Complete example with Firebase cache

[gcattan]

Issue #84
This PR completed the example with MOABB and ERP classification.
It demonstrates how we can use Firebase admin to synchronize MOABB results.
A simple use case is to split the number of subjects in a dataset between two computers.
Then the two MOABB evaluations are run in parallel and the results are synchronized on the Firebase server.
It is also useful when we want to store and retrieve the results of past computation on a same computer.

@toncho11 FYI

[toncho11]

Does Firebase require any login credentials?
Can it work locally?
Does the connector require Internet and Google access?

[gcattan]

1. Yes, it requires credentials. The wiki page here https://github.com/pyRiemann/pyRiemann-qiskit/wiki/Use-a-Firebase-cache explains how to get credentials from Firebase.
2. It can work locally. The generate_caches function accepts a dictionary object that, if provded, is used instead of the remote database. E.g.:

fake = {
   "Brain Invaders 2012":{
      "RG+QuantumSVM":{
         "Brain Invaders 2012":{
            "1":{
               "RG+QuantumSVM":{
                  "true_labels":0.3433986306190491,
                  "predicted_labels":0.5
               }
            },
         }
      },
      "RG+LDA":{
         "Brain Invaders 2012":{
            "1":{
               "RG+LDA":{
                  "true_labels":0.13163451850414276,
                  "predicted_labels":0.6777283549308777
               }
            },
            "2":{
               "RG+LDA":{
                  "true_labels":0.13357791304588318,
                  "predicted_labels":0.8143579959869385
               }
            }
         }
      }
   }
}

3. Yes, if you want to connect to Firebase, you need internet and google credential. There is no replication of the database on local storage.

[toncho11]

I think there should be a separate example. We should not modify classify_P300_bi.py.

[toncho11]

What is the dictionary object accepted by generate_caches? An empty one that is going to be filled by generate_caches?

[gcattan]

What is the dictionary object accepted by generate_caches? An empty one that is going to be filled by generate_caches?

It accepts a list of datasets, a list of pipelines and optionally a set of mock data.

[gcattan]

Thanks for the review :)