Re: [Gnumed-devel] Demo data

[Karsten Hilbert]

On Fri, May 31, 2013 at 10:40:03PM +0000, Jim Busser wrote:

> Instead of identifying what to bootstrap, is it more
> practical to identify what *not* to bootstrap, and how most
> easily to enable this selectivity?

Likely, yes.

> 1) does one disable the following (except if it matches one's own country)
> 
>       in:
>             
> ~/Downloads/gnumed-server.18.4/server/bootstrap/bootstrap-latest.conf
> 
>       the lines:
>             bootstrap-de.conf
>             bootstrap-es.conf
>             bootstrap-ca.conf
>             bootstrap-au.conf
>             bootstrap-test_data.conf

Yes.

> 2) does disabling the above line 'bootstrap-test_data.conf' enable to avoid 
> to bootstrap, from
> 
>       ~ /gnumed-server.18.4/server/sql/test-data
> 
>       the following:
> 
>         gmDemographics-Person-test_data.sql
>         test_data-Kirk-procedure-dynamic.sql
>         test_data-Leonard_McCoy.sql
>         test_data-Christine_Chapel.sql
>         test_data-Kirk-vaccinations.sql
>         test_data-Spock.sql
>         test_data-James_Kirk.sql
>         test_data-Kirk_hospital_stay-dynamic.sql
>         test_data-USS_Enterprise.sql
>         test_data-Julian_Bashir.sql
>         test_data-Kirk_medication-dynamic.sql
>         test_data-lab_regression.sql
>         test_data-Kirk-diagnostic_certainty-dynamic.sql
>         test_data-Kirk_multi_results-dynamic.sql

Yes.

Later upgrades, however, do not quite so clearly separate
some of the demo data from the rest of the upgrade scripts.

So, one would need to run that and identify where it fails
and properly deal with that. Patches to separate out those
parts into standalone scripts are welcome.

> 3) some of the same .conf files (as listed above) appear also in
> 
>       ~/gnumed-server.18.4/server/bootstrap/redo-v2.conf
> 
> … however I was unable to find 'redo-v2.conf' referenced inside any of the 
> scripts and so am unclear of the purpose that it serves (or whether it is 
> deprecated?)

It allows to setup a v2 database and stop there. You will
notice that there is no *-v1.conf anywhere to be seen (as it
was an internal version never version tracked and never
release but, nonetheless, run at my parents office
successfully for many years. You will also notice many
consecutive update_db-vX_vX+1.conf used to bring databases
up to the latest version.

> 4) how to avoid to bootstrap the test accounts in 
> 
>       ~/gnumed-server.18.4/server/sql/gmTestAccounts.sql

Comment out the parts you don't want.

Be careful not to comment out at least one acocunt or else
you won't have any access, initially, unless you manually
create a user, which is fairly tedious.

If you want to forego the whole gmTestAccounts.sql you'd
find that name inside one of the .conf files and remove it
from the list of .sql files (move it out of the list option
and comment it out - you cannot comment it out
*inside* the list option).

> 5) how does GNUmed know or decide whether to bootstrap
> from each of the subdirectories within sql/country.specific
> or just one subdirectory (or no subdirectories)?

It doesn't. It just does what it is told by the .conf
scripts.

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346