Closed wehuang16 closed 1 year ago
Thanks for reporting this issue. It looks like we're not clearing the shapes graph when a library is over written.
If you don't actually want to overwrite the library you can pass overwrite=False
when calling the load function. This will return the existing library in the database if one exists.
Thank you for your information, Tobias! My small suggestion is that we include the "overwrite=False" to BuildingMOTIF's tutorial page (https://nrel.github.io/BuildingMOTIF/tutorials/model_validation.html), so that others who use BuildingMOTIF will not run into the same error and wonder what to do.
I believe this is fixed in https://github.com/NREL/BuildingMOTIF/pull/168
Body: This issue may be related to previous issues #168 and #109 . Basically what I did was, I copied the “model validation” tutorial (https://nrel.github.io/BuildingMOTIF/tutorials/model_validation.html) to a Jupyter notebook. The Jupyter notebook is attached here and is named “model_validation_tutorial.ipynb”. I stored this Juypyter notebook in the BuildingMOTIF/docs/tutorials path. I ran the Jupyter notebook without problems for the first time. However, when I ran it the second time, it returned a Shape Load error shown at the below screenshot. I did some preliminary troubleshooting and found that when I ran the “libarary-load-directory” coding line “g36 = Library.load(directory="../../libraries/ashrae/guideline36")” twice, the shape load error was returned. However, when I ran any other “library-load-ontology_graph” coding line twice, such as the line “brick = Library.load(ontology_graph="../../libraries/brick/Brick-subset.ttl")”, it did NOT have the shape load error. Thus, I would like to bring this issue up, and seek solutions to make BuildingMOTIF easier to use in Jupyter notebook.