Pull dictionary xml file from S3 into the Rule Manager

[rhystmills]

What does this change?

This PR gives the Rule Manager access to a dictionary XML file from s3, and parses it to a list of unique words.

More specifically, it:

• Enables the Rule Manager to fetch the lemmatised dictionary XML file from S3 when a user hits the refreshDictionaryRules endpoint. The list is parsed and converted to a list of unique words - for now this list is only returned to the user - in a follow-up PR we will convert the words to rules in our database.
• Adds tests for methods used in the above process.
• Replaces the setup-hooks script with a generic setup script that runs our dev-nginx setup and pulls the dictionary file for local use.
• Modifies our Docker yaml and the localstack init script to make the dictionary XML file available in our local S3 instance.

How to test

Run the rule manager locally.

Send a POST request to https://manager.typerighter.local.dev-gutools.co.uk/api/refreshDictionary - I use Postman to do this, using auth cookies from the network tab https://manager.typerighter.local.dev-gutools.co.uk/ in the browser (I think this just needs to be gutoolsAuth-assym but I tend to include PLAY_SESSION too).

This should return a list of unique words.

1. Are the words unique?
2. What happens if you do this before running the setup script? (you can delete the dictionary from ~/.gu/typerighter/ if you've already run it)
3. Does it work if you deploy to CODE?

[rhystmills]

Code looks great! Lovely to see tests. Tested locally, works as expected – I've made a couple of changes to init-aws.sh to get the paths working that I hope makes sense in 8e06f65.

I think line 9 was sort of right (this one):

awslocal s3 cp /etc/.gu/typerighter/typerighter-dictionary.xml s3://typerighter-app-local/local/dictionary/typerighter-dictionary.xml

... it's copying the file from a location that Docker copies the file to in the localstack container (etc/.gu...) as specified in the volumes bit of the docker-compose.yml - but I've realised the convention is /etc/gu rather than /etc/.gu so I'll update that.

I should have remove line 8 because it wasn't doing anything - the localstack container was failing to access composer which led me down the Docker volumes path. Line 8 was:

aws s3 cp --profile composer --region eu-west-1 s3://typerighter-app-dev/dictionary/typerighter-dictionary.xml ~/.gu/typerighter/typerighter-dictionary.xml

[rhystmills]

(tested that in CODE and locally - working nicely so merged now)

[jonathonherbert]

Ace, thanks @rhystmills! I've a memory that adding config to the user dir ~ is more common now (I can see composer, atom workshop and workflow in there), but unsure if others can confirm, it's totally fine your way too 👍