Stream: pascal
Topic: import terminologies
Jose Costa Teixeira (Oct 25 2021 at 23:41):
After I get my snomed text files reconciled, I need to put then through the console
Jose Costa Teixeira (Oct 25 2021 at 23:41):
(to create a cache, in this case a SNOMED-BE cache)
Grahame Grieve (Oct 25 2021 at 23:42):
reconciled?
Jose Costa Teixeira (Oct 25 2021 at 23:43):
the BE release comes as an extension
Jose Costa Teixeira (Oct 25 2021 at 23:44):
Rob explained that I need to combine some of the BE snapshot files and the corresponding INT snapshot files, in order to create a full SNOMED-CT_BE file
Jose Costa Teixeira (Oct 25 2021 at 23:45):
in doing that , there are some duplicates that need to be removed. supposing I need to do that, and supposing that i will do that, what do i do next to create a cache file for the server?
Grahame Grieve (Oct 25 2021 at 23:48):
I don't think you need to combine them yourself
Jose Costa Teixeira (Oct 25 2021 at 23:49):
ok I'll pick this up tomorrow. Meanwhile pinging @Rob Hausam
Jose Costa Teixeira (Oct 25 2021 at 23:50):
Perhaps what you mean @Grahame Grieve is that this exists ?
image.png
Grahame Grieve (Oct 25 2021 at 23:50):
oh no, not that combine. that's something else
Jose Costa Teixeira (Oct 25 2021 at 23:50):
ok
Jose Costa Teixeira (Oct 25 2021 at 23:51):
what i want is to get to the point i can enter it here:
image.png
Grahame Grieve (Oct 25 2021 at 23:51):
I think you just put the belgian in the top and then pick the intl it's based on. @Rob Hausam is that what we're doing?
Jose Costa Teixeira (Oct 25 2021 at 23:53):
ok meanwhile I'm trying that
Jose Costa Teixeira (Oct 25 2021 at 23:53):
After that, and after putting that on the server (i think it's just adding to config.json) how do I test it?
Jose Costa Teixeira (Oct 25 2021 at 23:54):
is there an easy request to see if the Belgian content is there?
Jose Costa Teixeira (Oct 25 2021 at 23:55):
perhaps I just do a $lookup ?
Grahame Grieve (Oct 25 2021 at 23:56):
or a validate-code, and nominate the snomed version
Jose Costa Teixeira (Oct 26 2021 at 00:00):
Jose Costa Teixeira (Oct 26 2021 at 00:00):
I have 32GB...
Grahame Grieve (Oct 26 2021 at 00:00):
64bit?
Jose Costa Teixeira (Oct 26 2021 at 00:01):
yes
Jose Costa Teixeira (Oct 26 2021 at 00:01):
I just built it. I think the build process builds for 64-bit, right?
Jose Costa Teixeira (Oct 26 2021 at 00:03):
(I'll be back tomorrow morning and pick this up)
Grahame Grieve (Oct 26 2021 at 00:08):
yes build si 64
Rob Hausam (Oct 26 2021 at 00:42):
@Jose Costa Teixeira This is the second part of what I want to work through with you. I’ll try to catch you when you get up - before I go to bed.
Jose Costa Teixeira (Oct 26 2021 at 07:52):
I'm back, perhaps we catch up in a few hours
Rob Hausam (Oct 26 2021 at 11:40):
Yes. I'm available now.
Jose Costa Teixeira (Oct 26 2021 at 19:27):
While I test the converted output (Thanks Rob for this), is there anything we can do so that the SNOMED import doesn't require 100+ GB of RAM?
Grahame Grieve (Oct 26 2021 at 19:29):
it doesn't require that for me, but nothing easy, no
Rob Hausam (Oct 26 2021 at 19:30):
What does it require for you?
Grahame Grieve (Oct 26 2021 at 19:30):
i have 64 and I haven't had to pay attention
Rob Hausam (Oct 26 2021 at 19:34):
Ok. The last version of fhirconsole that I tested the SNOMED CT import on (including today) is 1.9.363 (on Windows) - because that works, and that's where we were previously when I was able to work around the issue. I think that's pretty out of date now, and some of your recent changes may be helping with the memory utilization. I'll update to the latest version and will see how that is doing.
Grahame Grieve (Oct 26 2021 at 20:51):
well, what should I test with
Rob Hausam (Oct 26 2021 at 20:53):
I'm working on building the latest revision in Lazarus and then I'm going to check again.
Last updated: Apr 12 2022 at 19:14 UTC
