create Anchor Screen Dashboard #243
Conversation
| @@ -574,3 +575,45 @@ def get_column_types(): | |||
| return cols | |||
|
|
|||
| return DataTableData(get_column_types, get_data) | |||
|
|
|||
|
|
|||
| def get_anchor_screen_metadata_table(): | |||
There was a problem hiding this comment.
oh wow. This is a lot more hassle then I was expecting you to have to go through.
I was imagining the react component could just fetch the data from breadbox. (I was actually just having a similar conversation with Jessica yesterday and advocated for bypassing the portal backend.)
I know this is historically how we've handled tables, but do you see advantages to keeping this pattern? It has always struck me as extra complicated.
There was a problem hiding this comment.
The way WideTable works currently wants a static link to point to for the CSV download. I didn't want to mess with that because there's something nice about it being a proper resource. We have some code to transform JSON to CSV on the frontend but it's not particularly well-written. So I just stuck with the tried and truth method.
| <p> | ||
| Below is a table of the anchor screens which have data analyzed and | ||
| loaded into the portal. The “volcano” links will show a volcano plot | ||
| of the differential analysis produced by “Chronos compare.” The | ||
| “scatter” links will show a scatter plot of the drug arm vs the | ||
| control arm. | ||
| </p> | ||
| <p> | ||
| <b>Note</b>: the screens shown in the scatter plot are from | ||
| processing all screens together with Chronos, whereas the Chronos | ||
| compare volcano plots processed only the anchor screens. As a | ||
| result, the differential effect size in the volcano plot should be | ||
| similar but not exactly the same as the difference shown in the | ||
| scatter plot. | ||
| </p> |
If it was set to a given_id and you tried to select a different feature/sample, it would get confused.
Asana:
https://app.asana.com/1/9513920295503/project/1165651979405609/task/1209952772841305