rapidpro-pull is an open-source command-line tool for pulling data from RapidPro servers, printing it in the JSON format and caching it in local or remote relational databases (any database supported by SQLAlchemy - e.g. PostgreSQL, Oracle, SQLite). It has been developed for UNICEF to create a foundation for a new ETL subsystem for IMAMD but it is a standalone tool which can be used independently.
To install and get started:
$ pip install rapidpro-pull $ rapidpro-pull --help
Usage:
rapidpro-pull --flow-runs --api-token=<api-token> [--address=<address>]
[--before=<before> --after=<after>]
[--with-contacts --with-flows]
[--cache=<database-url>]
rapidpro-pull --flows --api-token=<api-token> [--address=<address>]
[--before=<before> --after=<after>]
[--uuid=<uuid> ...]
[--cache=<database-url>]
rapidpro-pull --contacts --api-token=<api-token> [--address=<address>]
[--before=<before> --after=<after>]
[--uuid=<uuid> ...]
[--cache=<database-url>]
rapidpro-pull --help
Options:
--flow-runs download flow runs
--flows download flows
--contacts download contacts
-a, --address=<address> a RapidPro server [default: rapidpro.io]
-t, --api-token=<api-token> a RapidPro API token
-h, --help display this help and exit
--before=<before> download all older than ISO 8601 date/time
--after=<after> download all newer than ISO 8601 date/time
--uuid=<uuid> fetch objects matching UUID(s) (repeatable)
--with-flows download associated flows, too
--with-contacts download associated contacts, too
--cache=<database-url> use database-url as cache (store retrieved
objects in cache; retrieve objects from
cache instead of downloading from RapidPro
when possible)
Examples:
- rapidpro-pull --api-token=a-token -flow-runs >all-flow-runs.json
- Use a RapidPro API token a-token to download all flow runs and save them into a JSON file all-flow-runs.json.
- rapidpro-pull -t a-token --address https://rapidpro.kotarba.net --flow-runs
- Use a RapidPro API token a-token to download all flow runs from an alternative RapidPro service rapidpro.kotarba.net over HTTPS and print them in the JSON format.
- rapidpro-pull -t a-token --flow-runs --with-flows --cache=sqlite:////tmp/rp.db
- Use token a-token to download all flow runs and their associated flows. Do not download flows already cached in the provided SQLite database in file /tmp/rp.db and do not overwrite cached flow runs. Add all newly downloaded objects to the database for later processing.
- rapidpro-pull --flow-runs --with-flows --with-contacts --api-token=a-token
- Use a RapidPro API token a-token to download all flow runs together with all associated flows and contacts
- rapidpro-pull -t a-token --flows --after 2016-01-01T12:12:12.596000Z
- Use token a-token to download all flows newer than 2016-01-01T12:12:12.596Z.
- rapidpro-pull -t a-token --contacts --uuid=a --uuid=b --uuid=c
- Use token a-token to download contacts with UUIDs a, b or c.
Working on rapidpro-pull requires the installation of a small number of development dependencies (in addition to the dependencies required to just run the program). These dependencies are listed in tests_require in the setup.py file but one does not need to install them by hand unless one chooses to invoke the project test runners manually (see: alternative ways to run tests). In order to get started one may want to do the following:
$ # Create a virtualenv and activate it, e.g.: mkvirtualenv rapidpro-pull $ git clone https://github.com/system7-open-source/rapidpro-pull.git $ cd rapidpro-pull $ pip install --editable .
To use the alternative ways of running tests one needs to explicitly install the aforementioned additional dependencies (this step is optional and not required to run tests):
$ pip install --editable .[development]
The project has been developed using the BDD / outside-in TDD approach and there are two separate groups of tests: features and scenarios describing the high-level/system behaviour using the Gherkin syntax (and, underneath, Python), and the low-level unit tests (the author is not a mockist but a classicist which means that mocking and stubbing is used where it seems to make sense instead of everywhere ;) ). The provided unit tests ensure 100% code coverage (statement + branch). Apart from the coverage reports printed after each execution of unit tests, one can view the latest HTML report stored in the htmlcov directory.
The functional tests (features/scenarios) are found in the features/ directory. To execute them:
$ python setup.py behave_test # please use -b to pass arguments to behave $ behave # an alternative way of running tests, please see: behave --help
The unit tests are found in the tests/ directory. To execute them:
$ python setup.py pytest # please use -p to pass arguments to py.test $ python setup.py test # an alias for pytest $ py.test # an alternative way of running tests, please see: py.test -h
Alternatively, to run all tests on all supported implementations and versions of Python, one can just execute the following command:
$ tox
We use tox together with various continuous integration services to analyse the code quality and test rapidpro-pull on all supported platforms (Linux, MacOS, Windows) and on all supported implementations and versions of Python. The status of the current stable release can be easily checked by looking at the status badges at the top of this document (rapidpro-pull). For developers, the status of the develop branch is displayed in the Development section.
Please feel free to use this project issue tracker where appropriate, fork this repository and generate pull requests. The author can also be contacted via e-mail: Tomasz J. Kotarba <tomasz@kotarba.net>.
Special thanks to Robert Johnston (a crusading saint of UNICEF, always ready to fight dragons to save those in need) without whom this project would never be.