Export files to s3 and crons to scripts

[antoineludeau]

WARNING - pre-requisite : before merging this PR, need to have deployed the build-addok-bundle cloud service.

Context

In order to make ban-plateforme cloud ready, we need to modify some mechanisms that won't be possible on the cloud. This PR is targeting the file storage system. For now, the storage used in the server storage system. We need to migrate to an S3 storage system. As a consequence, we need to migrate the way export files are created and stored. The export files targeted are :
1- Department files (csv, csv-bal, csv-with-ids, addok)
2- National files :

• csv : adresses-france.csv.gz, lieux-dits-beta-france.csv.gz
• csv-with-ids : adresses-with-ids-france.csv.gz, lieux-dits-with-ids-beta-france.csv.gz
• csv-bal : adresses-france.csv.gz
• addok : adresses-addok-france.ndjson.gz, addok-france-bundle.zip

Enhancements

1- Departement files : modify the current behavior that write the files on the server file storage to a behavior that does not right anything on the server but that upload the files on a S3. In order to work locally in dev mode, this PR also adds a "minio" container to the docker-compose.

2- National files : modify the current behavior using crons outside ban-plateforme (on the server) to a behavior that starts scripts using crons on the ban-plateforme worker.

How to test

1- Clone this branch.
2- Enter the right values for the following env variables :

MINIO_ROOT_USER=minio
MINIO_ROOT_PASSWORD=minio123
MINIO_PORT=9000
MINIO_FRONT_PORT=9001

3- Start the docker compose :
docker compose up --build -d

4- For department export files, start :
yarn dist

5- For national files, start :
node scripts/create-and-upload-national-files-to-s3.js