Stream: committers
Topic: Reference or inclusion of source documents
Lloyd McKenzie (May 18 2020 at 02:14):
Continued from https://chat.fhir.org/#narrow/stream/179297-committers.2Fnotification/topic/Reference.20or.20Inclusion.20of.20Source.20Documents
Grahame Grieve (May 18 2020 at 02:16):
Any document where an immutable reference is required from an IG should be included in the IG itself.
Lloyd McKenzie (May 18 2020 at 02:16):
A source document is the same as any other binary, such as a graphic. The IGPublisher expects those to be posted in an input folder designated with path-pages. Using the standard HL7 template, the appropriate folder is source/images. Any content that's there gets copied directly into the output without further processing.
Grahame Grieve (May 18 2020 at 02:17):
unless it's a common type - e.g. pdf - it should be zipped up (reason is because the HL7 server (hl7.org, as opposed to build.fhir.org) only servers files for which it knows the mime types, and I don't want a maintenance hassle).
Lloyd McKenzie (May 18 2020 at 02:17):
The issue is that Trifolia wasn't set up to handle source files that aren't resources - and that aren't explicitly referenced by the ImplementationGuide. That's why they tried the "include the binary in the href) as a workaround.
Grahame Grieve (May 18 2020 at 02:17):
I would think that source/images is not the best name for this kind of content
Grahame Grieve (May 18 2020 at 02:18):
but yes, it primarily sounds like a Trifolia issue
Lloyd McKenzie (May 18 2020 at 02:18):
images is what the most common content is
Lloyd McKenzie (May 18 2020 at 02:19):
I presume that .xls or at least .xlsx is a recognized file format? (It must be given that the publisher itself spits out .xlsx files...)
Grahame Grieve (May 18 2020 at 04:46):
yes xlsx is
Saul Kravitz (May 18 2020 at 14:45):
Looks like this one will be done with sushi, so inclusion of documents directly in the IG isn't a challenge.
I think the files in question are all PDFs.
Lisa Nelson (May 19 2020 at 17:40):
@Saul Kravitz In our case, the file in question is not a pdf. It is an .xslx, so based on what Grahame and Lloyd confirmed, we should be able to do this without errors. I would just mention that .xslx would be a key use case because a data file with many columns can't easily be turned into a pdf. This is a case where it is very useful to offer the .xlsx as a downloadable sheet from within the IG.
@Cheng Liu Please note the proposed solution so we can confirm if this works for our IG.
Saul Kravitz (May 19 2020 at 20:56):
including the xlsx file in the IG publisher image directory, and referencing it from the narrative leads to the desired result -- the xlsx file is included in the IG, and can be downloaded via a browser
image.png
Cheng Liu (May 27 2020 at 16:18):
@Saul Kravitz I imported an .xlsx file and follow the format described in this page: https://trifolia-fhir-dev.lantanagroup.com/help/Import.html.
However, the imported xlsx file was converted to a valueset in Trifolia. How do I make the xlsx file downloadable? Thank you
Saul Kravitz (May 27 2020 at 16:33):
Can't speak to trifolia.
In the sushi/IG-publisher process, simply including the xlsx in the ig-data/input/images directory will cause it to be included in the IG. Then reference it from the narrative pages with an <href>.
Cheng Liu (May 27 2020 at 16:49):
@Saul Kravitz Got it. Thank you.
@Sean McIlvenna Could you help us to figure it out? Is there some way in Trifolia to import an .xlsx file and make it downloadable?
Sean McIlvenna (May 27 2020 at 17:53):
this isn't currently possible in ToF. I will add it to the backlog of things needing to be done, though. I'll let you know when I know more.
Sean McIlvenna (May 27 2020 at 17:55):
of course, having said that, you can always put the xlsx file in your GitHub repository, and the CI build should respect that it's there. ToF's GitHub export can be configured to ignore files in GitHub that don't exist in the ToF IG, so that it doesn't delete it.
Last updated: Apr 12 2022 at 19:14 UTC
