Re: [Gnumed-devel] Re: Lab importing Re: Bootstrap of (test) lab data

[Karsten Hilbert]

On Tue, Jan 22, 2008 at 06:49:34AM -0500, Dave Cramer wrote:

>> - document how to setup Mirth on Debian/testing in general
>>  - beyond what is documented already elsewhere
>>  - content might go to our Wiki under, say LegacyAppConfiguration
>>
>> - obtain sample hl7 data from Jim
>>
>> - make Mirth pickup those file from a directory
>>
>> - make Mirth "properly display" those files
>>
>> - match hl7 content to target (GNUmed) database fields
>>  - once you have a list of which data I can fairly
>>    easily match them up and add necessary fields
>
> Excuse me if this is a really naive question, but won't the hl7 data be 
> different for every lab ?

Content-wise, sure. But HL7 claims to sort of define a data
format so their will be synergies.

>> - export the Mirth channel for inclusion in our CVS
> Certainly we can export the channels,but if my assumption above is correct 
> then there are a number of channels required, no ?

One per (sufficiently different) lab. The use case currently
is Jim's lab(s) so we might add this to CVS as an example
for others. Or to the wiki if it really proves very
individual.

> Finally, as I eluded to in a previous message, it appears that the 
> "scripting" is ECMA javascript. Is this desirable ? Mirth seems like an 
> awful lot of infrastructure to take an hl7 file and parse it to put it into 
> your database ?
It is an awful lot. But, as you noted, parsing hl7 is a
major mess particularly when the number of players grows.
So, let's Mirth do it's thing, I dare say.

If you want to write a specific Jim's-HL7-to-GNUmed
importer, fine.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346