Orbital Technical Assessment Solution

Submission by Éanna Morley

Deployed Solution

A fully deployed solution that can be run in the browser is available at: https://orbital-steel.vercel.app/

This consists of a simple NextJS, React frontend for displaying the data with a FastAPI python backend.

The web app is deployed to Vercel and the API is deployed with Render.

If there is existing parsed data, you can remove it by clicking the 'Reset' button and run the solution again by clicking the 'Parse' button.

You can search the returned parsed data using the search bar provided.

You can download the parsed data as a JSON file using the download button.

Quickstart

To run the solution, navigate to the backend folder and simply run:

python solution.py

To run unit tests, run:

python -m unittest test.py

Overview of Solution

This approach takes advantage of the fact that each of the fields we wish to extract begins a consistent number of characters from the end of the string. This is made visible by the trailing whitespace that pads each line. By reversing each line, we end up with right-aligned, fixed-width fields. This makes it possible to define a regular expression using a positive lookahead that reliably targets each field based on its fixed position relative to the new (reversed) start of the string. After extracting the fields, we can concatenate them appropriately to reconstruct the complete, structured information in a usable format.

As notes seemingly are always found at the end of the array of strings and begin with some variation of "NOTE", "NOTE 1", etc. we can easily identify them and treat them independently of the other fields.

Regular expressions are a highly performant and scalable solution that satisfy the criteria set out as part of this assessment.

Assumptions

We assume that cancelled leases are to be ignored for the purposes of this task.

Next Steps

• Optimise the performance of the list component in the web app. Due to the number of records, loading and rendering the data as well as searching is slow. Some approaches for doing this would be: Use a virtualised list for more performant rendering (e.g. @tanstack/react-virtual), implement cursor-based pagination with server-side searching capabilities.
• Implement the outstanding "Report Issue" feature whereby flagged records that have been improperly parsed for whatever reason can be sent to a LLM for fixing and/or reviewed by a human to assess what went wrong.
• Implement a proper database (e.g. PostgreSQL) for persisting and querying records.
• Implement API authentication.
• Build out a more comprehensive suite of unit tests to make sure the solution is robust to any potential edge cases.

Additional Information

If you would like to run the web app locally:

• Start the API:
  • Navigate to /backend
  • Run pip install -r requirements.txt
  • Run fastapi run main.py
• Start the app:
  • Navigate to /frontend
  • Run npm install
  • Run npm run dev
  • Open http://localhost:3000

API docs can be found at: https://orbital-ia4z.onrender.com/docs