Update and fix examples

[PMeira]

They don't run as-is and there are some other problems reported in other issues.

We should also remove the PS Dataport or reduce its use in the examples due to issues reported in https://github.com/nilmtk/nilmtk/issues/873

[Rithwikksvr]

@PMeira @levaphenyl I can help with writing the notebooks. Just, one quick question. Currently, we have nilmtk-contrib which is under-development and nilmtk-contrib which is hosted on anaconda cloud.

For which of these nilmtk-contrib versions should I start writing the tutorial notebooks? Also, the same issue exists with NILMTK. Post, completion I can send a PR.

Also, I think we can use REDD dataset and DRED dataset. We can also link the notebooks for making REDD.h5 and DRED.h5.

Hopefully, that will make it easier to use the toolkit

[PMeira]

The development version would be better. If that works for you without major issues, we can release another version (maybe after #37?), with a matching NILMTK.

the same issue exists with NILMTK. Post, completion I can send a PR.

Yep, I believe I just copied the notebook at the time.

Also, I think we can use REDD dataset and DRED dataset. We can also link the notebooks for making REDD.h5 and DRED.h5.

OK. From https://github.com/nilmtk/nilmtk/pull/850, AMPds2 includes a HDF5. If that works, it could also be an alternative.

Some of the notebooks from the Buildsys 2019 paper don't require Dataport, right? We could probably link to those more explicitly too.

[levaphenyl]

I agree the example should work with the master branch.

Advantage of REDD is the stability: the dataset authors are not breaking the converter.