Port all existing (python, libshogun) examples to meta examples

[karlnapf]

#3000 is the full thing, example and cookbook page

This task is simple copy paste: Just take any of the existing python examples examples/uncodcumented/python_modular and create a meta example that does the same thing.

Remove the old example afterwards

See https://github.com/shogun-toolbox/docs/blob/master/EXAMPLES.md and https://github.com/shogun-toolbox/docs/blob/master/DEVELOPING.md for instructions.

This is a great task to get started with Shogun's examples, very easy.

There are a few hundred examples, so lots to do :)

[souravsingh]

@karlnapf I would like to take a stab at the issue.

[karlnapf]

Yeah sure, as said, there is hundreds

[karlnapf]

This is still really important

[spothound]

I have write a post in my blog commenting things I've seen while creating the example I've just submitted. If someone trying to add samples can find it useful I would be delighted :)

[karlnapf]

Great! :)

Keep in mind also that make meta_examples parses and translates all meta examples found by cmake (if you added a new one you have to rerun cmake), and that build_cpp_meta_examples builds the cpp versions of them, and finally, ctest -R generated_cpp-binary-linear_support_vector_machine runs your example. If you want more details on how it is executed, append a -V

[karlnapf]

There are still more examples to port as of early 2019: see examples/undocumented/python

[taichi-yagami]

@karlnapf I wish to work on making meta examples for remaining py files mentioned in examples/uncodcumented/python directory, I wanted to know that whether there are any documentation / tutorial/ blogs on how to convert those py file into and meta example, and how it works.

[karlnapf]

There is a readme

[taichi-yagami]

@karlnapf Ok I will take this task of building meta example for remaining python files.

[karlnapf]

We will need to port all examples in examples/undocumented/python/*.py to the new meta example system, using factories and base types rather than specialised types (i.e. using Machine m = machine("LibSVM") rather than LibSVM m = LibSVM(), and using m.apply().get("labels") rather than m.apply_binary().get_labels()

[Hephaestus12]

There is a readme

Where is this readme available?
Also since we are no longer writing cookbook pages (.rst) files do we still need to comment sections on the meta example (.sg) files?

[vigsterkr]

@Hephaestus12 i'm not so sure where did you read about not writing cookbook pages.... yes you need to comment the sections.
please check out our documentation and you'll find answers to most of your questions; if not please create an issue (https://github.com/shogun-toolbox/shogun/blob/develop/doc/readme/EXAMPLES.md#automatically-generated-examples-)

[Hephaestus12]

@vigsterkr Sorry, I was referring to the distance examples, it was mentioned here: #3000 (comment)
Anyway, thanks for the help :)

[vigsterkr]

@Hephaestus12 that's a specific case...

[festusojo123]

Wanted to upload .sg and .rst files for the Linear Time MMD example but couldn't make a new branch to start the patch. Also, is trying to make files for the exampes on the cookbook site still a good initial patch, or are there any others that you'd recommend that are new from the 2017 "getting started" page? Thanks1

[vigsterkr]

@festusojo123 plz read about how to submit your first PR on github (search for manual on the internet)

[Khalifa1997]

@karlnapf There are many cases where labels would be generated using np.random how do you suggest going about those cases

[karlnapf]

Load a dataset instead. We have many, check the other examples