ds_problem

Data science problem: Corrupted Ledger

To install:

source install.sh

This will leave you in the virtual environment. If you open a new shell you will need to do:

source venv/bin/activate

To generate the simulated legder and corrupted ledger:

python run_python_script.py simulate

To check any corrected ledger file against the original ledger:

python run_python_script.py validate path_to_your_corrected_ledger_file score

This will print out the validation info such as number and percent of lines that are correct and the amount recovered. The percent recovered is the thing we are trying to optimize for.

You can only do that if you actually have the ledger file. An applicant taking this coding test can do this instead:

python run_python_script.py validate.py path_to_your_corrected_ledger_file

This will check that the file has the right format but not actually check how correct it is. File must have the same number of lines as original ledger and the same three fields and they must sum up to the same amount. Only the assignment of the numbers can be different.

This also includes three solutions. You can run them like this

python run_python_script.py solution_1

python run_python_script.py solution_2

python run_python_script.py solution_3

You can run everything and also make some plots with

python run_python_script.py all

Think of that like an integration test for everything.