Elf M. Sternberg 2ce0a091fc | ||
---|---|---|
handy | ||
initial_data | ||
ranksearch | ||
static | ||
templates | ||
.gitignore | ||
README.md | ||
manage.py | ||
requirements.txt |
README.md
Handy
Handy is a simple Django app that provides the user with a working, if simple, search rank algorithm matching maintenance experts with the properties they've maintained. Handy starts with an initial dataset provided from a CSV file.
Installation
This script requires Python3 and Sqlite.
- Create a new virtual environment:
$ virtualenv --python=python3 handy
$ cd handy
$ source bin/activate ;# this is the bash instruction; use what your env demands
$ pip install django
$ git clone <this repository> handy
$ cd handy ;# Yes, twice: handy/handy.
$ rm db.sqlite3
$ ./manage.py migrate auth admin sessions
$ ./manage.py migrate ranksearch 0001_initial
$ ./manage.py recoverhandy
$ ./manage.py runserver
"Recoverhandy" is a custom command that imports the CSV into the database, creating owners, properties, maintainers, and jobs progressively. It does a lot of 'get_or_create' commands, so on SQLite it's pretty slow.
After entering the 'handy' app home, you can also
$ ./manage.py test
This will run a simple unit test that asserts both the database structure and client status are adequate.
Status
July 3, 2016: This project is completed. No new features are being considered. Bug reports will (probably) not be addressed.
Process
The first thing is to understand the data.
The problem:
- A maintainer starts with a score
- For the problem space, the score is artificial
- A rating is given to a maintainer on a given date
- More precisely: The maintainer gets a set of scores in order
- The maintainer's overall score is:
- score when there are zero jobs
- score + (sum ratings / 10) between 1 and 9 jobs
- (sum ratings / num sits) for 10 o more jobs
Tables:
- Property Owners
- Properties (owner_id)
- Maintainers
- Jobs (maintainer_id, property_id, start_date, end_date, rating, comments)
I'm big on slugs. I'm very big on navigable and bookmarkable URLs. I believe URLS are UX (user experience), and that readable URLs communicate to the user knowledge and reassurance that both parties know what they're doing. If a property owner comes to like a particular maintainer, that maintainer should have their own easy-to-understand URL the owner can always navigate to, to see if the maintainer is available.
I know it's bad form to write your own pagination kit, but I couldn't help myself; it's a bit of showing off, especially with the history thing in-lined.
Analysis
This was fun. I haven't worked in Django in almost three years, and the things that have changed since 1.4 and within Python 3 make development even more interesting than usual.
Postscript
If you're reading this and you can't figure out what this is really for, you probably don't need it.