Make it easier for everyone to install

[IsmailM]

Remove everything specific to the main setup.

• Remove NETLIFY_SITE_ID
• Replace S3_VCF_FILE_URL with the path to an included VCF file (or provide a curl command in the README on where to download from.
• Remove MAIL_USERNAME - i.e. make this SMTP completely optional...
• Include a local S3 compliant storage server (for data uploads etc.) - one example is MinIO - to use this we simply need to add an endpoint argument - which would point to MinIO instead of AWS.
• Create a installation/run script.

[alanwilter]

I'm playing with branch mig_auto and there I've moved NETLIFY_SITE_ID from public.env to private.env.

I hope to finally finish the changes in autocomplete.py so we can merge my PR.

[alanwilter]

We need a brainstorm here because the devil's always in the details.

• How can phenopolis browser be used by others with their own data?
  Currently we fill tables and rely on a VCF file for variant details.
• How others will fill their data in?
  Will others have to create their own VCF file to and use our import scripts to fill DB with their own data?
  In this sense, others could perfectly use GRCh38 assembly if their data are so.
• I've always a fan of the idea of getting completely rid of the VCF file and have variant data in a DB but I understand it may be challenging to achieve this. Shall we invest time to address it now knowing it may not pay off?

This is not a exhaustive list, please add yours if you feel like.