biolab / orange3-timeseries

🍊 :chart_with_upwards_trend: Orange add-on for analyzing, visualizing, manipulating, and forecasting time series data.
Other
62 stars 41 forks source link

TIMESERIES add-on v0.4.0 #212

Closed LizzyOberholzer2 closed 2 years ago

LizzyOberholzer2 commented 2 years ago

What's wrong?

ORANGE DATA MINING I installed the TIMESERIES add-on and it shows ticked, hence installed (restarted as well) but not showing up in my left pane.

How can we reproduce the problem?

I think there is a bug with v0.4.0, install it and let me know if you can access the Timeseries widgets

What's your environment?

thocevar commented 2 years ago

I've just successfully installed Orange 3.32 and the TimeSeries add-on on a clean Windows machine.

Could you try starting Orange from command line with C:\Users\<user>\AppData\Local\Programs\Orange\python.exe -m Orange.canvas -l4 (you can find the right installation folder from the target of your Orange shortcut) and report the output.

Avicebron1972 commented 2 years ago

I am using Orange3-3.32.0. in "Portable" version (unfortunately, the installation of the Miniconda version does not work on my PC). I also installed the TIMESERIES add-on v0.4.0. Everything works fine, except the two models ARIMA and VAR: when I insert them in the flowchart and try to open them I get these errors:

ARIMA

Traceback (most recent call last): File "C:\Users\info\Orange\lib\site-packages\orangecanvas\scheme\widgetmanager.py", line 236, in add_widget_for_node w = self.create_widget_for_node(node) File "C:\Users\info\Orange\lib\site-packages\orangewidget\workflow\widgetsscheme.py", line 300, in create_widget_for_node widget = self.create_widget_instance(node) File "C:\Users\info\Orange\lib\site-packages\orangewidget\workflow\widgetsscheme.py", line 413, in create_widget_instance widget.init() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets\owarimamodel.py", line 29, in init super().init() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets_owmodel.py", line 47, in init__ self.setup_layout() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets_owmodel.py", line 132, in setup_layout self.add_main_layout() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets\owarimamodel.py", line 43, in add_mainlayout gui.spin(None, self, 'p', 0, 100, **kwargs)) File "C:\Users\info\Orange\lib\site-packages\orangewidget\gui.py", line 783, in spin sbox = bi.control = b.control = \ AttributeError: 'NoneType' object has no attribute 'control'

VAR _------ Traceback (most recent call last): File "C:\Users\info\Orange\lib\site-packages\orangecanvas\scheme\widgetmanager.py", line 236, in add_widget_for_node w = self.create_widget_for_node(node) File "C:\Users\info\Orange\lib\site-packages\orangewidget\workflow\widgetsscheme.py", line 300, in create_widget_for_node widget = self.create_widget_instance(node) File "C:\Users\info\Orange\lib\site-packages\orangewidget\workflow\widgetsscheme.py", line 413, in create_widget_instance widget.init() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets_owmodel.py", line 47, in init__ self.setup_layout() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets_owmodel.py", line 133, in setup_layout self.add_bottom_buttons() File "C:\Users\info\Orange\lib\site-packages\orangecontrib\timeseries\widgets_owmodel.py", line 152, in add_bottom_buttons gui.spin(None, self, 'forecaststeps', 1, 100, File "C:\Users\info\Orange\lib\site-packages\orangewidget\gui.py", line 783, in spin sbox = bi.control = b.control = \ AttributeError: 'NoneType' object has no attribute 'control'

Can you help me, please? Thank you Regards

janezd commented 2 years ago

@Avicebron1972, the problem with ARIMA and VAR unrelated. We're aware of it and have fixed it; we thought we'd prepare a new release, but haven't gotten to it.

@LizzyOberholzer2 's problem is different; we'd need the log (see @thocevar comment) to understand what is going on.

Avicebron1972 commented 2 years ago

Thanks for the reply. One question: I would like to use Orange for my data science classes at the University of Venice (I think Orange is great for teaching), but I need to use ARIMA model for time series analysis. Are you planning to release a new version of ARIMA by the end of June or should I use another software? Thanks. Regards :)

ajdapretnar commented 2 years ago

Timeseries release 0.4.1 is now available on PyPi and should be available on conda within a few hours.

Avicebron1972 commented 2 years ago

Thanks! 👍