Skip to content
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

Create a new Plotly Visualizer #1308

Closed
umesh-timalsina opened this issue Nov 11, 2019 · 3 comments
Closed

Create a new Plotly Visualizer #1308

umesh-timalsina opened this issue Nov 11, 2019 · 3 comments
Assignees
Labels
enhancement job feedback Graphs, images, and other feedback generated during job execution.
Milestone

Comments

@umesh-timalsina
Copy link
Contributor

The current metamodel is updated to add sublot support as well and should be updated to incorporate image support as well (#1290). As it turns out plotly is easy to use visualizer with subplot and image trace support and would be preferred over d3 plotting library that we currently use. So, a separate visualizer should be developed for plotly with subplots and image support. deepforge-dev/deepforge-keras#138 could help immensely for JSON to GME conversion, a lot of it is done based on the meta-nodes to be created. Rather than having a standard conversion mechanism.

Possible limitations:
The current ImageTrace in plotly has primitive support for images and may not be used for advanced image visualization, but is actively under development. A way to convert(rgb) image into matrix in javascript might be necessary(Although this can be achieved in the custom_matplotlib_backend) we have. Either way, After #1290 is closed, this should be a priority to incorporate a new visualizer in the Execution visualization.

@umesh-timalsina umesh-timalsina added enhancement job feedback Graphs, images, and other feedback generated during job execution. labels Nov 11, 2019
@umesh-timalsina umesh-timalsina self-assigned this Nov 11, 2019
umesh-timalsina added a commit that referenced this issue Nov 11, 2019
@umesh-timalsina
Copy link
Contributor Author

This tree shows all the nodes being created. Now ExecutionIndexControl needs to be modified to send updated information to the plotly widget.

umesh-timalsina added a commit that referenced this issue Nov 11, 2019
@umesh-timalsina
Copy link
Contributor Author

umesh-timalsina commented Nov 11, 2019

image

74c6931 replaces the existing LineGraphWidget in ExecutionIndexControl with PlotlyGraphWidget. Now, there should be a standard way of doing three things:

  1. Creating Graph, Subgraph and Image nodes from backend_deepforge.py's json represenation of the plot (Add documentation for JSONImporter. Closes #137 deepforge-keras#138).
  2. Change Images in the metamodel to point to
  3. Also, incrementally support different plots in matplotlib in the visualizer.

But, the first priority should be changing the existing code to support subplots(should be st. forward) by changing ExecutionIndexControl PlotlyGraphWidget and PlotlyGraphPanel. This implementatioin should only resolve subplot issue (#1290) and should #1218 be addressed in a different PR?

@umesh-timalsina
Copy link
Contributor Author

umesh-timalsina commented Nov 11, 2019

This implementation should only resolve subplot issue (#1290) and should #1218 be addressed in a different PR?

The best way to do this is to have a separate feature branch for image support.

umesh-timalsina added a commit that referenced this issue Nov 20, 2019
umesh-timalsina added a commit that referenced this issue Nov 21, 2019
…data, Refactor GraphDescExtractor to send plotlyJSON Directly to widget (Widget Display problem for Failed execcutions)
umesh-timalsina added a commit that referenced this issue Nov 22, 2019
@brollb brollb added this to the 2.2.0 milestone Nov 26, 2019
@brollb brollb closed this as completed in e4585d5 Nov 27, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement job feedback Graphs, images, and other feedback generated during job execution.
Projects
None yet
Development

No branches or pull requests

2 participants