Vscode-jupyter: Jupyter used to show my plots in VS Code now it doesn't

[brando90]

Environment data

Version: 1.42.1
Commit: c47d83b293181d9be64f27ff093689e8e7aed054
Date: 2020-02-11T14:44:27.652Z
Electron: 6.1.6
Chrome: 76.0.3809.146
Node.js: 12.4.0
V8: 7.6.303.31-electron.0
OS: Darwin x64 19.3.0

Expected behaviour

Show my torchviz plots on VS code

Actual behaviour

Says:

<graphviz.dot.Digraph at 0x124421490>

instead of displaying the image.

Steps to reproduce:

self contained code to reproduce:

import torch

from torchviz import make_dot

x = torch.randn(1)

y = 3*x

make_dot(y)

instead of displaying image it prints:

<graphviz.dot.Digraph at 0x128cc40d0>

Logs

Output for Python in the Output panel (View→Output, change the drop-down the upper-right of the Output panel to Python)

EMPTY OUTPUT

Output from Console under the Developer Tools panel (toggle Developer Tools on under Help; turn on source maps to make any tracebacks be useful by running Enable source map support for extension debugging)

console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:01: Activating Environment to capture Environment variables, . /Users/brandomiranda/miniconda3/envs/automl-meta-learning/bin/activate && conda activate base && echo 'e8b39361-0157-4923-80e1-22d70d46dee6' && python /Users/brandomiranda/.vscode/extensions/ms-python.python-2020.2.63072/pythonFiles/printEnvVariables.py
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:01: parseEnvironmentOutput, Class name = S, completed in 0ms
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:01: getActivatedEnvironmentVariables, Class name = S, completed in 1579ms, Arg 1: undefined, Arg 2: {"architecture":3,"path":"/Users/brandomiranda/miniconda3/envs/automl-meta-learning/bin/python","version":{"options":{"loose":false,"includePrerelease":false},"loose":false,"raw":"3.7.6-final","major":3,"minor":7,"patch":6,"prerelease":["final"],"build":[],"version":"3.7.6-final"},"sysPrefix":"/Users/brandomiranda/miniconda3/envs/automl-meta-learning","fileHash":"d6fcec69eb45106b78c8b7e2bdc1a01e7ec3e18c1891b180796455cb288051a6c198219bb18a867ffb6d9a6488901a5ad21f316f65277c27de3c4ccef07d7d00","companyDisplayName":"Anaconda, Inc.","type":"Conda","envPath":"/Users/brandomiranda/miniconda3/envs/automl-meta-learning","envName":"base","displayName":"Python 3.7.6 64-bit"}, Arg 3: true
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Submitting code for 7627e4a1-eb0d-41d8-aab7-8f9e617fa3c3
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Waiting for jupyter server and web panel ...
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Cached data exists getEnvironmentVariables, <No Resource>
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Getting jupyter server options ...
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Connecting to jupyter server ...
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Connected to jupyter server.
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:11:29: Finished execution for 0651c08b-fab6-4f76-a9cf-aaae3959bedf
console.ts:137 [Extension Host] Info Python Extension: 2020-02-18 10:12:21: Cached data exists getEnvironmentVariables, tasks
log.ts:197   ERR No application in the Launch Services database matches the input criteria.: Error: No application in the Launch Services database matches the input criteria.

[brando90]

I'm unsure if I have the most recent version of everything cuz I am using this debugger: https://github.com/microsoft/ptvsd/issues/1706

although things were working perfectly well last night so I am unsure what happened overnight...

[brando90]

related bug, but VS code doesn't save my images either:

(automl-meta-learning) brandomiranda~/automl-meta-learning/prototyping_tests_playground/trainable_optimizers ❯ env PTVSD_LAUNCHER_PORT=55117 /Users/brandomiranda/miniconda3/envs/automl-meta-learning/bin/python /Users/brandomiranda/.vscode/extensions/ms-python.python-2020.2.63072/pythonFiles/lib/python/new_ptvsd/wheels/ptvsd/launcher /Users/brandomiranda/automl-meta-learning/prototyping_tests_playground/trainable_optimizers/torch_viz_graph.py 

with script:

import torch

from torchviz import make_dot

x = torch.randn(1)

y = 3*x

make_dot(y).render('image', format='png')

but running the file directly from the integrated terminal seems to work:

automl-meta-learning) brandomiranda~/automl-meta-learning/prototyping_tests_playground/trainable_optimizers ❯ python torch_viz_graph.py 
(automl-meta-learning) brandomiranda~/automl-meta-learning/prototyping_tests_playground/trainable_optimizers ❯ ls
__pycache__                             cifar10_tutorial.ipynb                  detached_deepcopy.ipynb                 differentiable_mdl.py                   image                                   sample_torchviz.ipynb                   torch_viz_same_graph.ipynb
backprop_through_optimizer.py           clone_playground.ipynb                  diff_over_backward.py                   extracting_gradients_as_data.ipynb      image.png                               scheduler.ipynb                         visualize_complicated_detach.ipynb
backward_twice.ipynb                    data                                    differentiable_mdl.ipynb                grad_test_sympy.py                      is_leaf.ipynb                           torch_viz_graph.py                      visualize_pytorch_graphs.ipynb

[IanMatthewHuff]

@brando90 thanks for the report on this. I see the issue. I think that we regressed something in our last release. I'll look at getting a fix out quick.

[rchiodo]

Validated