Fix issue with default renderer when ipython is installed

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).

[emilykl]

Closes#5257

Reverts change to plotly/io/_renderers.py added in #5096 to keep the original logic flow of the if statement.

Adds a test to ensure that the default renderer is "browser" while running the tests; since the test_optional workflow installs jupyter which installs ipython, this should prevent a regression.

Steps to test

1. Verify incorrect behavior on main:
   • pip install jupyter
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()". This SHOULD show a figure in the browser, but instead it prints the HTML/JS to the terminal output.
2. Verify correct behavior on this branch:
   • Run python -c "import plotly.express as px; fig = px.scatter(x=[1], y=[1]); fig.show()" and make sure the figure is shown in the browser
3. Make sure figure still renders correctly in Jupyter notebook:
   • This is a little tricky because Jupyter rendering seems to not play nicely with the local install (at least on my machine) and I'm not sure why yet. But it's not a problem introduced by this PR; the same is true on main.
   • For that reason I recommend downloading the build artifacts from the CI and installing the built wheel.
   • Run jupyter notebook to launch the notebook editor, and create a new notebook
   • Make sure the following code correctly renders a figure in the notebook:

importplotly.expressaspxfig=px.scatter(x=[1], y=[1])
fig.show()

[marthacryan]

LGTM!

[camdecoster]

I followed the test steps above and everything worked correctly (including the Jupyter notebook with a local package).