search

biolab / orange3

🍊 :bar_chart: :bulb: Orange: Interactive data analysis
https://orangedatamining.com
Other
4.78k stars 996 forks source link

Color Widget Borks My dataset #5485

Open bluewin4 opened 3 years ago

bluewin4 commented 3 years ago

If I remove the color widget I can replot things just fine with either the old or new data. But then I don't have the plot colours I wanted. image

bluewin4 commented 3 years ago

Update: I closed orange and opened it again and the color widget no longer borks my data but the colours I selected all got reset to generic in all of my color widgets. Of note, when I closed Orange, it generated an error, I had to rejigger which data sets were included to get the error again though

Exception: AttributeError: Can't pickle local object 'bind_variable.<locals>.cast'
Module: orangewidget.settings:774
Widget Name: Color
Widget Module: Orange.widgets.data.owcolor:1149
Version: 3.30.0.dev
Environment: Python 3.8.8 on Windows 10 10.0.18362 AMD64
Installed Packages: AnyQt==0.0.13, Bottleneck==1.3.2, CacheControl==0.12.6, Genesis-PyAPI==1.2.1, Orange-Spectroscopy==0.5.8, Orange3-Associate==1.1.8, Orange3-Bioinformatics==4.3.1, Orange3-Educational==0.4.0, Orange3-Explain==0.3.0, Orange3-Network==1.5.1, Orange3-Survival-Analysis==0.2.0, Orange3-Timeseries==0.3.10, Orange3==3.30.0.dev0+f7fd0a7, Pillow==8.1.2, PyQt5-sip==4.19.18, PyQt5==5.12.3, PyQtChart==5.12, PyQtWebEngine==5.12.1, PySocks==1.7.1, PyYAML==5.4.1, XlsxWriter==1.3.7, asteval==0.9.23, astor==0.8.1, autograd-gamma==0.5.0, autograd==1.3, baycomp==1.0.2, beautifulsoup4==4.9.3, brotlipy==0.7.0, catboost==0.24.4, certifi==2020.12.5, cffi==1.14.5, chardet==4.0.0, colorcet==2.0.6, commonmark==0.9.1, cryptography==3.4.6, cycler==0.10.0, decorator==4.4.2, dictdiffer==0.8.0, docutils==0.16, et-xmlfile==1.0.1, extranormal3==0.0.3, formulaic==0.2.3, future==0.18.2, h11==0.12.0, h2==4.0.0, h5py==3.2.1, hpack==4.0.0, httpcore==0.12.3, httpx==0.16.1, hyperframe==6.0.0, idna==2.10, importlib-metadata==4.0.1, interface-meta==1.2.3, itsdangerous==1.1.0, jdcal==1.4.1, joblib==1.0.1, keyring==22.3.0, keyrings.alt==4.0.2, kiwisolver==1.3.1, lifelines==0.25.11, llvmlite==0.36.0, lmfit==1.0.2, lockfile==0.12.2, lxml==4.6.3, matplotlib==3.3.4, more-itertools==8.7.0, msgpack==1.0.2, networkx==2.5, numba==0.53.1, numpy==1.20.2, olefile==0.46, openTSNE==0.6.0, openpyxl==3.0.7, orange-canvas-core==0.1.21, orange-widget-base==4.13.1, pandas-datareader==0.9.0, pandas==1.2.3, param==1.10.1, patsy==0.5.1, pip==21.0.1, point-annotator==2.0.0, pyOpenSSL==20.0.1, pyclipper==1.2.1, pycparser==2.20, pyct==0.4.8, pyparsing==2.4.7, pyqtgraph==0.12.1, python-dateutil==2.8.1, python-louvain==0.15, pytz==2021.1, pywin32-ctypes==0.2.0, qasync==0.13.0, renishawWiRE==0.1.14, requests-cache==0.6.3, requests==2.25.1, resdk==13.1.0, rfc3986==1.4.0, scikit-learn==0.23.2, scipy==1.5.3, serverfiles==0.3.0, setuptools==49.6.0.post20210108, shap==0.37.0, six==1.15.0, slicer==0.0.3, slumber==0.7.1, sniffio==1.2.0, soupsieve==2.2.1, spectral==0.22.2, statsmodels==0.12.2, threadpoolctl==2.1.0, tornado==6.1, tqdm==4.60.0, typing-extensions==3.10.0.0, tzlocal==2.1, uncertainties==3.1.5, url-normalize==1.4.3, urllib3==1.26.3, wheel==0.36.2, win-inet-pton==1.1.0, wincertstore==0.2, wrapt==1.12.1, xlrd==2.0.1, zipp==3.4.1
Machine ID: 28a35e85-dec8-4cd4-b940-6519b257496e
Stack Trace: Traceback (most recent call last):  File "C:\Program Files\Orange\lib\site-packages\orangewidget\workflow\widgetsscheme.py", line 512, in eventFilter    widget.saveSettings()  File "C:\Program Files\Orange\lib\site-packages\orangewidget\widget.py", line 1149, in saveSettings    self.settingsHandler.update_defaults(self)  File "C:\Program Files\Orange\lib\site-packages\orangewidget\settings.py", line 813, in update_defaults    self.write_defaults()  File "C:\Program Files\Orange\lib\site-packages\orangewidget\settings.py", line 530, in write_defaults    self.write_defaults_file(settings_file)  File "C:\Program Files\Orange\lib\site-packages\orangewidget\settings.py", line 774, in write_defaults_file    pickle.dump([add_version(context) for context in self.global_contexts],AttributeError: Can't pickle local object 'bind_variable.<locals>.cast'
Local Variables: OrderedDict([('class', <class 'orangewidget.settings.ContextHandler'>),             ('add_version',              <function ContextHandler.write_defaults_file.<locals>.add_version at 0x00000281487ABCA0>),             ('self',              <Orange.widgets.settings.PerfectDomainContextHandler object at 0x000002813980D490>),             ('settings_file',              <_io.BufferedWriter name='C:/Users/artim/AppData/Local\Orange\3.30.0.dev\widgets\Orange.widgets.data.owcolor.OWColor.pickle'>)])
bluewin4 commented 3 years ago

Update: I figured out a new way to break it, when changing the names of different variables in the color widget interface just keep pressing tab too many times and it will throw an error and then prevents the variables with color selection and name addition from being used in a plot

Note the problem with the appearing on plot only seems to occur if the names are all changed via the color widget

Traceback (most recent call last): File "c:\users\artim\onedrive\documents\orange\orange3\Orange\widgets\data\owcolor.py", line 330, in setData desc.set_value(col - 1, value) File "c:\users\artim\onedrive\documents\orange\orange3\Orange\widgets\data\owcolor.py", line 118, in set_value self.new_values[i] = value IndexError: list assignment index out of range

markotoplak commented 3 years ago

Thanks for finding new ways to break Orange. :) Your first problem seems especially interesting. I suspect something is not OK with how settings get handled.

We really appreciate that you attach workflows and data so that we can reproduce the problem quickly. Next time, if you have time, you could perhaps try to make the smallest workflow/dataset combination that can still reproduce the issue. Also, if you are openning data with Multifile, you can also save (Save Data) as a single file and then reproduce the problem with that file.