Given a publically readable video file via https://
1. Using ffmpeg, while preserving type, stream from 1 second marker, saving 1 second of video local to harddisk.
2. Upload 4 thumbnails of that clip in same S3 folder as original video.
a. _640x1096.jpg (640X1096)
b. _medium.jpg (150x150)
c. _small.jpg (100x100)
d. __tiny.jpg (48x48)
3. Preserving public_read acl @ Amazon

• AWS_KEY
• AWS_REGION
• AWS_SECRET
• BUCKET
• CONVERT_COMMAND
• LOG_LEVEL
• PORT
• PROCESSED_DIR
• REQUEST_TIMEOUT
• S3_ACL
• S3_STORAGE_CLASS
• SQS_QUEUE
• STAGING_DIR
• TMP_DIR
• WEBHOOK_PORT
• WEBHOOK_TOKEN

• Steps below require: ENVKEY defined in your shell; ie exports ENVKEY=
• To test: echo $ENVKEY

Build container: docker build . -t du-thumbd

Begin processing docker run -it -e ENVKEY=$ENVKEY du-thumbd .droneup/process_queue.sh

Build container: docker build . -t du-thumbd

Submit to SQS Queue: docker run -it -e ENVKEY=$ENVKEY -e REMOTE_IMAGE="https://url_to_video/vid.mov" du-thumbd .droneup/add_thumbnail.sh

Or all in one: docker run -it -e ENVKEY=$ENVKEY -e REMOTE_IMAGE="https://url_to_video/vid.mov" $(docker build -t du-thumbd .).droneup/add_thumbnail.sh

Build Container docker build . -t du-thumbd

Execute: docker run -it -e ENVKEY=$ENVKEY du-thumbd .droneup/process_all.sh

curl --header "Content-Type: application/json" --request POST --data '{"token":"webhook","obj_url":"https://somedomain.com/folder/video.mov"}' https://localhost:3000

• Untested, should work... Using messagebody:

  resources: [ 'https://s3.amazonaws.com/s3_bucket_name/folder1/full_file_name.mov' ],
  descriptions: '[\n\t{\n\t\t"suffix": "tiny",\n\t\t"width": 48,\n\t\t"height": 48\n\t},\n\t{\n\t\t"suffix": "small",\n\t\t"width": 100,\n\t\t"height": 100,\n\t\t"background": "red",\n\t\t"strategy": "fill"\n\t},\n\t{\n\t\t"suffix": "medium",\n\t\t"width": 150,\n\t\t"height": 150,\n\t\t"strategy": "bounded"\n\t},\n\t{\n\t\t"suffix": "640x1096",\n\t\t"width": 640,\n\t\t"height": 1096,\n\t\t"strategy": "matted"\n\t}\n]',
  aws_key: 'your s3 aws key',
  aws_secret: 'your s3 aws secret',
  aws_region: 'your s3 region',
  remote_image: 'https://s3.amazonaws.com/s3_bucket_name/folder1/full_file_name.mov',
  sqs_queue: 'leave empty',
  bucket: 'your s3 bucket name',
  log_level: 'info',
  custom_logger: 'logger'```

### Overview of processing

  1. For given bucket, list all objects `aws s3 ls --recursive`
  2. Loop through all results, for each do the following
  3. If string has a `.`, assume its a file to convert.
  4. Create an authenticated http url to s3 object, for purposes of streaming
  4. Use ffmpeg and stream 1 second of video, capture thumbnail as .jpg to local container
  5. Upload generated .jpg, and copy to same folder at s3, and with same name but with .jpg appended.