You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Many users of Latent Scope are probably working in some kind of jupyter notebook, and the data they want to visualize is already loaded as some kind of DataFrame (pandas, polars?).
It would be ideal if running the dataframe through the latent scope process was a very simple call to a function in the notebook that did not require opening up a separate web app. If we could put the setup process into a widget then the user could call a single function with their dataframe as input and results in a new dataframe with latent scope annotations. Perhaps we have another widget for the exploration visualization.
To implement this ipyreact looks promising, built on top of anywidget
Setup widget
Once #56 is completed, we can likely fit it into a widget. Especially if we design with this feature in mind.
A big consideration is how we deal with the file system, intermediate files in the process and passing/loading the results when the setup is complete.
Explore widget
Once #59 is completed, this should also be easier to implement. The widget format is more wide than tall which will likely suit our purposes. Some considerations would be what we want to export from the widget back into the notebook.
The text was updated successfully, but these errors were encountered:
Many users of Latent Scope are probably working in some kind of jupyter notebook, and the data they want to visualize is already loaded as some kind of DataFrame (pandas, polars?).
It would be ideal if running the dataframe through the latent scope process was a very simple call to a function in the notebook that did not require opening up a separate web app. If we could put the setup process into a widget then the user could call a single function with their dataframe as input and results in a new dataframe with latent scope annotations. Perhaps we have another widget for the exploration visualization.
To implement this ipyreact looks promising, built on top of anywidget
Setup widget
Once #56 is completed, we can likely fit it into a widget. Especially if we design with this feature in mind.
A big consideration is how we deal with the file system, intermediate files in the process and passing/loading the results when the setup is complete.
Explore widget
Once #59 is completed, this should also be easier to implement. The widget format is more wide than tall which will likely suit our purposes. Some considerations would be what we want to export from the widget back into the notebook.
The text was updated successfully, but these errors were encountered: