Dash: [BUG]ModuleNotFoundError: No module named '_plotly_future_'
Describe your context
Please provide us your environment so we can easily reproduce the issue.
I am using Python 3.6.6 from the Anaconda rpi repository. I am using berryconda to create a conda virtual environment.
dash 0.43.0
dash-core-components 0.48.0
dash-daq 0.1.5
dash-html-components 0.16.0
dash-renderer 0.24.0
dash-table 3.7.0
if frontend related, tell us your Browser, Version and OS
- OS: [Raspbian Stretch]
- Browser [Web ]
- Version [ 3.8.2]
Describe the bug
Whenever I run a python script which imports the Dash module, a ModuleNotFoundExceptionError is thrown:
Exception has occurred: ModuleNotFoundError
No module named '_plotly_future_'
File "/home/pi/tdunit/app.py", line 3, in <module>
import dash
File "/home/pi/berryconda3/envs/tdunitenv/lib/python3.6/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/pi/berryconda3/envs/tdunitenv/lib/python3.6/runpy.py", line 96, in _run_module_code
mod_name, mod_spec, pkg_name, script_name)
File "/home/pi/berryconda3/envs/tdunitenv/lib/python3.6/runpy.py", line 263, in run_path
pkg_name=pkg_name, script_name=fname)
Expected behavior
I expect the program to execute normally. I have even tried the getting started example for Dash and the same error occurs.
Screenshots
tsokadev
All 11 comments
weird - _plotly_future_ should have come in with plotly. What version of the plotly package do you have?
@jonmmease any idea what's going on?
alexcjohnson
on 25 Jun 2019
This kind of thing is usually caused by there being multiple copies of plotly.py installed at the same time. Can you check the output of
pip list
and
conda list
to see what versions of the plotly package are installed?
jonmmease
on 25 Jun 2019
@alexjohnson, @jonmmease it appears that I have only 1 version of plotly:
(tdenv) pi@tdunit:~ $ conda list
# packages in environment at /home/pi/berryconda3/envs/tdenv:
#
# Name Version Build Channel
attrs 19.1.0 <pip>
ca-certificates 2018.8.24 0
certifi 2018.8.24 py36_1
chardet 3.0.4 <pip>
Click 7.0 <pip>
dash 0.36.0 <pip>
dash-core-components 0.42.1 <pip>
dash-daq 0.1.4 <pip>
dash-html-components 0.13.4 <pip>
dash-renderer 1.0.0 <pip>
DateTime 4.2 <pip>
decorator 4.4.0 <pip>
Flask 1.0.3 <pip>
Flask-Compress 1.4.0 <pip>
gunicorn 19.8.1 <pip>
idna 2.8 <pip>
ipython-genutils 0.2.0 <pip>
itsdangerous 1.1.0 <pip>
Jinja2 2.10.1 <pip>
jsonschema 3.0.1 <pip>
jupyter-core 4.5.0 <pip>
libgfortran 3.0.0 0
MarkupSafe 1.1.1 <pip>
nbformat 4.4.0 <pip>
ncurses 6.1 h4f752ac_1
numpy 1.14.0 py36h73aad69_0
numpy 1.16.1 <pip>
openblas 0.2.19 0
openssl 1.0.2r hdff2a78_0
pandas 0.23.4 <pip>
pip 18.0 py36_1
plotly 3.1.0 <pip>
pyrsistent 0.15.2 <pip>
python 3.6.6 hd0568c0_1
python-dateutil 2.8.0 <pip>
pytz 2019.1 <pip>
readline 7.0 hcb560eb_1
requests 2.22.0 <pip>
retrying 1.3.3 <pip>
scipy 1.0.0 py36h741f5fb_0
setuptools 40.2.0 py36_0
six 1.12.0 <pip>
sqlite 3.24.0 hfcb1bcf_1
tk 8.6.8 h849d6a0_0
traitlets 4.3.2 <pip>
urllib3 1.25.3 <pip>
Werkzeug 0.15.4 <pip>
wheel 0.31.1 py36_1
xz 5.2.4 hdff2a78_1
zlib 1.2.11 hdff2a78_1003
zope.interface 4.6.0 <pip>
I get a segmentation fault if I do pip list inside the virtual environment.
(tdenv) pi@tdunit:~ $ pip list
Segmentation fault
Outside the virtual environment:
pi@tdunit:~ $ pip list
DEPRECATION: The default format will switch to columns in the future. You can use --format=(legacy|columns) (or define a format=(legacy|columns) in your pip.conf under the [list] section) to disable this warning.
asn1crypto (0.22.0)
astroid (2.0.2)
certifi (2018.8.24)
cffi (1.10.0)
chardet (3.0.4)
conda (4.5.11)
cryptography (1.9)
idna (2.5)
isort (4.3.4)
lazy-object-proxy (1.3.1)
mccabe (0.6.1)
packaging (16.8)
pip (9.0.1)
pycosat (0.6.3)
pycparser (2.17)
pylint (2.1.1)
pyOpenSSL (16.2.0)
pyparsing (2.2.0)
requests (2.18.1)
ruamel.yaml (0.11.7)
setuptools (36.0.1)
six (1.10.0)
typed-ast (1.1.0)
urllib3 (1.21.1)
wheel (0.29.0)
wrapt (1.10.11)
Conda list outside virtual environment:
pi@tdunit:~ $ conda list
# packages in environment at /home/pi/berryconda3:
#
# Name Version Build Channel
asn1crypto 0.22.0 py36_0
astroid 2.0.2 py36_0
ca-certificates 2018.8.24 0
certifi 2018.8.24 py36_1
cffi 1.10.0 py36_0
chardet 3.0.4 py36_0
conda 4.5.11 py36_0
conda-env 2.6.0 1
cryptography 1.9 py36_0
idna 2.5 py36_0
isort 4.3.4 py36_0
lazy-object-proxy 1.3.1 py36_0
libffi 3.2.1 0
mccabe 0.6.1 py_1
ncurses 5.9 0
openssl 1.0.2r hdff2a78_0
packaging 16.8 py36_0
pip 9.0.1 py36_0
pycosat 0.6.3 py36hdff2a78_1
pycparser 2.17 py36_0
pylint 2.1.1 py36_0
pyopenssl 16.2.0 py36_0
pyparsing 2.2.0 py36_0
python 3.6.1 0
readline 6.3 0
requests 2.18.1 py36_0
ruamel_yaml 0.11.14 py36_0
setuptools 36.0.1 py36_0
six 1.10.0 py36_0
sqlite 3.15.0 0
tk 8.6.6 0
typed-ast 1.1.0 py36hdff2a78_0
urllib3 1.21.1 py36_0
wheel 0.29.0 py36_0
wrapt 1.10.11 py36hf843537_0
xz 5.2.2 0
yaml 0.1.6 0
zlib 1.2.11 0
Hi @tsokadev ,
Something odd is going on with the plotly vesions. The virtual environment reports plotly version 3.1.0, but this version predates the _plotly_future_ module. So the plotly.py file at site-packages/plotly/__init__.py somehow came from a different plotly installation. You could try uninstall plotly using pip and then reinstalling version 3.10.0.
Also, getting a segmentation fault when calling pip list from your virtual environment is a bit concerning. Perhaps there is some inconsistency in the environment that is causing both problems.
jonmmease
on 26 Jun 2019
Hi, I've got the same problem as @tsokadev and when checking pip list and conda list, I have Plotly 3.10.0 on both.
I've uninstalled and reinstalled Plotly but the error still persists:
(base) Florence:dash florence.townend$ python app.py
Traceback (most recent call last):
File "app.py", line 2, in <module>
import dash
File "/Users/florence.townend/anaconda3/lib/python3.7/site-packages/dash/__init__.py", line 1, in <module>
from .dash import Dash, no_update # noqa: F401
File "/Users/florence.townend/anaconda3/lib/python3.7/site-packages/dash/dash.py", line 22, in <module>
import plotly
File "/Users/florence.townend/anaconda3/lib/python3.7/site-packages/plotly/__init__.py", line 30, in <module>
from _plotly_future_ import _future_flags
ModuleNotFoundError: No module named '_plotly_future_'
Could there be another reason it's not working?
florencejt
on 26 Jun 2019
Hi @florencejt, how did you install plotly in the first place? Here's the uninstall/install sequence I mean.
$ conda uninstall plotly
$ pip uninstall plotly
$ pip install plotly
Then if things don't work after running that sequence, add the output of conda list and pip list here. Also, check the contents of your /Users/florence.townend/anaconda3/lib/python3.7/site-packages/ directory to see if there is a _plotly_future_ directory in there.
jonmmease
on 26 Jun 2019
When I tired coda uninstall Plotly I got this:
(base) Florence:~ florence.townend$ conda uninstall plotly
Collecting package metadata: done
Solving environment: failed
PackagesNotFoundError: The following packages are missing from the target environment:
- plotly
But when I did the pip uninstall and install again, for some reason it worked this time even though I think I did the same thing as before?
Well, I'm not getting an error anymore and the demo worked so thank you.
florencejt
on 26 Jun 2019
@tsokadev there's also something weird with the dash versions you've got. These ones:
dash 0.36.0 <pip>
dash-core-components 0.42.1 <pip>
dash-daq 0.1.4 <pip>
dash-html-components 0.13.4 <pip>
dash-renderer 1.0.0 <pip>
[email protected] (the latest version, out last week) is only expected to work with [email protected], but you have much older dash, dash-core-components, and dash-html-components
alexcjohnson
on 26 Jun 2019
The Plotly future flags only showed up in v3.8 or so and I see a v3.1 up there in the list. Best to upgrade to plotly v3.10 no?
nicolaskruchten
on 26 Jun 2019
I reinstalled everything in a new environment and the problem has gone away. @alexcjohnson the reason for the weird dash versions is that I installed them from a requirements.txt file which had fixed version numbers.
Thanks for all the help.
tsokadev
on 26 Jun 2019
I'm running into the same issue. Can I ask for step by step on a fresh install?
Assume I'm not using conda.
@tsokadev would you mind posting a pip list?
isConic
on 2 Jul 2019
Related issues
Mondrik
路
3Comments
lochbrunner
路
3Comments
Chris8080
路
3Comments
T4rk1n
路
3Comments
Ricardo-C-Oliveira
路
4Comments