Create APIs for Aggregations and Joins on MongoDB in Beneath 15 Minutes

Rockset has teamed up with MongoDB so you may construct real-time apps with information throughout MongoDB and different sources. When you haven’t heard of Rockset or know what Rockset does, you’ll by the top of this information! We’ll create an API to find out air high quality utilizing ClimaCell information on the climate and air pollution.

Air high quality has been documented to impact human well being (sources on the backside). Specifically, ranges of particulate matter (PM), sulfur dioxide (SO2), nitrogen dioxide (NO2), the ozone molecule (O3), and carbon monoxide (CO) are measured in an effort to recommend an individual’s beneficial exercise stage open air. What results PM, SO2, NO2, O3, and CO ranges are topic to in depth research: scientists study temperature, humidity, site visitors congestion, wind gust, and so forth to see how the air high quality index (AQI) adjustments with respect to those pollution.

It’s necessary to notice that this can be a pattern app to point out how MongoDB can combine with Rockset and demo Rockset’s tremendous powers of constructing APIs. This isn’t a scientific challenge that’s meant to be exhaustive in its conclusion. Rather more information is required, and plenty of variables are usually not considered when doing the question. For extra on air air pollution, see under on sources. To leverage Rockset and MongoDB in your real-time purposes (scientific or not), proceed studying!

In subsequent tutorials, I can present how this dataset can probably be used to know how temperature and humidity impacts AQI of PM and O3. You’re additionally welcome to do the challenges on the finish of the tutorial to earn Rockset credit score, swag, and extra!

On this information, we’re going to:

1. Combine MongoDB Atlas and Rockset

   • Construct collections on MongoDB that may every map to Rockset’s collections
2. Construct a Python script that may repeatedly get the climate and air high quality information from ClimaCell (Powered by ClimaCell) and put it into MongoDB
3. Use Rockset’s Question Editor to write down a question on real-time information coming in from MongoDB
4. Create an API to the question we wrote
5. Execute the API in our app and print out a press release in regards to the air high quality

Setup

Let’s get began on MongoDB

1. After you’ve created an account on MongoDB Atlas, go forward and navigate to Database Entry → Customized Roles and click on on Add New Customized Function. The picture under is what actions and roles needs to be added for rockset-role. Rockset has safe read-only entry to MongoDB Atlas.

2. Navigate to the Database Customers tab and click on on Add New Database Consumer. Keep in mind the password right here, as a result of we might want to use it once more, once we create an integration on Rockset.

3. Go forward and create one other database consumer that has Atlas admin privileges. We might be utilizing this database consumer in our Python app. You’ll be able to identify this consumer yourName-admin. Be sure to bear in mind the password right here, as a result of we are going to use this in our Python app.
4. Navigate to the Community Entry click on on Add IP Tackle and whitelist these IPs:
   Embedded content material: https://gist.github.com/nfarah86/c6014ea1d60ec6113948d889afb16fdf

5. Navigate to Clusters and click on on Collections then click on on Create database. We’re going to create a weather_pollution_db and it’s going to have weather_data assortment.
   
   
   
6. Beneath the weather_pollution_db, there may be going to be a plus signal. Click on on the plus signal, and add one other assortment, air_pollution_data.

7. Return to Clusters and click on on Join and click on on Join your utility. Copy the string, as a result of we are going to use it in Rockset. Once we are in our Rockset account, the username is rockset-user and the password is the password you used once you created rockset-user. In our Python app, the username might be yourName-admin and the password related to that database consumer.

8. That’s it for MongoDB! Let’s go forward and write our Python app!

Let’s construct our Python app

1. Create a challenge folder, and in that challenge, create a file .env.

2. Within the .env file add this:

   • Mongo_URI=“<uri string>”
   • The "<uri string>" is your connection string from MongoDB. Be sure to exchange the username and password within the connection string with yourName-admin and the password you used once you created that database consumer.
   • It ought to look one thing like this:
     mongodb://nadine-role-admin:password....

3. When you use a virtualenv go forward activate a env for this challenge. Be sure to’re below Python 3.7 or greater.

   • I personally use Pyenv, however be at liberty to make use of no matter you need!

4. Set up python-dotenv:

   • $ pip set up python-dotenv

5. Set up [pymongo] and [dnspython]:

   • $ pip set up pymongo
   • $ pip set up dnspython==1.16.0
6. Inside our challenge folder, go forward and create settings.py This file ought to appear like this: Embedded content material: https://gist.github.com/nfarah86/f87a9d37f1f72bb2d4a73d9b73dc87b4.
7. Create one other file within the challenge folder referred to as mongo_config.py. It ought to appear like this: Embedded content material: https://gist.github.com/nfarah86/1fc7bc9987d27edbec0fa9b32be95163
8. Within the challenge folder, go forward and create file referred to as script.py. All we’re going to do is make certain our Python app is connecting to MongoDB: Embedded content material: https://gist.github.com/nfarah86/4d8e87ff6e70e1da1c017e80b8daeef2
9. Beneath Clusters, click on on the collections button. Go to weather_pollution_db and click on on weather_data. It’s best to see this:

10. Now that we all know we are able to insert information into MongoDB, let’s go forward and create a ClimaCell developer account and get an API KEY.

11. In settings.py go forward and add this:

    • CLIMACELL_API_KEY = os.environ.get('CLIMACELL_API_KEY')
12. I selected ClimaCell as a result of they provide realtime information for climate and air air pollution. We’re going to work with this api. They’ve completely different parameters that may be added to the request. You’ll be able to discover these right here.

13. In our challenge folder go forward and pip set up a number of libraries:

    • $ pip set up requests
    • $ pip set up timeloop

14. In script.py go forward modify the packages we’re going to make use of: Embedded content material: https://gist.github.com/nfarah86/a49cbaa033239c636ef4f3bbe1dca2d0

    • Timeloop a library that may run jobs at designated intervals.

15. Hold insert_to_mongo() and add this perform in script.py to get the climate information: Embedded content material: https://gist.github.com/nfarah86/d2e3cc9236547e2fa630fd368dfee994

    • That lat and lon correspond to Beijing.
16. Now, we’re going so as to add this perform to get the air high quality: Embedded content material: https://gist.github.com/nfarah86/c598dbea0274d43215f15c9f01eca672
17. We’ll modify insert_to_mongo() to appear like this: Embedded content material: https://gist.github.com/nfarah86/e43f4ad2d8f7e3ca4b8d761408bc853c
18. To ensure we’re operating repeatedly, write this: Embedded content material: https://gist.github.com/nfarah86/959d875ad5ffcc08e16e3bf25358385a
19. After, write principal() like this: Embedded content material: https://gist.github.com/nfarah86/831e295b663aceb93603d9986c815b43
20. This is a gist of what your script.py ought to appear like: Embedded content material: https://gist.github.com/nfarah86/85caee5b14639e238e34715094cc5436

21. Now, run:

    • $ python script.py to populate MongoDB.
22. Whereas the script is operating, let’s get began on Rockset.

Let’s get began on Rockset

1. Login to Rockset and navigate to the Integrations tab on the left. Click on on Add Integration. Click on on MongDB and click on on begin:

2. Test the primary field MongoDB Atlas. We’re going to call this integration Rockset-Mongo-Integration. For the username and password, go forward and put rockset-user and the password you utilize once you created this database consumer. Paste the connection string within the subsequent field and click on on Save Integration.

   • Every integration can be utilized to entry a number of databases and collections in the identical MongoDB cluster

3. Beneath Collections click on on Create Assortment. Choose MongoDB because the supply.

4. Click on on the rockset-mongo-integration.

5. We’re going to call our new assortment on Rockset weather_data_collection. This isn’t tied to MongoDB. Go forward and fill out the remainder of the web page with the database and assortment we created on MongoDB. We’re going so as to add 2 collections, however let’s begin with the weather_data from MongoDB.

   • You see, Rockset is in a position to hook up with MongoDB. You’ll be able to confirm what information might be ingested into the Rockset assortment on the right-hand aspect. If you’ve created a set and operating a data-driven app in real-time, Rockset will repeatedly sync with MongoDB so your information can have the most recent info.
   • Let’s click on Create on the backside.

6. Observe the identical steps, step 3-5, to create the gathering, air_pollution_data_collection. On the finish, it ought to appear like this:

• Word permissions may be modified within the MongoDB UI at any time with out the necessity to drop and recreate integration. Besides when username and/or password or connection string adjustments – then the consumer might want to drop and recreate the Rockst integration

Assemble a Question on Rockset

1. On the left bar, let’s navigate to the Question Editor.

2. On the tab if we write:

   • Choose * from commons.air_pollution_data_collection; we must always see some output:

  - Go forward and do that for the `weather_data_collection`

3. We’re going to write down this pattern question: Embedded content material: https://gist.github.com/nfarah86/2d9c5bc316d55cfd0fcf17b4ded9141f

   • We’re averaging the PM10 information and the climate temperature information. We’re going to hitch each of those collections based mostly on the date. When you observed the timestamp within the JSON, the date is in ISO 8601 format. So as to be a part of on the times (and eliminates the minutes, hours, and seconds), we’re going to do an extraction.
4. Run the question.

5. After we run this question, we need to embed it in our app, so we are able to notify our customers when the degrees fluctuate, and probably predict, based mostly on climate, what PM10 ranges might appear like the subsequent day.

   • We’re going to want much more information than what we’ve now to foretell based mostly on temperature, however this can be a begin!

Construct an API on our question on Rockset

1. On the highest nook, click on on Create Question Lambda. A Question Lambda is a solution to make an API endpoint to the SQL question you write. Within the Python app, we received’t have to write down client-side SQL, stopping safety vulnerabilities.

2. Give your Question Lambda a reputation and outline. Afterwards, it’s best to see some code snippets on the subsequent display.
3. Let’s navigate again on the Question Editor and write one other question to get present climate in a brand new tab. Typically we might get a null area, so let’s go forward and write this within the Question Editor: Embedded content material: https://gist.github.com/nfarah86/4581c6bc09d30045ae75a5f330a8d72f
4. Create one other new Question Lambda.
5. If we need to seize the code snippet or URL of the Question Lambdas we simply created, navigate on the left aspect menu to Question Lambda and click on on the lambda you created.

Execute APIs on our app

1. When you create a Question Lambda, you’ll see one thing like this:
   
   
   
   

2. There are two methods I’ll present how we are able to execute a lambda:

   • Make an HTTP Request
   • Rockset’s Python consumer (backside field the place my API is boxed out)

3. Make an HTTP Request:

   • Let’s go forward and make an HTTP request to get the current_weather information. Listed here are the steps to do that:
   • Go forward and set your ROCKSET_API_KEY in your .env. Import it in settings.py like we did earlier than.
   • On Rockset, navigate to the Question Lambda that has the current_weather question. Copy the question lambda endpoint.
   • We’re going to write down this perform that may make an HTTP request to that endpoint:Embedded content material: https://gist.github.com/nfarah86/3a0ef9b1524532247e3ea7c504489d23

4. Use the Rockset Consumer to ship a request:

5. Then, we’re going to show the consequence:Embedded content material: https://gist.github.com/nfarah86/a0d1e15319bc117ef55ce35187fb6480
6. We’re going to vary sample_job_every_120s()so as to add make_requests so we are able to execute the Question Lambdas and show the info:Embedded content material: https://gist.github.com/nfarah86/0a54e082c9026aa5c9940b24836d9c65
7. Write make_requests() so it appears to be like like this:Embedded content material: https://gist.github.com/nfarah86/dea06329b25887bb58a0ef74c4a12fb0
8. After you run the script, it’s best to see this:Embedded content material: https://gist.github.com/nfarah86/32b35bd3269fbd1701dc57252fa783e4
9. That’s it! This wraps it up for the MongoDB-Rockset Python App!

Challenge Code

You’ll find the total challenge, together with the SQL statements right here. You probably have questions in regards to the challenge, Rockset, or MongoDB, you may attain out in our group.

Sources:

Different MongoDB sources: