The goal of the PML2PythonTranslator is to translate PML-code (PML stands for Prochitecture Markup Language) into Python code that describes the style of a building and is used in the blender-osm plugin for Blender. A description of the concept of this style can be found at here. An additional goal of this development step was to study, how classes could made reusable for the node-based visual editor, which can be found in the repository blosm-nodes, and for other side projects of the blender-osm plugin. Only a subset of PML has been realized in this development step in order to get a feeling of the problems of this concept and to enable a discussion.

pml_grammar (folder): This folder contains the file pml.g4, which defines the grammar of PML, written for the ANTLR parser generator. All the other files in this folder have been generated by the ANTLR4 parser generator (see below) for the target language Python. Do not change the files in this folder.

example.pml: This file contains an example description of a style, written in PML, as far as the translator already has been developed.

PML2PythonTranslator.py: This file contains a demo main program that starts the translation process. The parser created by ANTLR4 creates a parse tree from any file containing PML style code and the walker class then calls the translating functions using the generated listener class.

PythonListener.py: This file contains an interface class between the ANTLR4 listnener pmlListener and the coder class PythonCoder. It just calls the parser rule functions with the same name in the coder class, transferring the retrieved content from the parse tree. In principle, all the code in PythonCoder could be written directly in PythonListener, but as PythonListener depends on the ANTLR4 classes, it could not be reused in other programs without the ANTLR4 runtime for Python. PythonListener provides a method getCode() that gets the result of the translation for the PythonCoder.

PythonCoder.py: This class provides all rule functions that have to create Python code for the blender-osm plugin. Actually, it provides a write() method, that writes the translated text into a string, accessible by the getCode() method of the class. The write() method could be overwritten by a derived class to produce other kind of output. As for instance parsing the nodes in the node-based visual editor is very similar to parsing PML style code, PythonCoder.py can be reused by this editor to create the same code.

Dictionaries.py: This file contains only a fragement of code. It holds the data types and ranges of OSM attributes required by the osm-plugin. It could later be used to hold the information about allowed attribute names of the style blocks, to check if they are allowed in the context of the actual element, and so on. This will be object of future discussions.

To run the PML2PythonTranslator, a Python 3 interpreter is required.

If you like to change the grammar, ANTLR4 has to be installed (see here). The parser files for the Python 3 runtime can then be generated using

java org.antlr.v4.Tool -Dlanguage=Python3 pml.g4

To run the translator using the example style example.pml, execute

python PML2PythonTranslator.py example.pml

The translated code will be written to stdout.