masaki925/friday_address_provider

Address Provider

FRIDAY code challenge: https://gist.github.com/frank-kutzey-friday/76f9bf6b5b8901f7a88cab059fe606c9

Overview

• Street notation depends on the address of each country/region
• Even in the same country, there are some different notation for same address
  • e.g. 3丁目 == 三丁目
• So the conversion process could be complex.
• To mitigate the complex, I've divided converters for each country/region, then contained complex logic into each converter.
• I've applied the Facade pattern so that the main program can use converters without knowing the complex logics, but just call the convert() method and get the MyData result.

How to use

You can run with make commands as below:

$ cd address_provider
$ make run ADDRESS="Winterallee 3"
poetry run python src/address_provider/main.py --address "Winterallee 3"
{'street': 'Winterallee', 'housenumber': '3'}

You can run test as well:

$ make test
poetry run pytest -v tests
========================================================= test session starts =========================================================
platform darwin -- Python 3.11.1, pytest-7.2.1, pluggy-1.0.0 -- ${HOME}/Library/Caches/pypoetry/virtualenvs/address-provider-lv8EGNWZ-py3.11/bin/python
cachedir: .pytest_cache
rootdir: ${HOME}/address_provider
plugins: describe-2.0.1
collected 5 items

tests/test_main.py::test_main[Winterallee 3-expected0] PASSED                                                                   [ 20%]
tests/test_main.py::test_main[Musterstrasse 45-expected1] PASSED                                                                [ 40%]
tests/test_main.py::test_main[Blaufeldweg 123B-expected2] PASSED                                                                [ 60%]
tests/test_main.py::test_main2[Am B\xe4chle 23-expected0] PASSED                                                                [ 80%]
tests/test_main.py::test_main2[Auf der Vogelwiese 23 b-expected1] PASSED                                                        [100%]

========================================================== 5 passed in 0.07s =========================================================

System structure & tools

• Poetry to manage the dependencies
• Pytest to write a unit test
• abc to define Abstract Model to have Polymophic classes
  • By using it, we can ensure that each derivative converter has the same I/F such as convert() takes in string to convert, and returns MyData as the result.
• Facade pattern for Open Closed principle.
  • By using it, we can define new converters following a new format of address notation without changing existing converters.

What I want to have in future

• Creating Detector that can detect which kind of notation is the address, and assign a suitable Converter.
  • We need to take into account the same address but in other country?
  • As far as the requirement, we don't have to. creating the Converter for each languages is one another option.
  • If possible, getting country code as additional information would be nice to have.
• Docker environment
• Define CI/CD process
  • Run the test each CI process.
  • Currently, we only have test_main.py but we can separate it into the one for each converter.
• Terraform or other IaC tools to manage the infrastructure

FYI

ChatGPT couldn't handle japanese address correctly :P