbol “|”.

So the above Chain is input_xml_file|transformer_xslt|output_file. The names of the component sections like [input_xml_file] are arbitrary.

Note: since v1.1.0 a datastream can be split (see below) to multiple Outputs using () like :

[etl]
chains = input_xml_file|transformer_xslt|(output_gml_file)(output_wfs)

Or multiple Input streams can be combined/merged like:

[etl]
chains = (input_http_api_1) (input_http_api_2) | data_transformer | output_db

It is even possible to have both Splitting and Merging together with filtering:

[etl]
chains = (input_http_api_1 | cleaner_filter) (input_http_api_2) | data_transformer | (output_db) (output_file)

Note: since version 2 of stetl it is required that the call to stetl components actually start with stetl. This is not necessary when you write your own components (see example 7)

Configuring Components

Most Stetl Components, i.e. inputs, filters, outputs, have properties that can be configured within their respective [section] in the config file. But what are the possible properties, values and defaults? This is documented within each Component class using the @Config decorator much similar to the standard Python @property, only with some more intelligence for type conversions, defaults, required presence and documentation. It is loosely based on https://wiki.python.org/moin/PythonDecoratorLibrary#Cached_Properties and Bruce Eckel’s http://www.artima.com/weblogs/viewpost.jsp?thread=240845 with a fix/hack for Sphinx documentation.

See for example the stetl.inputs.fileinput.FileInput documentation.

For class authors: this information is added via the Python Decorators much similar to @property. The stetl.component.Config is used to define read-only properties for each Component instance. For example,

class FileInput(Input):
    """
    Abstract base class for specific FileInputs, use derived classes.
    """

    # Start attribute config meta
    # Applying Decorator pattern with the Config class to provide
    # read-only config values from the configured properties.

    @Config(ptype=str, default=None, required=False)
    def file_path(self):
        """
        Path to file or files or URLs: can be a dir or files or URLs
        or even multiple, comma separated. For URLs only JSON is supported now.
        """
        pass

    @Config(ptype=str, default='*.[gxGX][mM][lL]', required=False)
    def filename_pattern(self):
        """
        Filename pattern according to Python ``glob.glob`` for example:
        '\\*.[gxGX][mM][lL]'
        """
        pass

    @Config(ptype=bool, default=False, required=False)
    def depth_search(self):
        """
        Should we recurse into sub-directories to find files?
        """
        pass

    # End attribute config meta

    def __init__(self, configdict, section, produces):
        Input.__init__(self, configdict, section, produces)

        # Create the list of files to be used as input
        self.file_list = Util.make_file_list(self.file_path, None, self.filename_pattern, self.depth_search)

This defines three configurable properties for the class FileInput. Each @Config has three parameters: ptype, the Python type (str, list, dict, bool, int), default (default value if not present) and required (if property in mandatory or optional).

Within the config one can set specific config values like,

[input_xml_file]
class = inputs.fileinput.XmlFileInput
file_path = input/cities.xml

This automagically assigns file_path to self.file_path without any custom code and assigns the default value to filename_pattern. Automatic checks are performed: if file_path (required=True) is present, if its type is string. In some cases type conversions may be applied e.g. when type is dict or list. It is guarded that the value is not overwritten and the docstrings will appear in the auto-generated documentation, each entry prepended with a CONFIG tag.

Running Stetl

The above ETL spec can be found in the file etl.cfg. Now Stetl can be run, simply by typing

stetl -c etl.cfg

Stetl will parse etl.cfg, create all Components by their class name and link them in a Chain and execute that Chain. Of course this example is very trivial, as we could just call XSLT without Stetl. But it becomes interesting with more complex transformations.

Suppose we want to convert the resulting GML to an ESRI Shapefile. As we cannot use GDAL ogr2ogr on the input file, we need to combine XSLT and ogr2ogr. See example 3_shape. Now we replace the output by using outputs.ogroutput.Ogr2OgrOutput, which can execute any ogr2ogr command, converting whatever it gets as input from the previous Filter in the Chain.

[etl]
chains = input_xml_file|transformer_xslt|output_ogr_shape

[input_xml_file]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input/cities.xml

[transformer_xslt]
class = stetl.filters.xsltfilter.XsltFilter
script = cities2gml.xsl

# The ogr2ogr command-line. May be split over multiple lines for readability.
# Backslashes not required in that case.
[output_ogr_shape]
class = stetl.outputs.ogroutput.Ogr2OgrOutput
temp_file = temp/gmlcities.gml
ogr2ogr_cmd = ogr2ogr
        -overwrite
        -f "ESRI Shapefile"
        -a_srs epsg:4326
        output/gmlcities.shp
        temp/gmlcities.gml

Using Docker

The most convenient way to run Stetl is via Docker. See the installation instructions at Install with Docker. A full example can be viewed in the Smart Emission project: https://github.com/Geonovum/smartemission/tree/master/etl.

In the simplest case you run a Stetl Docker container from your own built image or the Dockerhub-provided one, geopython/stetl:<version> stetl as follows (latest version):

sudo docker run -v <host dir>:<container dir> -w <work dir> geopython/stetl:latest stetl <any Stetl arguments>

For example within the current directory you may have an etl.cfg Stetl file:

WORK_DIR=`pwd`
sudo docker run -v ${WORK_DIR}:${WORK_DIR} -w ${WORK_DIR} geopython/stetl:latest stetl -c etl.cfg

# or leaner
sudo docker run --rm -v $(pwd):/work -w /work geopython/stetl:latest stetl -c etl.cfg

A more advanced setup would be (network-)linking to a PostGIS Docker image like kartoza/postgis:

# First run Postgis, remains running,
sudo docker run --name postgis -d -t kartoza/postgis:9.4-2.1

# Then later run Stetl
STETL_ARGS="-c etl.cfg -a local.args"
WORK_DIR="`pwd`"

sudo docker run --name stetl --link postgis:postgis -v ${WORK_DIR}:${WORK_DIR} -w ${WORK_DIR} geopython/stetl:latest stetl ${STETL_ARGS}

The last example is used within the SmartEmission project. Also with more detail and keeping all dynamic data (like PostGIS DB), your Stetl config and results, and logs within the host. For PostGIS see: https://github.com/Geonovum/smartemission/tree/master/services/postgis and Stetl see: https://github.com/Geonovum/smartemission/tree/master/etl.

Even better is to use docker-compose.

Stetl Integration

Note: one can also run Stetl via its main ETL class: stetl.etl.ETL. This may be useful for integrations in for example Python programs or even OGC WPS servers (planned).

Reusable Stetl Configs

What we saw in the last example is that it is hard to reuse this etl.cfg when we have for example a different input file or want to map to different output files. For this Stetl supports config parameter substitution.

Dynamic or secret (e.g. database credentials) parameters in etl.cfg are declared symbolically and substituted at runtime via the commandline or the OS environment.

A variable is declared between curly brackets like {out_xml}. See example 6_cmdargs.

[etl]
chains = input_xml_file|transformer_xslt|output_file

[input_xml_file]
class = stetl.inputs.fileinput.XmlFileInput
file_path = {in_xml}

[transformer_xslt]
class = stetl.filters.xsltfilter.XsltFilter
script = {in_xsl}

[output_file]
class = stetl.outputs.fileoutput.FileOutput
file_path = {out_xml}

Note the symbolic input, xsl and output files. We can now perform the ETL using the stetl -a option in two basic ways. One, passing the arguments on the commandline, like

stetl -c etl.cfg -a "in_xml=input/cities.xml in_xsl=cities2gml.xsl out_xml=output/gmlcities.gml"

Two, passing the arguments in a properties file, here called etl.args (the name of the suffix .args is not significant, could be .env as well).

stetl -c etl.cfg -a etl.args

Where the content of the etl.args properties file is:

# Arguments in properties file
in_xml=input/cities.xml
in_xsl=cities2gml.xsl
out_xml=output/gmlcities.gml

It is also possible to specify multiple -a arguments. This provides for situations where a default.args contains all default arguments and a my.args or explicit -a settings that override the default values in default.args. Overriding is determined by the order of the -a arguments. Examples:

stetl -c etl.cfg -a default.args -a my.args
stetl -c etl.cfg -a default.args -a "db_user=docker db_password=pass"
stetl -c etl.cfg -a default.args -a db_user=docker -a db_password=pass

It is also possible to pass these key/value pairs via OS Environment variables. This is especially handy in Docker-based deployments like Docker Compose and Kubernetes. In this case the variable names need to be prepended with STETL_ or stetl_ as to not mix-up with other non-related OS-env vars. A mixture of commandline args (file) and environment vars is possible. The rule is that OS Environment variables always override/overrule arguments specified with -a option(s).

For example, the above args could also be passed as follows:

export stetl_in_xml="input/cities.xml"
export stetl_in_xsl="cities2gml.xsl"
export stetl_out_xml="output/gmlcities.gml"
stetl -c etl.cfg

or only override the input file name in_xml from etl.args:

export stetl_in_xml="input/cities2.xml"
stetl -c etl.cfg -a etl.args

or even with multiple -a args:

export stetl_in_xml="input/cities2.xml"
stetl -c etl.cfg -a etl.args -a my.args

This makes an ETL chain highly reusable. A very elaborate Stetl config with parameter substitution can be seen in the Top10NL ETL.

Connection Compatibility

During ETL Chain processing Components typically pass data to a next stetl.component.Component . A stetl.filter.Filter Component both consumes and produces data, Inputs produce data and Outputs only consume data.

Data and status flows as stetl.packet.Packet objects between the Components. The type of the data in these Packets needs to be compatible only between two coupled Components. Stetl does not define one unifying data structure, but leaves this to the Components themselves.

Each Component provides the type of data it consumes (Filters, Outputs) and/or produces (Inputs, Filters). This is indicated in its class definition using the consumes and produces object constructor parameters. Some Components can produce and/or consume multiple data types, like a single stream of records or a record array. In those cases the produces or consumes parameter can be a list (array) of data types.

During Chain construction Stetl will check for compatible formats when connecting Components. If one of the formats is a list of formats, the actual format is determined by:

  1. explicit setting: the actual input_format and/or output_format is set in the Component .ini configuration
  2. no setting provided: the first format in the list is taken as default

Stetl will only check if these input and output-formats for connecting Components are compatible when constructing a Chain.

The following data types are currently symbolically defined in the stetl.packet.FORMAT class:

  • any - ‘catch-all’ type, may be any of the types below.
  • etree_doc - a complete in-memory XML DOM structure using the lxml etree
  • etree_element - each Packet contains a single DOM Element (usually a Feature) in lxml etree format
  • etree_feature_array - each Packet contains an array of DOM Elements (usually Features) in lxml etree format
  • geojson_feature - as struct but following naming conventions for a single Feature according to the GeoJSON spec: http://geojson.org
  • geojson_collection - as struct but following naming conventions for a FeatureCollection according to the GeoJSON spec: http://geojson.org
  • gdal_vsi_path - a single file path in the GDAL Virtual File System (VSI) format (via GDAL Python bindings)
  • ogr_feature - a single Feature object from an OGR source (via GDAL Python bindings)
  • ogr_feature_array - a Python list (array) of a single Feature objects from an OGR source
  • record - a Python dict (hashmap)
  • record_array - a Python list (array) of dict
  • string- a general string
  • struct - a JSON-like generic tree structure
  • xml_doc_as_string - a string representation of a complete XML document
  • xml_line_stream - each Packet contains a line (string) from an XML file or string representation (DEPRECATED)

Many components, in particular Filters, are able to transform data formats. For example the XmlElementStreamerFileInput can produce an etree_element, a subsequent XmlAssembler can create small in-memory etree_doc s that can be fed into an XsltFilter, which outputs a transformed etree_doc. The type any is a catch-all, for example used for printing any object to standard output in the stetl.packet.Component. An etree_element may also be interesting to be able to process single features.

Starting with Stetl 1.0.7 a new stetl.filters.formatconverter.FormatConverterFilter class provides a Stetl Filter to allow almost any conversion between otherwise incompatible Components.

TODO: the Packet typing system is still under constant review and extension. Soon it will be possible to add new data types and converters. We have deliberately chosen not to define a single internal datatype like a “Feature”, both for flexibility and performance reasons.

Multiple Chains

Usually a complete ETL will require multiple steps/commands. For example we need to create a database, maybe tables and/or making tables empty. Also we may need to do postprocessing, like removing duplicates in a table etc. In order to have repeatable/reusable ETL without any manual steps, we can specify multiple Chains within a single Stetl config. The syntax: chains are separated by commas (steps are sill separated by pipe symbols).

Chains are executed in order. We can even reuse the specified components from within the same file. Each will have a separate instance within a Chain.

For example in the Top10NL example we see three Chains:

[etl]
chains = input_sql_pre|schema_name_filter|output_postgres,
                input_big_gml_files|xml_assembler|transformer_xslt|output_ogr2ogr,
                input_sql_post|schema_name_filter|output_postgres

Here the Chain input_sql_pre|schema_name_filter|output_postgres sets up a PostgreSQL schema and creates tables. input_big_gml_files|xml_assembler|transformer_xslt|output_ogr2ogr does the actual ETL and input_sql_post|schema_name_filter|output_postgres does some PostgreSQL postprocessing.

Chain Splitting

In some cases we may want to split processed data to multiple Filters or Outputs. For example to produce output files in multiple formats like GML, GeoJSON etc or to publish converted (Filtered) data to multiple remote services (SOS, SensorThings API) or just for simple debugging to a target Output and StandardOutput.

See issue https://github.com/geopython/stetl/issues/35 and the Chain Split example.

Here the Chains are split by using () in the ETL Chain definition:

# Transform input xml to valid GML file using an XSLT filter and pass to multiple outputs.
# Below are two Chains: simple Output splitting and splitting to 3 sub-Chains at Filter level.

[etl]
chains = input_xml_file | transformer_xslt |(output_file)(output_std),
         input_xml_file | (transformer_xslt|output_file) (output_std) (transformer_xslt|output_std)


[input_xml_file]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input/cities.xml

[transformer_xslt]
class = stetl.filters.xsltfilter.XsltFilter
script = cities2gml.xsl

[output_file]
class = stetl.outputs.fileoutput.FileOutput
file_path = output/gmlcities.gml

[output_std]
class = stetl.outputs.standardoutput.StandardOutput

Chain Merging

In some cases we may want to merge (combine, join) multiple input streams.

For example to harvest data from multiple HTTP REST APIs, or to realize a Filter that integrates data from two data-sources.

See issue https://github.com/geopython/stetl/issues/59 and the Chain Merge example.

Here the Chains are merged by using () notation in the ETL Chain definition, possibly even combined with Splitting Outputs:

# Merge two inputs into single Filter.

[etl]
chains = (input_1) (input_2)|transformer_xslt|output_std,
                 (input_1) (input_2)|transformer_xslt|(output_file)(output_std)


[input_1]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input1/cities.xml

[input_2]
class = stetl.inputs.fileinput.XmlFileInput
file_path = input2/cities.xml

[transformer_xslt]
class = stetl.filters.xsltfilter.XsltFilter
script = cities2gml.xsl

[output_file]
class = stetl.outputs.fileoutput.FileOutput
file_path = output/gmlcities.gml

[output_std]
class = stetl.outputs.standardoutput.StandardOutput

Cases

This chapter lists various cases/projects where Stetl is used.

NLExtract

NLExtract https://nlextract.nl is a development project that aims to provide ETL-tooling for all Dutch Open Geo-Datasets, in particular the country wide “Key Registries” (Dutch: Basisregistraties) like Cadastral Parcels (BRK), Topography (BRT+BGT) and Buildings and Addresses (BAG). These datasets are provided as XML/GML. The ETL mostly provides a transformation to PostGIS. For all Key Registries, except for the BAG, Stetl is used, basically as-is, without extra (Python) programming. See also the NLExtract GitHub: https://github.com/nlextract/NLExtract

Adresses and Buildings (BAG v2)

BAG version 1 ETL was developed as a custom Python program. In 2021 Dutch Kadaster released BAG version 2. For NLExtract a moment to switch ETL for BAG to Stetl. In particular use is made of the recent (v3.2.1+) GDAL/OGR LVBAG Driver and Python bindings for GDAL VSI /vsizip file handling.

See https://github.com/nlextract/NLExtract/tree/master/bagv2/etl and the Stetl conf at https://github.com/nlextract/NLExtract/tree/master/bagv2/etl/conf/

Topography (BRT)

Includes ETL for 5 scale levels: TOP10NL through TOP1000NL.

See https://github.com/nlextract/NLExtract/tree/master/brt/top10nl/etl and the Stetl conf at e.g. https://github.com/nlextract/NLExtract/tree/master/brt/top10nl/etl/conf/

Detailed Topography (BGT)

This is a very large and heavy dataset based on CityGML. Stetl streaming ETL is here at its best.

See https://github.com/nlextract/NLExtract/tree/master/bgt and the Stetl conf at https://github.com/nlextract/NLExtract/blob/master/bgt/etl/conf/

INSPIRE

These were the origins of Stetl. This project was sponsored by Kadaster. See https://github.com/justb4/inspire-foss. The ETL involved the transformation of Dutch Key Registries (see above) to harmonized INSPIRE GML according to the Annexes.

Addresses

BAG to INSPIRE Addresses Annex II Theme.

See https://github.com/justb4/inspire-foss/blob/master/etl/NL.Kadaster/Addresses/

Ordnance Survey

A successful Proof-of-Concept to convert Ordnance Survey Mastermap GML to PostGIS:

https://github.com/geopython/stetl/tree/master/examples/ordnancesurvey

SOSPilot

A SensorWeb project by Geonovum, see http://sensors.geonovum.nl.

Dutch AQ to WFS/WMS(-Time) and SOS

Stetl was used for ETL from Dutch Air Quality Data from RIVM (XML) to WMS(-Time), WFS and SOS. The latter was effected by SOS-Transactional publication. Documentation at http://sospilot.readthedocs.org and ETL on GitHub at https://github.com/Geonovum/sospilot/tree/master/src/rivm-lml

Dutch AQ to EAI Reporting

Stetl was used to generate XML-based reports for the EU EAI:

https://github.com/Geonovum/sospilot/tree/master/src/aq-report

This involved the first use of Jinja2 templating for complex XML/GML generation.

Smart Emission

Sensors for air quality, meteo and audio at civilians. Project by University of Nijmegen/Gemeente Nijmegen with participation by Geonovum. Stetl is used to transform a low-level sensor API to PostGIS and later on WMS/WFS/SOS and the SensorThings API. Also InfluxDB output is developed here.

This is also an example how to use a Stetl Docker image:

See https://github.com/Geonovum/smartemission/tree/master/etl

API and Code

Below is the API documention for the the Stetl Python code.

Main Entry Points

There are several entry points through which Stetl can be called. The most common is to use the commandline script bin/stetl. This command should be available after doing an install.

In some contexts like integrations you may want to call Stetl via Python. The entries are then.

class stetl.etl.ETL(options_dict, args_dict=None)[source]

The main class: builds ETL Chains with connected Components from a config and let them run.

Usually this class is called via main but it may be called directly for direct integration.

env_expand_args_dict(args_dict, args_names)[source]

Expand values in dict with equivalent values from the OS Env. NB vars in OS Env should be prefixed with STETL_ or stetl_ as to get overrides by accident.

Returns:expanded args_dict or None

Core Framework

The core framework is directly under the directory src/stetl. Below are the main seven classes. Their interrelation is as follows:

One or more stetl.chain.Chain objects are built from a Stetl ETL configuration via the stetl.factory.Factory class. A stetl.chain.Chain consists of a set of connected stetl.component.Component objects. A stetl.component.Component is either an stetl.input.Input, an stetl.output.Output or a stetl.filter.Filter. Data and status flows as stetl.packet.Packet objects from an stetl.input.Input via zero or more stetl.filter.Filter objects to a final stetl.output.Output.

As a trivial example: an stetl.input.Input could be an XML file, a stetl.filter.Filter could represent an XSLT file and an stetl.output.Output a PostGIS database. This is effected by specialized classes in the subpackages inputs, filters, and outputs. New in 1.1.0: stetl.Splitter to split data to multiple Outputs and stetl.Merger to combine multiple Inputs.

class stetl.factory.Factory[source]

Object and class Factory (Pattern). Based on: http://stackoverflow.com/questions/2226330/instantiate-a-python-class-from-a-name

class_forname(class_string)[source]

Returns class instance specified by a string.

Args:
class_string: The string representing a class.
Raises:
ValueError if module part of the class is not specified.
new_instance(class_obj, configdict, section)[source]

Returns object instance from class instance.

Args:
class_obj: object representing a class instance. args: standard args. kwargs: standard args.
class stetl.component.Component(configdict, section, consumes='none', produces='none')[source]

Abstract Base class for all Input, Filter and Output Components.

after_chain_invoke(packet)[source]

Called right after entire Component Chain invoke.

after_invoke(packet)[source]

Called right after Component invoke.

before_invoke(packet)[source]

Called just before Component invoke.

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

init()[source]

Allows derived Components to perform a one-time init.

input_format()[source]
CONFIG
The specific input format if the consumes parameter is a list or the format to be converted to the output_format.
  • type: str
  • required: False
  • default: None
invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

output_format()[source]
CONFIG
The specific output format if the produces parameter is a list or the format to which the input format is converted.
  • type: str
  • required: False
  • default: None
timer_stop(start_time)[source]

Collect and calculate per-Component performance timing stats. :param start_time: :return:

class stetl.component.Config(ptype=<class 'str'>, default=None, required=False)[source]

Decorator class to tie config values from the .ini file to object instance property values. Somewhat like the Python standard @property but with the possibility to define default values, typing and making properties required.

Each property is defined by @Config(type, default, required). Basic idea comes from: https://wiki.python.org/moin/PythonDecoratorLibrary#Cached_Properties

class stetl.chain.Chain(chain_str, config_dict)[source]

Holder for single invokable pipeline of components A Chain is basically a singly linked list of Components Each Component executes a part of the total ETL. Data along the Chain is passed within a Packet object. The compatibility of input and output for linked Components is checked when adding a Component to the Chain.

add(etl_comp)[source]

Add component to end of Chain :param etl_comp: :return:

assemble()[source]

Builder method: build a Chain of linked Components :return:

get_by_class(clazz)[source]

Get Component instance from Chain by class, mainly for testing. :param clazz: :return Component:

get_by_id(id)[source]

Get Component instance from Chain, mainly for testing. :param name: :return Component:

get_by_index(index)[source]

Get Component instance from Chain by position/index in Chain, mainly for testing. :param clazz: :return Component:

run()[source]

Run the ETL Chain. :return:

class stetl.packet.FORMAT[source]

Format of Packet (enumeration).

Current possible values:

  • ‘none’
  • ‘xml_line_stream’
  • ‘line_stream’
  • ‘etree_doc’
  • ‘etree_element’
  • ‘etree_feature_array’
  • ‘xml_doc_as_string’
  • ‘string’
  • ‘record’
  • ‘record_array’
  • ‘struct’
  • ‘geojson_feature’
  • ‘geojson_collection’
  • ‘gdal_vsi_path`
  • ‘ogr_feature’
  • ‘ogr_feature_array’
  • ‘any’
class stetl.packet.Packet(data=None)[source]

Represents units of (any) data and status passed along Chain of Components.

class stetl.input.Input(configdict, section, produces)[source]

Bases: stetl.component.Component

Abstract Base class for all Input Components.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.output.Output(configdict, section, consumes)[source]

Bases: stetl.component.Component

Abstract Base class for all Output Components.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filter.Filter(configdict, section, consumes, produces)[source]

Bases: stetl.component.Component

Maps input to output. Abstract base class for specific Filters.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.splitter.Splitter(config_dict, child_list)[source]

Bases: stetl.component.Component

Component that splits a single input to multiple output Components. Use this for example to produce multiple output file formats (GML, GeoJSON etc) or to publish to multiple remote services (SOS, SensorThings API) or for simple debugging: target Output and StandardOutput.

after_chain_invoke(packet)[source]

Called right after entire Component Chain invoke.

after_invoke(packet)[source]

Called right after Component invoke.

before_invoke(packet)[source]

Called just before Component invoke.

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

init()[source]

Allows derived Components to perform a one-time init.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.merger.Merger(config_dict, child_list)[source]

Bases: stetl.component.Component

Component that merges multiple Input Components into a single Component. Use this for example to combine multiple input streams like API endpoints. The Merger will embed Child Components to which actions are delegated. A Child Component may be a sub-Chain e.g. (Input|Filter|Filter..) sequence. Hence the “next” should be coupled to the last Component in that sub-Chain with the degenerate case where the sub-Chain is a single (Input) Component. NB this Component can only be used for Inputs.

first(child)[source]

Get first Component in Child sub-Chain. :param child: :return: first Component

last(child)[source]

Get last Component in Child sub-Chain. :param child: :return: last Component

Components: Inputs

class stetl.inputs.dbinput.DbInput(configdict, section, produces)[source]

Bases: stetl.input.Input

Input from any database (abstract base class).

class stetl.inputs.dbinput.PostgresDbInput(configdict, section)[source]

Bases: stetl.inputs.dbinput.SqlDbInput

Input by querying records from a Postgres database. Input is a query, like SELECT * from mytable. Output is zero or more records as record array (array of dict) or single record (dict).

produces=FORMAT.record_array (default) or FORMAT.record

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

host()[source]
CONFIG
host name or host IP-address, defaults to ‘localhost’
  • type: str
  • required: False
  • default: localhost
init()[source]

Allows derived Components to perform a one-time init.

password()[source]
CONFIG
User password, defaults to ‘postgres’
  • type: str
  • required: False
  • default: postgres
port()[source]
CONFIG
port for host, defaults to ‘5432’
  • type: str
  • required: False
  • default: 5432
raw_query(query_str)[source]

Performs DB-specific query and returns raw records iterator.

schema()[source]
CONFIG
The postgres schema name, defaults to ‘public’
  • type: str
  • required: False
  • default: public
user()[source]
CONFIG
User name, defaults to ‘postgres’
  • type: str
  • required: False
  • default: postgres
class stetl.inputs.dbinput.SqlDbInput(configdict, section)[source]

Bases: stetl.inputs.dbinput.DbInput

Input using a query from any SQL-based RDBMS (abstract base class).

column_names()[source]
CONFIG
Column names to populate records with. If empty taken from table metadata.
  • type: str
  • required: False
  • default: None
database_name()[source]
CONFIG
Database name
  • type: str
  • required: True
  • default: None
do_query(query_str)[source]

DB-neutral query returning Python record list.

query()[source]
CONFIG
The query (string) to fire.
  • type: str
  • required: False
  • default: None
raw_query(query_str)[source]

Performs DB-specific query and returns raw records iterator.

read_once()[source]
CONFIG
Read once? i.e. only do query once and stop
  • type: bool
  • required: False
  • default: False
result_to_output(db_tuples)[source]

Convert DB-specific record tuples to single Python record (dict) or record array (list of dict).

table()[source]
CONFIG
Table name
  • type: str
  • required: False
  • default: None
tuples_to_records(db_tuples, columns=None)[source]

Convert tuple array (list of tuple) to list of records (list of dict’s) using list of column names.

class stetl.inputs.dbinput.SqliteDbInput(configdict, section)[source]

Bases: stetl.inputs.dbinput.SqlDbInput

Input by querying records from a SQLite database. Input is a query, like SELECT * from mytable. Output is zero or more records as record array (array of dict) or single record (dict).

produces=FORMAT.record_array (default) or FORMAT.record

init()[source]

Allows derived Components to perform a one-time init.

raw_query(query_str)[source]

Performs DB-specific query and returns raw records iterator.

class stetl.inputs.httpinput.ApacheDirInput(configdict, section, produces='record')[source]

Bases: stetl.inputs.httpinput.HttpInput

Read file data from an Apache directory “index” HTML page. Uses http://stackoverflow.com/questions/686147/url-tree-walker-in-python produces=FORMAT.record. Each record contains file_name and file_data (other meta data like date time is too fragile over different Apache servers).

file_ext()[source]
CONFIG
The file extension for target files in Apache dir.
  • type: str
  • required: False
  • default: xml
filter_file(file_name)[source]

Filter the file_name, e.g. to suppress reading, default: return file_name.

Parameters:file_name –
Return string or None:
 
init()[source]

Read the list of files from the Apache index URL.

next_file()[source]

Return a tuple (name, date, size) with next file info.

Return tuple:
no_more_files()[source]

More files left?.

Return Boolean:
read(packet)[source]

Read the data from the URL.

Parameters:packet –
Returns:
class stetl.inputs.httpinput.HttpInput(configdict, section, produces='any')[source]

Bases: stetl.input.Input

Fetch data from remote services like WFS via HTTP protocol. Base class: subclasses will do datatype-specific formatting of the returned data.

produces=FORMAT.any

add_authorization(request)[source]

Add authorization from config data. Authorization scheme-specific. May be extended or overloaded for additional schemes.

Parameters:request – the HTTP Request
Returns:
auth()[source]
CONFIG

Authentication data: Flat JSON-like struct dependent on auth type/schema. Only the type field is required, other fields depend on auth schema. Supported values :

type: basic|token

If the type is basic (HTTP Basic Authentication) two additional fields user and password are required. If the type is token (HTTP Token) additional two additional fields keyword and token are required.

Any required Base64 encoding is provided by HttpInput.

Examples:

# Basic Auth
url = https://some.rest.api.com
auth = {
    type: basic,
    user: myname
    password: mypassword
}

# Token Auth
url = https://some.rest.api.com
auth = {
    type: token,
    keyword: Bearer
    token: mytoken
}
  • type: dict
  • required: False
  • default: None
format_data(data)[source]

Format response data, override in subclasses, defaults to returning original data. :param packet: :return:

parameters()[source]
CONFIG

Flat JSON-like struct of the parameters to be appended to the url.

Example: (parameters require quotes):

url = http://geodata.nationaalgeoregister.nl/natura2000/wfs
parameters = {
    service : WFS,
    version : 1.1.0,
    request : GetFeature,
    srsName : EPSG:28992,
    outputFormat : text/xml; subtype=gml/2.1.2,
    typename : natura2000
}
  • type: dict
  • required: False
  • default: None
read(packet)[source]

Read the data from the URL.

Parameters:packet –
Returns:
read_from_url(url, parameters=None)[source]

Read the data from the URL.

Parameters:
  • url – the url to fetch
  • parameters – optional dict of query parameters
Returns:
url()[source]
CONFIG
The HTTP URL string.
  • type: str
  • required: True
  • default: None
class stetl.inputs.ogrinput.OgrInput(configdict, section)[source]

Bases: stetl.input.Input

Direct GDAL OGR input via Python OGR wrapper. Via the Python API http://gdal.org/python an OGR data source is accessed and from each layer the Features are read. Each Layer corresponds to a “doc”, so for multi-layer sources the ‘end-of-doc’ flag is set after a Layer has been read.

This input can read almost any geospatial dataformat. One can use the features directly in a Stetl Filter or use a converter to e.g. convert to GeoJSON structures.

produces=FORMAT.ogr_feature or FORMAT.ogr_feature_array (all features)

data_source()[source]
CONFIG
String denoting the OGR datasource. Usually a path to a file like “path/rivers.shp” or connection string to PostgreSQL like “PG: host=localhost dbname=’rivers’ user=’postgres’”.
  • type: str
  • required: True
  • default: None
init()[source]

Allows derived Components to perform a one-time init.

source_format()[source]
CONFIG

Instructs GDAL to use driver by that name to open datasource. Not required for many standard formats that are self-describing like ESRI Shapefile.

Examples: ‘PostgreSQL’, ‘GeoJSON’ etc

  • type: str
  • required: False
  • default: None
source_options()[source]
CONFIG
Custom datasource-specific options. Used in gdal.SetConfigOption().
  • type: dict
  • required: False
  • default: None
sql()[source]
CONFIG
String with SQL query. Mandatory for PostgreSQL OGR source.
  • type: str
  • required: False
  • default: None
class stetl.inputs.ogrinput.OgrPostgisInput(configdict, section)[source]

Bases: stetl.input.Input

Input from PostGIS via ogr2ogr command. For now hardcoded to produce an ogr GML line stream. OgrInput may be a better alternative.

Alternatives: either stetl.input.PostgresqlInput or stetl.input.OgrInput.

produces=FORMAT.xml_line_stream

in_pg_db()[source]
CONFIG
Database name input DB.
  • type: str
  • required: True
  • default: None
in_pg_host()[source]
CONFIG
Host of input DB.
  • type: str
  • required: False
  • default: localhost
in_pg_password()[source]
CONFIG
Password input DB.
  • type: str
  • required: False
  • default: postgres
in_pg_port()[source]
CONFIG
Port of input DB.
  • type: str
  • required: False
  • default: 5432
in_pg_schema()[source]
CONFIG
DB Schema name input DB.
  • type: str
  • required: False
  • default: None
in_pg_sql()[source]
CONFIG
The input query (string) to fire.
  • type: str
  • required: False
  • default: None
in_pg_user()[source]
CONFIG
User input DB.
  • type: str
  • required: False
  • default: postgres
in_srs()[source]
CONFIG
SRS (projection) (ogr2ogr -s_srs) input DB e.g. ‘EPSG:28992’.
  • type: str
  • required: False
  • default: None
init()[source]

Allows derived Components to perform a one-time init.

out_dimension()[source]
CONFIG
Dimension (OGR: DIM=N) of features in output stream.
  • type: str
  • required: False
  • default: 2
out_geotype()[source]
CONFIG
OGR Geometry type new layer in output stream, e.g. POINT.
  • type: str
  • required: False
  • default: None
out_gml_format()[source]
CONFIG
GML format OGR name in output stream, e.g. ‘GML3’.
  • type: str
  • required: False
  • default: None
out_layer_name()[source]
CONFIG
New Layer name (ogr2ogr -nln) output stream, e.g. ‘address’.
  • type: str
  • required: False
  • default: None
out_srs()[source]
CONFIG
Target SRS (ogr2ogr -t_srs) code output stream.
  • type: str
  • required: False
  • default: None

Components: Filters

class stetl.filters.xsltfilter.XsltFilter(configdict, section)[source]

Bases: stetl.filter.Filter

Invokes XSLT processor (via lxml) for given XSLT script on an etree doc.

consumes=FORMAT.etree_doc, produces=FORMAT.etree_doc

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

script()[source]
CONFIG
Path to XSLT script file.
  • type: str
  • required: True
  • default: None
class stetl.filters.xmlassembler.XmlAssembler(configdict, section)[source]

Bases: stetl.filter.Filter

Split a stream of etree DOM XML elements (usually Features) into etree DOM docs. Consumes and buffers elements until max_elements reached, will then produce an etree doc.

consumes=FORMAT.etree_element, produces=FORMAT.etree_doc

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.xmlelementreader.XmlElementReader(configdict, section)[source]

Bases: stetl.filter.Filter

Extracts XML elements from a file, outputs each feature element in Packet. Parsing is streaming (no internal DOM buildup) so any file size can be handled. Use this class for your big GML files!

consumes=FORMAT.string, produces=FORMAT.etree_element

element_tags()[source]
CONFIG
Comma-separated string of XML (feature) element tag names of the elements that should be extracted and added to the output element stream.
  • type: list
  • required: True
  • default: None
invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

strip_namespaces()[source]
CONFIG
should namespaces be removed from the input document and thus not be present in the output element stream?
  • type: bool
  • required: False
  • default: False
class stetl.filters.xmlvalidator.XmlSchemaValidator(configdict, section)[source]

Bases: stetl.filter.Filter

Validates an etree doc and prints result to log.

consumes=FORMAT.etree_doc, produces=FORMAT.etree_doc

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.sieve.AttrValueRecordSieve(configdict, section)[source]

Bases: stetl.filters.sieve.Sieve

Sieves by attr/value(s) in Record Packets.

attr_name()[source]
CONFIG
Name of attribute whose value(s) are to be sieved.
  • type: str
  • required: True
  • default: None
attr_values()[source]
CONFIG
Value(s) for attribute to be to sieved. If empty any value is passed through (existence of attr_name is criterium).
  • type: list
  • required: False
  • default: []
sieve(packet)[source]

Filter out Packets that are not matching designated attr value(s). :param packet: :return:

class stetl.filters.sieve.Sieve(configdict, section, consumes, produces)[source]

Bases: stetl.filter.Filter

ABC for specific Sieves that pass-through, “sieve”, Packets based on criteria in their data.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

sieve(packet)[source]

To be implemented in subclasses. :param packet: :return:

class stetl.filters.stringfilter.StringConcatFilter(configdict, section)[source]

Bases: stetl.filters.stringfilter.StringFilter

Concatenates a specified string with the input string (packet.data) either/or as prefix (prepend) or postfix (append) and outputs that concatenation.

consumes=FORMAT.string, produces=FORMAT.string

append_string()[source]
CONFIG

String to be appended.

Example: append_string = /002PND150904.xml

  • type: str
  • required: False
  • default: None
prepend_string()[source]
CONFIG

String to be prepended.

Example: prepend_string = /vsizip/

  • type: str
  • required: False
  • default: None
class stetl.filters.stringfilter.StringFilter(configdict, section, consumes, produces)[source]

Bases: stetl.filter.Filter

Base class for any string filtering

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.stringfilter.StringSubstitutionFilter(configdict, section)[source]

Bases: stetl.filters.stringfilter.StringFilter

String filtering using Python advanced String formatting. String should have substitutable values like {schema} {foo} format_args should be of the form format_args = schema:test foo:bar …

consumes=FORMAT.string, produces=FORMAT.string

format_args()[source]
CONFIG

Provides a list of format arguments used by the string substitution filter. Formatting of content according to Python String.format(). String should have substitutable values like {schema} {foo}.

Example: format_args = schema:test foo:bar

  • type: str
  • required: True
  • default: None
separator()[source]
CONFIG
Provides the separator to split the format argument names from their values.
  • type: str
  • required: False
  • default: :
class stetl.filters.templatingfilter.Jinja2TemplatingFilter(configdict, section)[source]

Bases: stetl.filters.templatingfilter.TemplatingFilter

Implements Templating using Jinja2. Jinja2 http://jinja.pocoo.org, is a modern and designer-friendly templating language for Python modelled after Django’s templates. A ‘struct’ format as input provides a tree-like structure that could originate from a JSON file or REST service. This input struct provides all the variables to be inserted into the template. The template itself can be configured in this component as a Jinja2 string or -file. An optional ‘template_search_paths’ provides a list of directories from which templates can be fethced. Default is the current working directory. Via the optional ‘globals_path’ a JSON structure can be inserted into the Template environment. The variables in this globals struture are typically “boilerplate” constants like: id-prefixes, point of contacts etc.

consumes=FORMAT.struct, produces=FORMAT.string

add_env_filters(jinja2_env)[source]

Register additional Filters on the template environment by updating the filters dict: Somehow min and max of list are not present so add them as well.

create_template()[source]

To be overridden in subclasses.

static geojson2gml_filter(value, source_crs=4326, target_crs=None, gml_id=None, gml_format='GML2', gml_longsrs='NO')[source]

Jinja2 custom Filter: generates any GML geometry from a GeoJSON geometry. By specifying a target_crs we can even reproject from the source CRS. The gml_format=GML2|GML3 determines the general GML form: e.g. pos/posList or coordinates. gml_longsrs=YES|NO determines the srsName format like EPSG:4326 or urn:ogc:def:crs:EPSG::4326 (long).

template_globals_path()[source]
CONFIG
One or more JSON files or URLs with global variables that can be used anywhere in template. Multiple files will be merged into one globals dictionary
  • type: str
  • required: False
  • default: None
template_search_paths()[source]
CONFIG
List of directories where to search for templates, default is current working directory only.
  • type: str
  • required: False
  • default: None
class stetl.filters.templatingfilter.StringTemplatingFilter(configdict, section)[source]

Bases: stetl.filters.templatingfilter.TemplatingFilter

Implements Templating using Python’s internal string.Template. A template string or file should be configured. The input record contains the actual values to be substituted in the template string as a record (key/value pairs). Output is a regular string.

consumes=FORMAT.record or FORMAT.record_array, produces=FORMAT.string

create_template()[source]

To be overridden in subclasses.

safe_substitution()[source]
CONFIG
Apply safe substitution? With this method, string.Template.safe_substitute will be invoked, instead of string.Template.substitute. If placeholders are missing from mapping and keywords, instead of raising an exception, the original placeholder will appear in the resulting string intact.
  • type: bool
  • required: False
  • default: False
class stetl.filters.templatingfilter.TemplatingFilter(configdict, section, consumes='any', produces='string')[source]

Bases: stetl.filter.Filter

Abstract base class for specific template-based filters. See https://wiki.python.org/moin/Templating Subclasses implement a specific template language like Python string.Template, Mako, Genshi, Jinja2,

consumes=FORMAT.any, produces=FORMAT.string

create_template()[source]

To be overridden in subclasses.

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

init()[source]

Allows derived Components to perform a one-time init.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

template_file()[source]
CONFIG
Path to template file. One of template_file or template_string needs to be configured.
  • type: str
  • required: False
  • default: None
template_string()[source]
CONFIG
Template string. One of template_file or template_string needs to be configured.
  • type: str
  • required: False
  • default: None
class stetl.filters.gmlfeatureextractor.GmlFeatureExtractor(configdict, section='gml_feature_extractor')[source]

Bases: stetl.filter.Filter

Extract arrays of GML features etree elements from etree docs.

consumes=FORMAT.etree_doc, produces=FORMAT.etree_feature_array

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.formatconverter.FormatConverter(configdict, section)[source]

Bases: stetl.filter.Filter

Converts (almost) any packet format (if converter available).

consumes=FORMAT.any, produces=FORMAT.any but actual formats are changed at initialization based on the input to output format to be converted via the input_format and output_format config parameters.

converter_args()[source]
CONFIG
Custom converter-specific arguments.
  • type: dict
  • required: False
  • default: None
static etree_doc2geojson_collection(packet, converter_args=None)[source]

Use converter_args to determine XML tag names for features and GeoJSON feature id. For example

converter_args = {
‘root_tag’: ‘FeatureCollection’, ‘feature_tag’: ‘featureMember’, ‘feature_id_attr’: ‘fid’ }
Parameters:
  • packet –
  • converter_args –
Returns:
static etree_doc2struct(packet, strip_space=True, strip_ns=True, sub=False, attr_prefix='', gml2ogr=True, ogr2json=True)[source]
Parameters:
  • packet –
  • strip_space –
  • strip_ns –
  • sub –
  • attr_prefix –
  • gml2ogr –
  • ogr2json –
Returns:
static etree_elem2geojson_feature(packet, converter_args=None)[source]
static etree_elem2struct(packet, strip_space=True, strip_ns=True, sub=False, attr_prefix='', gml2ogr=True, ogr2json=True)[source]
Parameters:
  • packet –
  • strip_space –
  • strip_ns –
  • sub –
  • attr_prefix –
  • gml2ogr –
  • ogr2json –
Returns:
init()[source]

Allows derived Components to perform a one-time init.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.execfilter.CommandExecFilter(configdict, section)[source]

Bases: stetl.filters.execfilter.ExecFilter

Executes an arbitrary command and captures the output

consumes=FORMAT.string, produces=FORMAT.string

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.execfilter.ExecFilter(configdict, section, consumes, produces)[source]

Bases: stetl.filter.Filter

Executes any command (abstract base class).

env_args()[source]
CONFIG

Provides of list of environment variables which will be used when executing the given command.

Example: env_args = pgpassword=postgres othersetting=value~with~spaces

  • type: str
  • required: False
  • default:
env_separator()[source]
CONFIG
Provides the separator to split the environment variable names from their values.
  • type: str
  • required: False
  • default: =
invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.nullfilter.NullFilter(configdict, section, consumes='any', produces='any')[source]

Bases: stetl.filter.Filter

Pass-through Filter, does nothing. Mainly used in Test Cases.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.packetbuffer.PacketBuffer(configdict, section)[source]

Bases: stetl.filter.Filter

Buffers all incoming Packets, main use is unit-testing to inspect Packets after ETL is done.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.packetwriter.PacketWriter(configdict, section)[source]

Bases: stetl.filter.Filter

Writes the payload of a packet as a string to a file.

consumes=FORMAT.any, produces=FORMAT.string

file_path()[source]
CONFIG
File path to write content to.
  • type: str
  • required: True
  • default: None
invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.regexfilter.RegexFilter(configdict, section, consumes='string', produces='record')[source]

Bases: stetl.filter.Filter

Extracts data from a string using a regular expression and returns the named groups as a record. consumes=FORMAT.string, produces=FORMAT.record

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

init()[source]

Allows derived Components to perform a one-time init.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

pattern_string()[source]
CONFIG
Regex pattern string. Should contain named groups.
  • type: str
  • required: True
  • default: None
class stetl.filters.fileextractor.FileExtractor(configdict, section, consumes='any', produces='string')[source]

Bases: stetl.filter.Filter

Abstract Base Class. Extracts a file an archive and saves as the configured file name.

consumes=FORMAT.any, produces=FORMAT.string

after_chain_invoke(packet)[source]

Called right after entire Component Chain invoke.

buffer_size()[source]
CONFIG
Buffer size for read buffer during extraction.
  • type: int
  • required: False
  • default: 1073741824
delete_file()[source]
CONFIG
Delete the file when the chain has been completed?
  • type: bool
  • required: False
  • default: True
file_path()[source]
CONFIG
File name to write the extracted file to.
  • type: str
  • required: True
  • default: None
invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.fileextractor.VsiFileExtractor(configdict, section)[source]

Bases: stetl.filters.fileextractor.FileExtractor

Extracts a file from a GDAL /vsi path spec, and saves it as the given file name.

Example paths: /vsizip/{/project/nlextract/data/BAG-2.0/BAGNLDL-08112020.zip}/9999STA08112020.zip’ /vsizip/{/vsizip/{BAGGEM0221L-15022021.zip}/GEM-WPL-RELATIE-15022021.zip}/GEM-WPL-RELATIE-15022021-000001.xml

See also stetl.inputs.fileinput.VsiZipFileInput that generates these paths.

Author: Just van den Broecke

consumes=FORMAT.gdal_vsi_path, produces=FORMAT.string

class stetl.filters.fileextractor.ZipFileExtractor(configdict, section)[source]

Bases: stetl.filters.fileextractor.FileExtractor

Extracts a file from a ZIP file, and saves it as the given file name. Author: Frank Steggink

consumes=FORMAT.record, produces=FORMAT.string

class stetl.filters.vsifilter.VsiFilter(configdict, section, vsiname)[source]

Bases: stetl.filter.Filter

Abstract base class for applying a GDAL/OGR virtual file system (VSI) filter.

invoke(packet)[source]

Components override for Component-specific behaviour, typically read, filter or write actions.

class stetl.filters.vsifilter.VsiZipFilter(configdict, section)[source]

Bases: stetl.filters.vsifilter.VsiFilter

Applies a VSIZIP filter to the input record.

consumes=FORMAT.record, produces=FORMAT.string

NB: stetl.filters.zipfileextractor is deprecated. ZipFileExtractor is now part of module stetl.filters.fileextractor.

Components: Outputs

class stetl.outputs.fileoutput.FileOutput(configdict, section)[source]

Bases: stetl.output.Output

Pretty print input to file. Input may be an etree doc or any other stringify-able input.

consumes=FORMAT.any

file_path()[source]
CONFIG
Path to file, for MultiFileOutput can be of the form like: gmlcities-%03d.gml
  • type: str
  • required: True
  • default: None
class stetl.outputs.fileoutput.MultiFileOutput(configdict, section)[source]

Bases: stetl.outputs.fileoutput.FileOutput

Print to multiple files from subsequent packets like strings or etree docs, file_path must be of a form like: gmlcities-%03d.gml.

consumes=FORMAT.any

class stetl.outputs.standardoutput.StandardOutput(configdict, section)[source]

Bases: stetl.output.Output

Print any input to standard output.

consumes=FORMAT.any

class stetl.outputs.standardoutput.StandardXmlOutput(configdict, section)[source]

Bases: stetl.output.Output

Pretty print XML from etree doc to standard output. OBSOLETE, can be done with StandardOutput

consumes=FORMAT.etree_doc

class stetl.outputs.ogroutput.Ogr2OgrOutput(configdict, section)[source]

Bases: stetl.output.Output

Output from GML etree doc to any OGR2OGR output using the GDAL/OGR ogr2ogr command

consumes=FORMAT.etree_doc

class stetl.outputs.ogroutput.OgrOutput(configdict, section)[source]

Bases: stetl.output.Output

Direct GDAL OGR output via Python OGR wrapper. Via the Python API http://gdal.org/python OGR Features are written.

This output can write almost any geospatial, OGR-defined, dataformat.

consumes=FORMAT.ogr_feature or FORMAT.ogr_feature_array

always_apply_lco()[source]
CONFIG
Flag to indicate whether the layer creation options should be applied to all runs.
  • type: bool
  • required: False
  • default: False
append()[source]
CONFIG
Add to destination destination if it extists (ogr2ogr -append option).
  • type: bool
  • required: False
  • default: False
dest_create_options()[source]
CONFIG

Creation options.

Examples: ..

  • type: list
  • required: False
  • default: []
dest_data_source()[source]
CONFIG
String denoting the OGR data destination. Usually a path to a file like “path/rivers.shp” or connection string to PostgreSQL like “PG: host=localhost dbname=’rivers’ user=’postgres’”.
  • type: str
  • required: True
  • default: None
dest_format()[source]
CONFIG

Instructs GDAL to use driver by that name to open data destination. Not required for many standard formats that are self-describing like ESRI Shapefile.

Examples: ‘PostgreSQL’, ‘GeoJSON’ etc

  • type: str
  • required: False
  • default: None
dest_options()[source]
CONFIG
Custom data destination-specific options. Used in gdal.SetConfigOption().
  • type: dict
  • required: False
  • default: None
init()[source]

Allows derived Components to perform a one-time init.

layer_create_options()[source]
CONFIG
Options for newly created layer (-lco).
  • type: list
  • required: False
  • default: []
new_layer_name()[source]
CONFIG
Layer name for layer created in the destination source.
  • type: str
  • required: True
  • default: None
overwrite()[source]
CONFIG
Overwrite destination if it extists (ogr2ogr -overwrite option).
  • type: bool
  • required: False
  • default: False
sql()[source]
CONFIG
String with SQL query. Mandatory for PostgreSQL OGR dest.
  • type: str
  • required: False
  • default: None
target_srs()[source]
CONFIG
SRS (projection) for the target.
  • type: str
  • required: False
  • default: None
class stetl.outputs.execoutput.CommandExecOutput(configdict, section)[source]

Bases: stetl.outputs.execoutput.ExecOutput

Executes an arbitrary command.

consumes=FORMAT.string

class stetl.outputs.execoutput.ExecOutput(configdict, section, consumes)[source]

Bases: stetl.output.Output

Executes any command (abstract base class).

env_args()[source]
CONFIG

Provides of list of environment variables which will be used when executing the given command.

Example: env_args = pgpassword=postgres othersetting=value~with~spaces

  • type: str
  • required: False
  • default:
env_separator()[source]
CONFIG
Provides the separator to split the environment variable names from their values.
  • type: str
  • required: False
  • default: =
class stetl.outputs.execoutput.Ogr2OgrExecOutput(configdict, section)[source]

Bases: stetl.outputs.execoutput.ExecOutput

Executes an Ogr2Ogr command. Input is a file name to be processed. Output by calling Ogr2Ogr command.

consumes=FORMAT.string

always_apply_lco()[source]
CONFIG
Flag to indicate whether the layer creation options should be applied to all runs.
  • type: bool
  • required: False
  • default: False
cleanup_input()[source]
CONFIG
Flag to indicate whether the input file to ogr2ogr should be cleaned up.
  • type: bool
  • required: False
  • default: False
dest_data_source()[source]
CONFIG
String denoting the OGR data destination. Usually a path to a file like “path/rivers.shp” or connection string to PostgreSQL like “PG: host=localhost dbname=’rivers’ user=’postgres’”.
  • type: str
  • required: True
  • default: None
dest_format()[source]
CONFIG

Instructs GDAL to use driver by that name to open data destination. Not required for many standard formats that are self-describing like ESRI Shapefile.

Examples: ‘PostgreSQL’, ‘GeoJSON’ etc

  • type: str
  • required: False
  • default: None
gfs_template()[source]
CONFIG
Name of GFS template file to use during loading. Passed to ogr2ogr as –config GML_GFS_TEMPLATE <name>
  • type: str
  • required: False
  • default: None
lco()[source]
CONFIG
Options for newly created layer (-lco).
  • type: str
  • required: False
  • default: None
options()[source]
CONFIG
Miscellaneous options to pass to ogr2ogr.
  • type: str
  • required: False
  • default: None
spatial_extent()[source]
CONFIG
Spatial extent (-spat), to pass as xmin ymin xmax ymax
  • type: str
  • required: False
  • default: None
class stetl.outputs.dboutput.DbOutput(configdict, section, consumes)[source]

Bases: stetl.output.Output

Output to any database (abstract base class).

class stetl.outputs.dboutput.PostgresDbOutput(configdict, section)[source]

Bases: stetl.outputs.dboutput.DbOutput

Output to PostgreSQL database. Input is an SQL string. Output by executing input SQL string.

consumes=FORMAT.string

database()[source]
CONFIG
Database name.
  • type: str
  • required: True
  • default: None
host()[source]
CONFIG
Hostname for DB.
  • type: str
  • required: False
  • default: None
password()[source]
CONFIG
DB Password for user.
  • type: str
  • required: False
  • default: None
schema()[source]
CONFIG
Postgres schema name for DB.
  • type: str
  • required: False
  • default: public
user()[source]
CONFIG
DB User name.
  • type: str
  • required: False
  • default: None
class stetl.outputs.dboutput.PostgresInsertOutput(configdict, section, consumes='record')[source]

Bases: stetl.outputs.dboutput.PostgresDbOutput

Output by inserting a single record in a Postgres database table. Input is a Stetl record (Python dict structure) or a list of records. Creates an INSERT for Postgres to insert each single record. When the “replace” parameter is True, any existing record keyed by “key” is attempted to be UPDATEd first.

NB a constraint is that the first and each subsequent each record needs to contain all values as an INSERT and UPDATE query template is built once for the columns in the first record.

consumes=[FORMAT.record_array, FORMAT.record]

exit()[source]

Allows derived Components to perform a one-time exit/cleanup.

init()[source]

Allows derived Components to perform a one-time init.

key()[source]
CONFIG
The key column name of the table, required when replacing records.
  • type: str
  • required: False
  • default: None
replace()[source]
CONFIG
Replace record if exists?
  • type: bool
  • required: False
  • default: False
table()[source]
CONFIG
Table for inserts.
  • type: str
  • required: False
  • default: public
class stetl.outputs.deegreeoutput.DeegreeBlobstoreOutput(configdict, section)[source]

Bases: stetl.output.Output

Insert features into deegree Blobstore from an etree doc.

consumes=FORMAT.etree_doc

init()[source]

Allows derived Components to perform a one-time init.

class stetl.outputs.deegreeoutput.DeegreeFSLoaderOutput(configdict, section)[source]

Bases: stetl.output.Output

Insert features via deegree using deegree’s FSLoader tool from an etree doc.

consumes=FORMAT.etree_doc

Contact

The website stetl.org is the main entry point for all of Stetl.

All development is done via GitHub: see https://github.com/geopython/stetl.

Contact the main author Just van den Broecke via email at just@justobjects.nl.

Online chat via Gitter: https://gitter.im/geopython/stetl

Stetl Projects/Cases

Known uses of Stetl. More detail in the chapter on Cases.

Tools

Tools/components used by/with Stetl.

Indices and tables