Rabobank Customer Statement Processor

"Rabobank receives monthly deliveries of customer statement records. This information is delivered in two formats, CSV and XML. These records need to be validated."

How to use

# install dependencies
npm i

# run tests
npm run test

# build
npm run build

# install the CLI globally
npm install -g .

# run the CLI
rabo -h

Simply run the CLI in a folder containing the input files (XML or CSV). The CLI will scan the folder and generate a JSON report of the failed transactions in the same folder. For more options, see the help menu (rabo -h).

You can also run the CLI with npx:

npx rabo

The challenge

The challenge is to implement an I/O API that can process the customer statement records and generate a report on any invalid records.

The customer statement records can be delivered in two file formats: CSV or XML. Records within these files need to be parsed and validated.

Validations:

1. all records should have a unique transaction reference.
2. all records should show a balance that matches the sum of all the mutations to the initial balance.

Any record that does not meet these requirements should be written to a report.

Approach

My approach consists of three steps:

1. Read the input file and parse the records.
2. Validate the records.
3. Generate a report.

Requirements and missing requirements

As this is a challenge, there are some requirements missing. Additional to the requirements I received initially, I was also told:

"A requirement we would put to you is that the report should be easy to use for further processing. Think about what is necessary for a user/another application to best navigate this report, what information should be in it and how can it best be presented and used."

Statements and assumptions

1. My goal with this is to create a simple API that could easily be implemented in other applications.
2. I will not implement a UI as the challenge is to create a processor that generates a report that can be used elsewhere. 1
3. I will use a command line interface to demonstrate the functionality and write the reports.
4. Based on the additional requirement, I will use a JSON format for the report as a default.

1: I imagine, and am working under the assumption, that a statement processor would more likely exist in an automated environment, e.g. as part of a pipeline, rather than be implemented in a UI. Of course, it would still be easy to support a manual file upload too.

Focus points

1. Code quality
2. Testing and coverage
3. Type safety (TS) and documentation (JSDoc)