Juan Salamanca - Recruiting Test

Question A

Your goal for this question is to write a program that accepts two lines (x1,x2) and (x3,x4) on the x-axis and returns whether they overlap. As an example, (1,5) and (2,6) overlaps but not (1,5) and (6,8).

Solution

Tests

Example

from juan.point_a import check_overlap
a = (3, 8)
b = (6, 11)
output = check_overlap(a, b)
print(output)

Observations

1. For every input vector a or b, it does not matter if a[0] is greater or less than a[1]

Question B

The goal of this question is to write a software library that accepts 2 version string as input and returns whether one is greater than, equal, or less than the other. As an example: “1.2” is greater than “1.1”. Please provide all test cases you could think of.

Solution

Tests

Example

from juan.point_b import check_version
v1 = '10.1.3.0.0.0'
v2 = '10.1.3'
output = check_version(v1, v2)
print(output)

Observations

check_version supports alpha and beta versions with literals a or b at the end of every version string.

For example '10.1.3a' is a valid version code

Question C

We want to optimize every bits of software we write. Your goal is to write a new library that can be integrated to our stack. Dealing with network issues everyday, latency is our biggest problem. Thus, your challenge is to write a new Geo Distributed LRU (Least Recently Used) cache with time expiration. This library will be used extensively by many of our services so it needs to meet the following criteria: 1 - Simplicity. Integration needs to be dead simple. 2 - Resilient to network failures or crashes. 3 - Near real time replication of data across Geolocation. Writes need to be in real time. 4 - Data consistency across regions 5 - Locality of reference, data should almost always be available from the closest region 6 - Flexible Schema 7 - Cache can expire

Solution

The proposed solution is a composition of two flask web apps: routes and caches. Routes act as a sharding component routing to the nearest region cache given a gps point. Caches App just receives objects and bring them to its users. Both apps are backed with a redis database for high performance and low latency due to its in-memory nature.

The routes app must be located as a central component where every user needing a cache end-point will query where to send its objects. After retrieving a region cache end-point, the user will send their objects to it and retrieve her objects as many times as she requires.

Criteria compliance:

Simplicity: As simple as send a rest api request over HTTP

Resilient: Resilient to crashes by its error handler design, but not resilient to network failures at request-response time. It could be solved with an event broker such as RabbitMQ, but in a next time.

Near real time replication of data across geolocation & Data consistency across regions: Every cache region end-point is backed by a redis database which one is capable of cluster deployment bringing consistency and high availability.

Locality of reference: Every cache region must be deployed over an independent network location.

Flexible Schema: The cache admin can add as many regions as possible to the routes app

Cache can expire: Cache expire time must be configured as a OS Environment variable at deployment time.

Tests

Examples

a. Get the nearest region cache end-point which distance is less than 600 km from a geo-location (lat, long)

curl -X GET "http://localhost:8080/v1/routes?lat=38.561387&long=-121.498287&radius=600&unit=km"

response must come in json format with details such as region name, lat, long and url to access

b. Add a new cache region end-point to the routes end-point

curl -H "Content-Type: application/json" -d '{"region": "AU", "lat": "-37.810745", "long": "144.965207", "url": "http://au.mycompany.com/v1/caches"}' -X POST "http://localhost:8080/v1/routes"

c. Put a object to a cache region with key juan and content body {"name": "juan", "id": "1000"}

curl -H "Content-Type: application/json" -H "X-Key: juan" -d '{"name": "juan", "id": "1000"}' -X POST "http://localhost:8181/v1/caches"

If the key already exists, the object will be overwritten with the new content.

d. Retrieve the object with key juan from a given region cache end-point

curl  -X GET "http://localhost:8181/v1/caches/juan"

The retrieved object must come with the same content-type as it was posted originally

Observations

For both web apps (routes and caches) config variables are passed as OS Environment variables as described in router/config.py for example.

TODO

1. Add feature to remove cache regions with a bad status from the routes app
2. Add auto register of cache region over the routes app
3. Add authentication and authorization
4. Make cache expire time a property of every cached object