This lesson's interactive features are locked, please to keep using them

SDKs and S3

An SDK or "Software Development Kit" is just a collection of tools (often involving an importable library) that helps you interact with a specific service or technology.

AWS has official SDKs for most popular programming languages. They're usually the best way to interact with AWS services.

Don't roll your own crypto, and don't roll your own AWS SDK.

When you as a human interact with AWS resources, you'll typically use the web console (GUI) or the CLI. When your code interacts with AWS resources, you'll use the SDK within your code.

Assignment

You've escaped writing code for a while - this one will have you write a lot of code and might take a bit of time, that's okay!

Install the AWS S3 Go SDK

go get github.com/aws/aws-sdk-go-v2/service/s3 github.com/aws/aws-sdk-go-v2/config

Configure an S3 client in main.go
1. Add an s3Client field to apiConfig of type *s3.Client
2. Use config.LoadDefaultConfig to auto load the default AWS SDK config (the keys you set with aws configure)
  - As arguments, give it an empty Context and pass config.WithRegion(s3Region) to use the region that's set in your .env file.
3. Create a client with your config using s3.NewFromConfig
4. Assign the client to the s3Client field
5. Update the S3_BUCKET and S3_REGION variables in your .env file. They'll be saved into your apiConfig.
Complete the (currently empty) handlerUploadVideo handler to store video files in S3. Images will stay on the local file system for now. I recommend using the image upload handler as a reference.
1. Set an upload limit of 1 GB (1 << 30 bytes) using http.MaxBytesReader.
2. Extract the videoID from the URL path parameters and parse it as a UUID
3. Authenticate the user to get a userID
4. Get the video metadata from the database, if the user is not the video owner, return a http.StatusUnauthorized response
5. Parse the uploaded video file from the form data
  - Use (http.Request).FormFile with the key "video" to get a multipart.File in memory
  - Remember to defer closing the file with (os.File).Close - we don't want any memory leaks
6. Validate the uploaded file to ensure it's an MP4 video
  - Use mime.ParseMediaType and "video/mp4" as the MIME type
7. Save the uploaded file to a temporary file on disk.
  - Use os.CreateTemp to create a temporary file. I passed in an empty string for the directory to use the system default, and the name "tubely-upload.mp4" (but you can use whatever you want)
  - defer remove the temp file with os.Remove
  - defer close the temp file (defer is LIFO, so it will close before the remove)
  - io.Copy the contents over from the wire to the temp file
8. Reset the tempFile's file pointer to the beginning with .Seek(0, io.SeekStart) - this will allow us to read the file again from the beginning
9. Put the object into S3 using PutObject. You'll need to provide:
  - The bucket name
  - The file key. Use the same <random-32-byte-hex>.ext format as the key. e.g. 1a2b3c4d5e6f7890abcd1234ef567890.mp4
  - The file contents (body). The temp file is an os.File which implements io.Reader
  - Content type, which is the MIME type of the file.
10. Update the VideoURL of the video record in the database with the S3 bucket and key. S3 URLs are in the format https://<bucket-name>.s3.<region>.amazonaws.com/<key>. Make sure you use the correct region and bucket name!
11. Restart your server and test the handler by uploading the boots-video-vertical.mp4 file. Make sure that:
  - The video is correctly uploaded to your S3 bucket.
  - The video_url in your database is updated with the S3 bucket and key (and thus shows up in the web UI)

Run and submit the CLI tests.