Python bindings for go text/template

go_template works Python 2.7, 3.5, 3.6, 3.7.

pip install go_template

Content of sample.tmpl

{{.Count}} items are made of {{.Material}}

Content of values.yml

Count: 12
Material: Wool

1. Print rendered output to stdout

>>> import go_template
>>> go_template.render_template('tests/sample.tmpl','tests/values.yml','')
12 items are made of Wool

2. Get rendered output in a file

>>> import go_template
>>> go_template.render_template('tests/sample.tmpl','tests/values.yml','output.txt')

Content of output.txt

12 items are made of Wool

NOTE: Paths provided to render_template should either be absolute path or relative to directory where it is ran.

For building a fresh shared object of text/template, you must have golang^1.5 installed.

./build.sh

This will create template.so in the bind folder.

Currently, there is no python package which exposes golang text/template functionality to python. And I am in the process of learning about interoperability between different languages. So, I started working on this as a learning project.

Golang library cannot be directly used in python. Firstly, we have to compile it as shared object or archive for interoperability with C. And then create python bindings for this C object.

CPython is the original Python implementation and provides cpython API for creating python wrapper, but the wrapping code is in C. There is a library gopy which exactly uses this approach. But it works only on go1.5 and for python2.

If we want to write the wrapping code in python, there are Cython and ctypes. Ctypes allow directly importing the C library and calling functions through its interface. This project uses ctypes for calling go functions.

When a golang library is compiled as shared object, cgo handles exposing functions and data type conversion. Using ctypes, we can only modify simple data type, string, int, float, bool. I tried converting python class to golang struct, but it failed.

So, I created a golang wrapper over text/template library, which takes simple datatypes. And handles complex operation in this layer. Then a python wrapper over this layer using ctypes.

It is far from complete and doesn't use the best approach. Currently, it has only one function which takes path of template and value file. And depending on the third argument, either writes to stdout if empty or to file if given its path.

This project is licensed under MIT License.