Run Stata's do-files with pytask.

pytask-stata is available on PyPI and Anaconda.org. Install it with

$ pip install pytask-stata

# or

$ conda install -c conda-forge pytask-stata

You also need to have Stata installed on your system and have the executable on your system's PATH. If you do not know how to do it, here is an explanation.

Similarly to normal task functions which execute Python code, you define tasks to execute scripts written in Stata with Python functions. The difference is that the function body does not contain any logic, but the decorator tells pytask how to handle the task.

Here is an example where you want to run script.do.

import pytask


@pytask.mark.stata(script="script.do")
@pytask.mark.produces("auto.dta")
def task_run_do_file():
    pass

When executing a do-file, the current working directory changes to the directory where the script is located. This allows you, for example, to reference every data set you want to read with a relative path from the script.

Dependencies and products can be added as with a normal pytask task using the @pytask.mark.depends_on and @pytask.mark.produces decorators. which is explained in this tutorial.

The decorator can be used to pass command line arguments to your Stata executable. For example, pass the path of the product with

@pytask.mark.stata(script="script.do", options="auto.dta")
@pytask.mark.produces("auto.dta")
def task_run_do_file():
    pass

And in your script.do, you can intercept the value with

* Intercept command line argument and save to macro named 'produces'.
args produces

sysuse auto, clear
save "`produces'"

The relative path inside the do-file works only because the pytask-stata switches the current working directory to the directory of the do-file before the task is executed.

To make the task independent from the current working directory, pass the full path as an command line argument. Here is an example.

# Absolute path to the build directory.
from src.config import BLD


@pytask.mark.stata(script="script.do", options=BLD / "auto.dta")
@pytask.mark.produces(BLD / "auto.dta")
def task_run_do_file():
    pass

You can also parametrize the execution of scripts, meaning executing multiple do-files as well as passing different command line arguments to the same do-file.

The following task executes two do-files which produce different outputs.

for i in range(2):

    @pytask.mark.task
    @pytask.mark.stata(script=f"script_{i}.do", options=f"{i}.dta")
    @pytask.mark.produces(f"{i}.dta")
    def task_execute_do_file():
        pass

pytask-stata can be configured with the following options.

stata_keep_log

Use this option to keep the .log files which are produced for every task. This option is useful to debug Stata tasks. Set the option via the configuration file with

[tool.pytask.ini_options]
stata_keep_log = true

The option is also available in the command line interface via the --stata-keep-log flag.

stata_check_log_lines

Use this option to vary the number of lines in the log file which are checked for error codes. It also controls the number of lines displayed on errors. Use any integer greater than zero. Here is the entry in the configuration file

[tool.pytask.ini_options]
stata_check_log_lines = 10

and here via the command line interface

$ pytask build --stata-check-log-lines 10

Consult the release notes to find out about what is new.