Open PRijnbeek opened 3 years ago
If you are running simple counts, these are simple SQL scripts at the moment - it is not wrapped into R code (even though we are planing to wrap everything into R wrapper at some point).
It has all supported permutations because Data Node is not required and people can still run it via their R enviroment and send the results back - hence, there is no guarantee that SQL Renderer to be on a server side. But these SQL dialects are generated using SQL Renderer
Since the execution is manual, the following is the workflow:
1) Submit a job to the data node 2) Data node admin gets a notification, accepts and downloads the submitted source code 3) Data node admin executes the code, reviews the results and uploaded it back and set the status to approved 4) Analysis is marked as ready
All these are described in this OHDSI ARACHNE slide deck that is stored in EHDEN Sharepoint in mode details
Thanks yes I know the offline workflow did that.
So in that case the user has to take the SQL fill in all the schema's etc manually and then copy a count in a file and upload?
R wrapper is quite simple to make if we want it
If you select the sql file as EntryPoint than it runs the SQL file it seems. This needs to be explained in the Documentation
1) The follow files are generated:
Diabetes Cohort_T2DM Cohort Count_code.zip
Trying to understand how this works. It has the OHDSI SQL and in a zip file translated versions. Why are there translated versions needed is this not handled in R with DatabaseConnector?
2) There is no main.R that gets executed on the Database to create the cohort and return the count. How does this work?
3) When i press Submit i see this:
What process does this start? Please explain in detail. It is unclear if the study lead now has to do anything else. i assume that since i am the owner of EHDEN_SYNPUF i can click on the Execute green symbol (Synthia i cannot click since i am probably not the owner)
it now say Executing but what is the process that is now started? I think it is now running automatically in this data node (this one is in network mode) since after a while it changes to
However with 0 documents because there is nothing to run (see above)
I receive an message:
In this message it is not clear on which database this is requested. Since I am owner of multiple databases which one do i accept if I click that button? (seems i do not accept just acknowledge i saw the message, is that correct?)
I then see this:
However, this datasource is not connected to a datanode in network mode. Suggestion is to show for each database in which mode they operate. I need to know that there is a human in het middle in this case.
End of workflow for now since there is no code to run.