-
-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Decoupled scda #534
Decoupled scda #534
Conversation
Code Coverage Summary
Diff against main
Results for commit: 137fcc7 Minimum allowed coverage is ♻️ This comment has been updated with latest results |
Code Coverage Summary
Diff against main
Results for commit: b071192 Minimum allowed coverage is ♻️ This comment has been updated with latest results |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hey @kartikeyakirar I left a few general comments that I would like you to apply in multiple places. This is good enough, but requires small changes.
Also, can you put the code that you used to create datasets in dev/data-creation.R or in data-raw/data-creation.R file? And add this file to .Rbuildignore? See this section . This way we can track properties of data and we can recreate data in the future if they need slight changes
@kartikeyakirar one more thing. Look at this PR where we decoupled |
I think the process of data creation will reintegrate SCDA dependencies, as data is stored from the SCDA package. This is consistent across all packages. However, it is worth considering the possibility of consolidating all the data into a single package. While the current approach is in use, it may be beneficial to have a discussion about centralizing the data within a specific package. |
@kartikeyakirar this is fine. Just use the code and then comment it out complitely so we can track how the data was created. Can you also document the |
I initially thought to format them but there are more or less 55 cols in ADSL itself. I don't have descriptions for all of them as well. |
Ok, so let's forget about the format section
…On Tue, Jul 11, 2023 at 11:21 AM kartikeya kirar ***@***.***> wrote:
I initially thought to format them but there are more or less 55 cols. I
don't have description for all of them as well.
—
Reply to this email directly, view it on GitHub
<#534 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/A74AIEPGHCIQGHJLBQUAYLLXPULKFANCNFSM6AAAAAA2B2TEVM>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Yes , for this I will raise PR along with teal.data to remove downstream dependencies. |
Great, since the merge here needs to happen at the same time as for scda/scda.2022 |
…eering/teal.modules.general into decouple_scda@main
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
requested changes were not solved
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@kartikeyakirar a couple more minor comments.
1
Mostly, are you able to append init
functions with proper package prefix teal
? The same for modules/module
function? In all examples? The same goes for actual usage of tm_g_*
and tm_a_*
functions. If they are used, are you able to append package name in front of them (teal.modules.general
)?
2
One more though, are you able to rewrite examples, so that we only use rADSL and rADLB? This way we have less datasets created and we can try removing that on other packages, so that we use only those two. This is just a loos thought and I'd like your honest opinion on that.
@gogonzo @kartikeyakirar do we need to create datasets inside this package ( |
@m7pr @kartikeyakirar let's not make links between packages because of data. Ideally in the future most of the packages should be independent from each other and data-dependency would be another thing to decouple. |
The examples were created using these datasets, and choices and selections were made accordingly. It is essential to thoroughly examine all these datasets to understand how they are used in the examples. Additionally, the RDA files' overall size was quite small, so I chose not to refactor them. What is your thought @gogonzo? We also have eight datasets in teal.data. |
adding namesapce call :: for teal.modules.general and teal.tranform
@kartikeyakirar let's leave those 5 datasets for now |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
great job. team work makes the dream work
A follow-up after #139 that closed #133 After I had a chance to do a scda decoupling for goshawk insightsengineering/goshawk#198 and had a chance to review scda decoupling for teal.modules.general insightsengineering/teal.modules.general#534 I realized some changes need to be applied in PRs that were already merged in other packages. Main changes: - I prepend dataset names with package names as now we are having the same data in multiple packages (`teal.transform::rADAE` and `teal.modules.general::rADAE` for example) - added a `data-raw/data.R` file to show how the `data/` folder was created - extended `.RBuildignore` file to omit `data-raw/data.R` while building the package
fixes #532
Related PRs
insightsengineering/scda#128
insightsengineering/scda.2022#127