SendGrid

Informatics 117 : Team Optimize Prime

• Brenda La
• David Pham
• Eric Chou
• Jose Gomez
• Sheila Truong

Installation

$ go get -u github.com/govend/govend                 # make sure govend is installed
$ git clone https://github.com/blics18/SendGrid.git  # grab all the code from github
$ govend -v                                          # download all the dependencies in the vendor.yml file

Endpoints

• healthBF check the status of the server
• populateBF add data into the Bloom Filter
• checkBF check if data exists in the Bloom Filter
• clearBF clear the Bloom Filter

Curl command examples

• healthBF curl localhost:1234/healthBF
• populateBF curl localhost:1234/populateBF -d "[{\"UserID\": 123, \"Email\": [\"test@gmail.com\"]}]"
• checkBF curl localhost:1234/checkBF -d "{\"UserID\": 123, \"Email\": [\"test@gmail.com\"]}"
• clearBF curl localhost:1234/clearBF

Environment Variables

• BLOOM_SIZE size of the Bloom Filter
• BLOOM_PORT port that the server will run on
• BLOOM_NUM_USERS number of users that will be generated
• BLOOM_NUM_EMAILS up to this number of emails that each user will have
• BLOOM_NUM_HASH_FUNCTIONS number of hash functions
• BLOOM_NUM_TABLES is NOT configurable, because createTables.sql is hard coded to create only 5 tables.

Files & Description:

• client.go

  • Populate() - grabs data from MySQL, packages it as a JSON, and sends it to populateBF endpoint
  • Clear() - calls clearBF endpoint to clear the bloom filter
  • Check() - passes userid and list of emails to checkBF endpoint; calls checkBF to check if an input exists in the bloom filter

• runClient.go the client "main", and what we used to demo. Calls on all relevant endpoints (health, populate, check)

• server.go

  • healthBF() - checks to see if the server is up and running
  • populateBF() - retreives the JSON data, parses it, and populates the bloom filter with it
  • checkBF() - retreives the JSON data, parses it, and checks the data against the bloom filter. Also cross checks against the MySQL database if the data exists within the bloom filter
  • clearBF() - clears the bloom filter

• createTables.sql SQL file that creates the database and tables

• database.go creates and inserts randomly-generated data into the MySQL database

• server_test.go unit testing

• benchmark_test.go benchmarking for each endpoint except healthBF

• generateEmail.go generates random userIDs and emails

• data.txt file with randomly generated data (filled with suppression data and NON-suppresion data)

How to do unit testing:

1. Change into directory of server_test.go
2. In the terminal, type go test

How to do benchmarking:

1. Change into directory of benchmark_test.go
2. In the terminal, type go test -run=XXX -bench=name_of_function()

Graphite Visualization

1. Make sure that Docker is downloaded and running
2. Set up Graphite in docker:

docker run -d\
 --name graphite\
 --restart=always\
 -p 80:80\
 -p 2003-2004:2003-2004\
 -p 2023-2024:2023-2024\
 -p 8125:8125/udp\
 -p 8126:8126\
 hopsoft/graphite-statsd

3. Run server.go, and test using curl commands
4. Go to localhost in any browser to check and configure the metrics

Setting up MySQL

1. Create the docker container: docker run -v $(pwd)/createTables.sql:/root/createTables.sql --name bfmysql -it -d -e MYSQL_ALLOW_EMPTY_PASSWORD=1 -p 3306:3306 percona
2. Run this command on the container: docker exec -it bfmysql /bin/bash
3. Change directories to where the .sql file is stored: cd /root
4. Run the file that creates the schema with mysql < createTables.sql

Additional Notes

• Hits: Number of ACCURATE predictions. This has nothing to do if the value is in the Bloom Filter or not.
• Misses: Number of INACCURATE predictions. This has nothing to do if the value is in the Bloom Filter or not.