/hotjar-api

Primary LanguagePython

Welcome to the Backend Engineer Task!

This README file contains the instructions you will need to complete your task in the form of a User Story – the format regularly used by Hotjar engineers in our weekly sprints.

Instructions

  • This master branch already contains code which you will use for your task. Please add any new code to a new feature branch. When you’re done with the task, please submit a GitLab merge request to master.

  • If you cannot satisfy all the acceptance criteria within the given timeframe, simply prioritize what you feel is important and mention what you have excluded and why in your README file (task-README.md).

  • For the parts of the task which you feel are unclear, we encourage you to make your own assumptions as long as they are documented in your README file (task-README.md). However, if you have a question or concern that is blocking you from completing the work, please reach out to us via email.

A week after your task start date, a Hotjar engineer will review your merge request and you will receive an email within 48 hours notifying you of the next steps.

User Story

As a Hotjar user, I want to be able to search for error logs using advanced filters.

Background

The objective of this user story is to implement an API endpoint for an existing log storage service written in Python.

Error logs are stored in a database table with the following structure:

Field name Type
id Integer
created DateTime
browser String
page_url String
country String
message String

An endpoint, allowing for searches by browser or country, already exists.

Acceptance Criteria

  • Acceptance Criteria 1
    • The functionality must be implemented as an endpoint separate from the existing simple search endpoint.
    • The endpoint must accept a search expression represented by a json tree
      • The tree has three types of inner nodes:
        • grouping nodes (labelled as 'AND', 'OR')
          • a grouping node is used to combine multiple operation and control operation priorities
            • the AND node only allows a log message if it's allowed by all its children
            • the OR node only allows a log message if it's allowed by at least one of its children
          • a grouping node can have an arbitrary number of children (>1) and children can be of any type
        • negating nodes (labelled as 'NOT')
          • a negating node has only one child. The child can either be a grouping node or an operation node
          • a negating node only allows a log message if it's not allowed by its child
        • operation nodes (labelled as 'IS', 'CONTAIN')
          • operation nodes are the only parents for leaf nodes. The left leaf of an operation node is a field name. The right leaf is a value
          • operation nodes allow all log messages for which the result of applying the operation on the 'field' with the 'value' is true
    • Search tree examples:
      • {"CONTAINS":{"message": "error"}}
        • returns all records with the field "message" containing the word "error"
      • {"IS":{"browser": "chrome"}}
        • returns all records with the field "browser" being exactly the word "chrome"
      • {"NOT":{"IS":{"country": "Italy"}}}
        • returns all records with the field "country" NOT exactly equal to the word "Italy"
      • {"NOT": {"OR": [{"AND": [{"IS": {"browser": "safari"}},{"IS": {"country": "Germany"}}]},{"CONTAINS": {"message": "stacktrace"}}]}}
        • returns all records that DO NOT match one or both of these conditions:
          • the field "message" contains the word "stacktrace"
          • the field "browser" is the word "safari" AND the field "country" is the word "Germany"
    • The endpoint must return a JSON list of all records matching the search expression.
  • Acceptance Criteria 2
    • The existing repository must be refactored using software development principles that are adequate to the size and nature of the project

Important Notes

  • Solutions and future enhancements which would require too much time to be implemented can be described in your README file (task-README.md) instead.
  • In addition to Python, any other technologies may be used to complete the task as long as choices can be justified.

Utilities

Make sure you have a recent version of docker and docker-compose installed before running the utilities.

Good luck!