pkalita-lbl
commented
1 month ago Thanks for the feedback! Let me try to tease apart some of this and see what might be actionable.
it looks like we must select an environmental template - what of studies that require more than one sample type...like studies that happen at an interface - say water and soil, or soil and host-associated...can this be moved to the level of each sample, or can you select multiple (and be allowed to edit?)
This is a limitation imposed by the portal backend at the moment. However, that will be changing shortly. The following issues track changes to the portal and the app, respectively, which will allow a study to be associated with multiple environmental templates:
- https://github.com/microbiomedata/nmdc-server/issues/1367
- https://github.com/microbiomedata/nmdc-field-notes/issues/177
Once the changes for those two issues are deployed you will be able to select, say, both water and soil for a single study. In that case when you attach a new sample to that study the app will ask you whether you'd like to use the water or soil template for that sample.
name and ORCID can auto-populate - can you do the same with email (even if over-ridden by user). Perhaps if not, can you check the studies within the app with the same ORCID/user and autopopulate email....in case they mistype their email?
The portal backend does not collect user email addresses at the moment. All we know about the logged-in user is their name and ORCID iD. There is planned work to address that on the portal side:
My opinion is we should wait for the portal backend to sort that out and then utilize what the backend provides in the app.
Can you NOT make analysis/data type required? some simply collect for future use. Also some can collect for multiple uses, can you make this multiple choice? and perhaps add free text as well in case they want to use for something else? Should this not be linked instead to actual data generated instead, and have that data file point back to this sample?
This is one of those instances where NMDC being relentlessly schema-driven comes back to bite us a bit. The sample metadata entry screens are entirely driven by the submission schema. The submission schema marks slots (like analysis/data type) as required if we require them before finalizing a submission (i.e. before saying "here you go NMDC, I'm done with this, you look at it now"). However, as you're getting at, the app is mainly about collecting what you can in the field, assuming that you'll fill out the rest and finalize later via the portal.
Probably what we need to do is to annotate the schema in some way where we can explicitly mark slots (like, say, depth) as "you should almost certainly collect this in the field" as opposed to just dual-purposing the slot's required attribute.
The analysis/data type slot is already multivalued, so you should be able to select multiple options:
If you aren't seeing that behavior please let me know. As far as the other changes to that slot like allowing free-text entries or deriving it from generated data, that's a much bigger schema and workflow question that feels outside of the scope of the app to me.
is there any control or lookup to make sure that entries can be correct - i entered false info into broad-scale environmental context, and in local environmental context...
The three ENVO slots (broad-scale environmental context, local environmental context, and environmental medium) should (in theory) be offering you options to select from instead of asking you to type something in. This is a known bug:
oddly, i entered in the example given in environmental medium but it says it's invalid!
i like the use GPS feature - but when i entered geographic location by name, i again entred the exact example listed but it says invalid.
why does depth in meters not accept 0? hmmm...also doesn't accept 1, which is another example provided. but 0-1 works.
depth for soil and sediment samples requires a range (i.e. not just 0). Why? That's the way some soil scientist said it should be :grin: (https://github.com/microbiomedata/submission-schema/issues/95)
However the examples for all slots should definitely be valid. The examples come from the schema itself, so I created an issue in the submission schema repository to address this:
i seem to have lost my prior study/samples...oh well.
That is definitely not expected. If you can reproduce that issue I'd be happy to investigate that.
patrickchain
Bug summary
just a list of comments
Bug details
General questions for study creation: it looks like we must select an environmental template - what of studies that require more than one sample type...like studies that happen at an interface - say water and soil, or soil and host-associated...can this be moved to the level of each sample, or can you select multiple (and be allowed to edit?)
name and ORCID can auto-populate - can you do the same with email (even if over-ridden by user). Perhaps if not, can you check the studies within the app with the same ORCID/user and autopopulate email....in case they mistype their email?
sample creation - (soil template study) Can you NOT make analysis/data type required? some simply collect for future use. Also some can collect for multiple uses, can you make this multiple choice? and perhaps add free text as well in case they want to use for something else? Should this not be linked instead to actual data generated instead, and have that data file point back to this sample?
is there any control or lookup to make sure that entries can be correct - i entered false info into broad-scale environmental context, and in local environmental context... oddly, i entered in the example given in environmental medium but it says it's invalid!
Can you default to auto-fill when only one entry is possible - like in Ecosystem - the only thing possible is 'environmental' - and then same with subclasses of ecosystem....
i like the use GPS feature - but when i entered geographic location by name, i again entred the exact example listed but it says invalid.
why does depth in meters not accept 0? hmmm...also doesn't accept 1, which is another example provided. but 0-1 works.
Steps to reproduce
No response
Anything else?
app auto-updated, that's nice.
i seem to have lost my prior study/samples...oh well.
App version
No response
Device type
Apple
Operating system version
No response