Some examples of CARTO integrations using Lambda AWS. These examples are implemented using Zappa so it's quite straight forward to deploy them on your own AWS account.

Clone the repository, create a virtual environment, install the dependencies and try zappa -h to check you have the tool-set installed.

$ virtualenv env
[...]
$ source env/bin/activate
(env)$ pip install -r requirements.txt
[...]
(env)$ zappa -h
usage: zappa [-h] [-v] [-a APP_FUNCTION] [-s SETTINGS_FILE]
             {certify,deploy,init,invoke,manage,rollback,schedule,status,tail,undeploy,unschedule,update}
             ...

Zappa - Deploy Python applications to AWS Lambda and API Gateway.
[...]

You also need of course an account on AWS and the aws command line tool with an account configured, more details [here](http://docs.aws.amazon.com/cli/latest/userguide/cli-chap-getting-set- up.html).

Source code · Blog post

This script converts responses from Flickr API to GeoJSON feature collections so you can feed them to CARTO as a Synchronized Table or as a normal (one shot) upload. It's quite straight forward to read, just go to the @app.route... line as the starting point and follow the procedure to parse the parameters. Main logic for Flickr is at the getPhotos function.

If you want to run this script on your own account you need to register your own Flickr application to get an API key and put into the zappa_settings.json file. You can also configure the default name of the file that the end point will put in the HTTP header. Then you only need to deploy it to AWS with:

(dev)$ zappa deploy dev
(python-dateutil 2.5.3 (.../carto-lambda-examples/env/lib/python2.7/site-packages), Requirement.parse('python-dateutil==2.6.0'), set([u'zappa']))
Calling deploy for environment dev..
Downloading and installing dependencies..
100%|█████████████████████████████████████████████████████████████████████████| 37/37 [00:12<00:00,  2.92pkg/s]
Packaging project as zip..
Uploading flickrtogeojson-dev-1482693879.zip (33.7MiB)..
100%|█████████████████████████████████████████████████████████████████████████| 35.3M/35.3M [01:04<00:00, 454KB/s]
Uploading flickrtogeojson-dev-template-1482693970.json (1.6KiB)..
100%|█████████████████████████████████████████████████████████████████████████| 1.60K/1.60K [00:00<00:00, 6.40KB/s]
Waiting for stack flickrtogeojson-dev to create (this can take a bit)..
 75%|██████████████████████████████████████▎                                  | 3/4 [00:13<00:04,  4.94s/res]
Deploying API Gateway..
Deployment complete!: https://you-api-entrypoint.amazonaws.com/dev

NOTE: You may need to tweak the configuration file if you are using different profiles on your AWS credentials, want to deploy in other zones, etc. Check the Advanced settings section of the official documentation.

This will create the Lambda function and the API Gateway entry and will give you a URL of your API end point, say https://you-api-entrypoint.amazonaws.com/dev. If you go to that url without any parameters it will download a GeoJSON file from the Project Weather Flickr Group with the last five entries.

From this point you can use this endpoint as a proxy to any Flickr API request that retrieves photos. You'll need to provide the same parameters that the method you are using. If you don't provide a extras parameter sensible defaults will be used.

The URL for the entry point has been exported as FLICKR2GEOJSON for convenience.

Retrieve the last 10 photos of the Your Best Shot 2016 group using the method [flickr.groups.pools.getPhotos](https://www.flickr.com/services/api/flickr.gro ups.pools.getPhotos.html):

$ curl -s "$FLICKR2GEOJSON??method=flickr.groups.pools.getPhotos&group_id=1546513@N25&per_page=10&file_name=flickr_foss4g" \
| jq ".features[].properties.title"

"Καινούριους τόπους δεν θα βρεις, δεν θά βρεις άλλες θάλασσες. Η πόλις θα σε ακολουθεί. Στους δρόμους θα γυρνάς τους ίδιους.."
"More the Merrier"
"kyoto."
"Donald Trump street art in Copenhagen"
"sea shells in the morning light"
"Soirée Lightpainting"
"e-m1 16#009"
"Banhof - Berlin"
"Hide And Seek"
"_DESVAN_bn"

Retrieve the last 100 most interesting pictures using the method flickr.interestingness.getList:

curl -s "$FLICKR2GEOJSON??method=flickr.interestingness.getList&per_page=100&file_name=flickr_interesting" \
> interesting.geojson

Using the last example you can define a CARTO Sync Table that will update every day with the last 500 images. Of course only a number of them have coordinates so every day the map will have a different number of photos.

https://team.carto.com/u/jsanz/builder/2df0ff3a-cacc-11e6-9fb5-0e3ff518bd15/embed

Set up a scheduled event that performs a maintenance task on a CARTO table.

TO DO

A proxy to allow some operations for anonymous users that would need an API key.

TO DO

Listen to Foursquare push API calls for real time maintenance of a CARTO checkins table.

TO DO