OHDSI Home | Forums | Wiki | Github

Quickest way to install fresh SynPUF CDM with vocabs?

Implementers
#1

Is there a quick way nowadays to get a minimal postgres install of a SynPUF CDM with vocabularies? Or is it still, download DDL, load from Lee’s SynPUF files, get Vocabs from ATHENA, load those…?

#2

Sounds like a great topic for a tutorial on the OHDSI infrastructure!!!

1 Like
#3

Hi @Patrick_Ryan / @Sigfried_Gold

Wondering if there are any updates here.
We’re looking to make reference to the SynPUF 1K data in demonstrating how a PLP study is executed.

Having synpuf hosted on the public version of ATLAS takes one part of the way there - i.e. to demonstrate cohort generation. Is there anyway to point to those hosted cohort tables / CDM from an RStudio instance to actually run a study?

Or would the only way to do this be to stand up our own postgres install? Has anyone done that, as far as you know, on a commercially available cloud service?

#4

Hi @Evan_Minty,

The automated execution of a study from ATLAS is not currently available, but our team is working on enabling this feature in the future ATLAS release. We will be enabling the Execution Engine (currently used in ARACHNE) that would enable a user to execute those PLP (and PLE) R generated scripts against databases registered in ATLAS.

#5

@gregk Thanks!

#6

I’ve put together a few scripts intended to help automate Synpuf-CDMV5 PostgreSQL setup using the premade data available from the OHDSI FTP.

Though still necessary to download the vocabulary files on your own (I’ve been using Athena, haven’t looked for static links to those files), the repository located here will take you a good bit of the way to your own Postgres CDM instance (if not all the way there).

Note that:

  1. I’ve only tested it on Ubuntu 16.04 virtual machines, and not to completion
  2. You’ll have to run ‘make’ from the directory where you download the project. Once make is finished, you’ll have to configure the installation to suit your needs
  3. You will have to interact with the terminal, specifically to run scripts as sudo, accept Java license terms, indicate you’ve downloaded the vocabulary, and indicate whether to download full or partial datasets
  4. I’m happy to add a bit of documentation if there’s interest

Pull requests are encouraged, don’t hesitate to fix anything.

t