SynData for Robot Framework

The purpose of this library is to generate synthetic test data that is as consistent as possible. The library achieves this by providing specialized generators for different countries, which are used to generate the test data.

In addition, there are two functions that characterize this library:

• The context defines a "data space" in which previously generated data is stored, so that repeating a keyword returns the value of the first call.
• The library logs the generated data in a separate file, and a corresponding switch allows the library to use a log file as a data source. This allows, for example, a test case to be re-executed with the data from a previous execution. To do this, only the import statement of the library needs to be changed, and the test case can remain unchanged.

See keyword documentation for more details.

Installation instructions

The library can be installed as usual using pip:

pip install robotframework-syndata

Examples

... to follow at a later date

Notes on the structure of the library

The library should support the concept of localization from the outset so that test data can be generated specifically for a country. Since only a specialized generator is initially provided for Germany, there is a general generator that uses Faker to generate test data.

The generated data should be consistent. When using a context, a stored date may need to be returned, and in some cases, only data from a defined data source should be returned and no data should be generated at all. This short list makes it clear that there are a number of factors that determine whether a date is generated or taken from a data source. In order to largely eliminate sources of error, the dependencies are modeled in decision tables and the Python code is generated from the decision tables.

This project uses the LF-ET decision table editor from LOHRFINK software engineering GmbH & Co. KG to create and maintain the decision tables. In LF-ET, the rules can be filtered in almost any way, so that "excerpts" from a large set of rules can also be analyzed. LF-ET provides automated verification of completeness, absence of redundancies and contradictions. From such a decision table, LF-ET can generate the corresponding source code for different programming languages, which is also used in this project.

The classes generated from a decision table are always introduced with the same comment lines:

# *** WARNING: DO NOT MODIFY *** This is a generated Python source code!
#
# Generated by LF-ET 2.4.1 (260127a), https://www.lohrfink.de/lfet

This project is supported by LOHRFINK software engineering GmbH & Co. KG, and a current version of the decision table editor LF-ET is available at the following link: https://www.lohrfink.de/lfet/lfet.latest.inst

The specialized generators and the required data are stored in separate packages that can be assigned using the name of the region. For Germany, the code and data can be found in the folder src/SynData/de_DE.

Special thanks

I would like to thank LOHRFINK software engineering GmbH & Co. KG for providing a project license for this project. Thanks to the project license for the LF-ET decision table editor, the process logic for generation can be modeled in decision tables and the corresponding Python code can be generated.

This library is also the result of constructive discussions with the Robot Framework community. Special thanks go to René, who gave me the idea that the library should log the generated data and that the log should be used as a source for repeating test cases.