algorithmia

Algorithmia Interview, 5th May 2018

Introduction

Invite

Hi Jeffrey,

It was great speaking with you, and I'm excited about exploring a fit with Algorithmia. We'd like to invite you to complete our technical take-home exercise at your earliest convenience. You can find it here: http://algo.work/interview

Once you submit your solution to the marketplace, please click the link below and paste your link in the notes.

Please let me know if you have any questions.

Thanks!

Chris Armstrong Director of Engineering Algorithmia

Please submit here: https://app2.greenhouse.io/tests/613c9acbc43a2fb2e656adf65c8628ee

Details: http://algo.work/interview

Algorithmia Interview

The purpose of this exercise is to give an opportunity to all candidates to work through a realistic scenario as an engineer at Algorithmia. This is also an opportunity to showcase your ability working in your most comfortable environment. The goal here is to see how you approach engineering problems, your understanding of tools and frameworks, and ability to prioritize and creatively solve problems.

One of the trickiest, and yet most common, challenges in programming comes from dealing with parallelization and asynchronous code. We have a server already set up with some API endpoints to query. An example starting point is:

GET http://algo.work/interview/a Each endpoint returns JSON of the form:

{
  "children": [
    "http:\/\/algo.work\/interview\/b",
    "http:\/\/algo.work\/interview\/c"
  ],
  "reward": 1
}

Your challenge is to write an algorithm on the Algorithmia platform, in the language of your choice, that traverses the entire tree of API endpoints and returns the sum total rewards.
The input to the algorithm will be a starting url, such as "http://algo.work/interview/a".
For each page listed in the "children" field, you are to make a request to the given endpoint and recursively follow down the tree until you have explored all descendants of the starting page.
As you visit each page, add up the "reward" fields of the whole tree.
Your algorithm should return this sum as its result.

Notes:

Some of the endpoints take time to perform their job.
You should implement your algorithm in a way that explores the tree in as short a time as possible.
When creating your algorithm, be sure to make it closed source.
When publishing it, set the visibility to private.
Please take your time to implement a sound and efficient solution.
Once you're happy with your implementation, send us a link for review.

Initial Review: Checkout Endpoint

Approach

To checkout interview endpoint http://algo.work/interview, the simplest and fastest is to use available tools locally command-terminal.

I work on a macOS laptop, so what is default is different from other Unix environments, and it has its own utilities for environment and tools installation, brew.

Recommended Tools Setup

The following setup is assuming the reader is developing on macOS.

Brew install jq

jq is a lightweight and flexible command-line JSON processor. Reference: https://stedolan.github.io/jq/

By default, macOS does not come with jq.

Install jq using Homebrew:

brew install jq

Brew install Bash 4

Reference: http://wiki.bash-hackers.org/bash4

By default, macOS does not come with the most current bash.

Install latest bash using Homebrew:

$ brew update
$ brew install bash
$ sudo chown -R `whoami` /usr/local
$ brew link --overwrite bash
$ brew upgrade bash

Curl checkout of Algorithmia interview endpoint

Created a simple bash script using both curl and jq to explore Algorithmia interview endpoint:

scripts
├── algo_work_interview.sh
└── sources
    └── curl_algo_work_interview.sh

Bash `algo_work_interview.sh`

./algo_work_interview.sh --help

Usage: ./algo_work_interview.sh
 [-v|--verbose]
 [-h|--help]
 [--endpoint <string>, default: 'a']

./algo_work_interview.sh

{
  "children": [
    "http://algo.work/interview/b",
    "http://algo.work/interview/c"
  ],
  "reward": 1
}
0.197475

./algo_work_interview.sh --endpoint c

{
  "children": [
    "http://algo.work/interview/f",
    "http://algo.work/interview/g"
  ],
  "reward": 3
}
1.070582

Walk The Tree

fmt_endpoint: str: http://algo.work/interview/ENDPOINT

Request Endpoint: fmt_endpoint
Response Endpoints: array fmt_endpoint (bag, not unique, expect duplicates): "children"
Reward: float
- Expect negative values: "reward"
- Returned value is consistent with requested fmt_endpoint if asked multiple times.
Processing Time: time_total provided by curl
- Reference: https://blog.josephscott.org/2011/10/14/timing-details-with-curl/
- Processing time is consistent with requested fmt_endpoint if asked multiple times.

Request Endpoint	Processing Time	Response Endpoints	Reward	Calls	Total Rewards
a	0.197475	b, c	1	2	1
b	6.734861	d, e	2	2	3
c	1.070582	f, g	3	2	6
d	0.394293	NULL	4	0	10
e	0.938208	NULL	5	0	15
f	1.338257	h	6	1	21
g	2.137113	i (8)	7	8	28
h	0.195248	NULL	-1	0	27
i	0.256074	j (8)	0	64	27
j	0.203789	k (8)	0	512	27
k	0.194828	NULL	0.25	0	155

Name	Value
Total Expected Calls	592
Total Expected Rewards	155

Observations

Either this is a early shallow tree, or I am missing something:

Python Coding: Worker Factory

Resolved problem without Algorithmia libraries and created my own Worker Thread Class based upon Requests, https://pypi.org/project/requests/:

$ cd py_algo_worker
$ make local-run
======================================================
clean:
======================================================
======================================================
install
======================================================
======================================================
local-run
======================================================
{"asctime": "2018-05-05 22:55:17 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Example Worker: Starting"}
{"asctime": "2018-05-05 22:55:17 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Run: Starting"}
{"asctime": "2018-05-05 22:55:17 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Worker: Start"}
{"asctime": "2018-05-05 23:55:06 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/a': Request: Finished", "request_time_msecs": 229}
{"asctime": "2018-05-05 23:55:07 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/c': Request: Finished", "request_time_msecs": 1303}
{"asctime": "2018-05-05 23:55:08 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/f': Request: Finished", "request_time_msecs": 1226}
{"asctime": "2018-05-05 23:55:13 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/b': Request: Finished", "request_time_msecs": 6633}
{"asctime": "2018-05-05 23:55:13 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/h': Request: Finished", "request_time_msecs": 96}
{"asctime": "2018-05-05 23:55:13 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/d': Request: Finished", "request_time_msecs": 390}
{"asctime": "2018-05-05 23:55:13 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/e': Request: Finished", "request_time_msecs": 812}
{"asctime": "2018-05-05 23:55:14 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/g': Request: Finished", "request_time_msecs": 1693}
***
{"asctime": "2018-05-05 22:56:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 91}
{"asctime": "2018-05-05 22:56:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 92}
{"asctime": "2018-05-05 22:56:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 91}
{"asctime": "2018-05-05 22:56:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 93}
{"asctime": "2018-05-05 22:56:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 93}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "Example Worker", "version": "0.0.1", "message": "TOTAL REWARD: 155.0"}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "Example Worker", "version": "0.0.1", "message": "TOTAL CALLS: 592"}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Worker: Finish"}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Run: Completed", "exit_code": 0, "exit_desc": "Successfully completed", "exit_name": "Success", "run_time_msec": 14712}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "requests_worker", "version": "0.0.1", "message": "Example Worker: Success", "exit_code": 0, "exit_desc": "Successfully completed", "exit_name": "Success", "run_time_msec": 14712}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "requests_worker.worker_factory", "version": "0.0.1", "message": "Completed: Success", "exit_code": 0, "exit_desc": "Successfully completed", "exit_name": "Success", "run_time_msec": 14712}
{"asctime": "2018-05-05 23:55:20 -0700", "levelname": "INFO", "name": "requests_worker.worker_factory", "version": "0.0.1", "message": "Finished", "Exit Code": 0}

Name	Value
Total Calls	592
Total Rewards	155
Without Parallelism	69111 msec
With Parallelism	14712 msec

Python Coding: Worker Algorithm

$ cd py_algo_client
$ make local-run
======================================================
clean:
======================================================
======================================================
install
======================================================
======================================================
local-run
======================================================
{"asctime": "2018-05-06 01:05:23 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/a': Request: Finished", "request_time_msecs": 236}
{"asctime": "2018-05-06 01:05:24 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/c': Request: Finished", "request_time_msecs": 1222}
{"asctime": "2018-05-06 01:05:26 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/f': Request: Finished", "request_time_msecs": 1227}
{"asctime": "2018-05-06 01:05:30 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/b': Request: Finished", "request_time_msecs": 6549}
{"asctime": "2018-05-06 01:05:30 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/h': Request: Finished", "request_time_msecs": 101}
{"asctime": "2018-05-06 01:05:30 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/d': Request: Finished", "request_time_msecs": 395}
{"asctime": "2018-05-06 01:05:30 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/e': Request: Finished", "request_time_msecs": 591}
{"asctime": "2018-05-06 01:05:31 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/g': Request: Finished", "request_time_msecs": 1695}
{"asctime": "2018-05-06 01:05:32 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/i': Request: Finished", "request_time_msecs": 107}
***
{"asctime": "2018-05-06 01:05:37 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 94}
{"asctime": "2018-05-06 01:05:37 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 91}
{"asctime": "2018-05-06 01:05:37 -0700", "levelname": "INFO", "name": "requests_fortified", "version": "0.2.0", "message": "request_url 'http://algo.work/interview/k': Request: Finished", "request_time_msecs": 97}
{"asctime": "2018-05-06 01:05:37 -0700", "levelname": "INFO", "name": "Algo Rewards", "version": "0.0.1", "message": "TOTAL REWARD: 155.0"}
{"asctime": "2018-05-06 01:05:37 -0700", "levelname": "INFO", "name": "Algo Rewards", "version": "0.0.1", "message": "TOTAL CALLS: 592"}

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
AlgoRewards		AlgoRewards
images		images
py_algo_rewards		py_algo_rewards
py_requests_worker		py_requests_worker
scripts		scripts
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

algorithmia

Introduction

Invite

Details: http://algo.work/interview

Initial Review: Checkout Endpoint

Approach

Recommended Tools Setup

Brew install jq

Brew install Bash 4

Curl checkout of Algorithmia interview endpoint

Bash `algo_work_interview.sh`

Walk The Tree

Observations

Python Coding: Worker Factory

Python Coding: Worker Algorithm

About

Releases

Packages

Languages

License

jeff00seattle/algo-rewards

Folders and files

Latest commit

History

Repository files navigation

algorithmia

Introduction

Invite

Details: http://algo.work/interview

Initial Review: Checkout Endpoint

Approach

Recommended Tools Setup

Brew install jq

Brew install Bash 4

Curl checkout of Algorithmia interview endpoint

Bash algo_work_interview.sh

Walk The Tree

Observations

Python Coding: Worker Factory

Python Coding: Worker Algorithm

About

Resources

License

Stars

Watchers

Forks

Releases

Packages 0

Languages

Bash `algo_work_interview.sh`

Packages