🌐 Unique Geohash 🌐

Requirements

• make (if not present, apt install make on Linux Debian or brew install make on Mac)
• the current sources (python-test-sebastienhoarau)

A Makefile provides convenient shortcuts for most tasks.

Setup

Create a dedicated Python virtual environment, ie. with conda:

conda create -n geo-transformer python=3.9 pip

Activate the virtual environment:

conda activate geo-transformer

Install minimum requirements and package locally with:

make install

at the root of the project.

Usage

Change directory to geo_transformer directory:

cd geo_transformer

then run:

python app.py data/test_points.txt.gz

to use the provided sample data and print output to the console.

You can also use any file respecting the same schema as the one provided, compressed in gzip format for the INPUT_FILE argument.

List all available commands with:

python app.py --help

Project structure

python-test-sebastienhoarau
├── LICENSE
├── Makefile
├── README.md            # this file
├── geo_transformer      # main package
│   ├── __init__.py
│   ├── app.py           # application entrypoint
│   ├── data             # sample application data
│   ├── io.py            # input/output functions
│   ├── models.py        # datas structures
│   ├── tests            # tests
│   └── transformer.py   # transformer functions (geohash encoding, unique prefix)
├── pyproject.toml
├── requirements-dev.in  # dev requirements
├── requirements-dev.txt # compiled dev requirements with pip-compile
├── requirements.in      # requirements
├── requirements.txt     # compiled requirements with pip-compile
└── setup.py

Python source code is formatted with Black.

Development environment

Similar to Setup, while at the root of the project, install development requirements and package locally with:

make install-dev

Run tests

Tests can be run with:

make test

The console prints tests outputs as well as code coverage.

Run benchmarks

Benchmarks can be run with:

make benchmark

Creating a standalone binary

A standalone binary for your platform can be created with:

make standalone

And will be found as geo-transformer in newly created dist directory. The produced binary does not require any Python installation or dependencies and can be run just like in Usage with:

./geo-transformer <input file>

A downloadable version (ELF / Linux only) is available in the Releases section.

Updating dependencies

The requirements files requirements.txt and requirements-dev.txt are generated with pip-compile from pip-tools.

• update requirements.in and requirements-dev.in as needed
• run pip-compile requirements.in and pip-compile requirements-dev.in to update the compiled requirements.

Original Problem Statement

Your task is to transform the set of longitude, latitude coordinates provided in the test_points.txt.gz file into corresponding GeoHash codes. For each pair of coordinates only the shortest geohash prefix that uniquely identifies this point must be stored. For instance, this 3 points dataset will store these unique prefixes:

latitude	longitude	geohash	unique_prefix
41.388828145321	2.1689976634898	sp3e3qe7mkcb	sp3e3
41.390743	2.138067	sp3e2wuys9dr	sp3e2wuy
41.390853	2.138177	sp3e2wuzpnhr	sp3e2wuz

The solution must be coded in Python and you can use any public domain libraries. It should work with any file respecting the same schema as the one provided. The executable must output the solution on stdout in CSV format with 4 columns following the structure of the example, ie:

lat,lng,geohash,uniq
41.388828145321,2.1689976634898,sp3e3qe7mkcb,sp3e3
41.390743,2.138067,sp3e2wuys9dr,sp3e2wuy
41.390853,2.138177,sp3e2wuzpnhr,sp3e2wuz

🤓 We value in the solution

• Good software design
• Proper documentation
• Compliance to Python standards and modern usages (eg.: PEP8)
• Proper use of data structures
• Ergonomy of the command line interface
• Setup/Launch instructions if required