return + arguments is ok with Python 3.x but not Python 2.x

This module allows you to download the following PDF documents:

* Your employees monthly payslips
* Your monthly direct debits and social contributions details
* Your monthly registration records
* Your yearly tax certificates

This module allows you to download the following PDF documents:
* Your employees payslips
* Your monthly direct debits
* Your yearly tax certificates

When using `boobill documents` to list all documents, I had this error:
AttributeError: 'str' object has no attribute 'id'

This small fix gets rid of that.

The bank's website has added a new ".eir_showxs" class to display
elements only on mobile devices, only containing the transaction's date
(without the "Date de valeur" prefix). Remove it if we see it in the
list of text elements in the raw label.

It works well with the connection I checked.

We decided to keep only PER and handle different kinds with specific
attributes later on.

Summary transactions do not have any details.

The exception is temporary and retrying usually fix the error.

They change the json response, add a check to know whether
we are going to get an SCA or not.

Sometimes when going to the market space, the website returns an
error 404 or 403. This is usually fixed the next time.

The market_login.go() request is skipped if we already landed on the
market home page after the first location.

Added new PhenixBrowser to the module to handle the new site.
All connections with type 1 (cardholder) will be redirected towards the new website.

The link we retrieve from the HTML is really unstable and has a lot of
HTTP Error 500. The new link (used by the website for elements that do
not have `href` attributes) works fine.

The 'titulaire' browser missed a condition to catch wrongpass errors.
The 'entreprise' browser now retrieve the error message when raising a
wrongpass.

It seems that the previous code was filtering transactions because the website was returning all transaction
from all cards in one request, the website seems to returning only the good transactions per request,
so i've removed what it seems to be depreciated code.

In some case, the subscription details page, where we get the
information for the subscription lab is not available.
We get a 401 error when getting it.

In this case, we can't get the phone numbers associated to the
subscriptions.

We set then the subscription label to the subscription id without the
phone numbers.

Sometimes there is no phone numbers associated to a subscription.
In this case the label was empty like " - "

We now use the subscription id as label, and concat phone numbers list
to it if we can get them.

Currently, account labels have no informations on who the accounts
belongs to. For Organization connections with accounts belonging to
different organization this caused multiple accounts with the same label
but no way to identify which organization it belonged to.

The website seems to enjoy creating new error messages and concatenate
others.
I now try to use generic parts of messages.
This way, I try to handle as much different messages as possible.

- 'Dépassement du montant':
	- 'Dépassement du montant journalier autorisé'
	- 'Dépassement du montant autorisé par virement' (new)
	- 'Dépassement du montant mensuel autorisé' (new)
- 'Opération non conforme':
	- 'Opération non conforme, les virements entre comptes Epargne ne sont pas autorisés.'
	- 'Opération non conforme: entre titulaires différents, seul un
compte courant peut être débité'
	- 'Opération non conforme, entre titulaires différents, seul un compte courant peut être crédité.' (new)
- 'Virement interdit':
	- 'Virement interdit sur compte clos'
	- 'Virement interdit sur compte inexistant'
	- 'Virement interdit sur compte inexistant ou clos' (new)
- 'Montant maximum autorisé':
	- 'Montant maximum autorisé au débit pour ce compte'
	- 'Montant maximum autorisé au crédit pour ce compte'

Some transactions were typed as bank instead of card.
It is because the typing in the module is too restrictive.

"XX/XX/XXXX PAIEMENT CB YYYY PLACE SHOP CARTE NUMBER -ZZ,ZZ EUR"

research of document by using form doesn't work anymore (even on website)

for example, if you have a document at 28/09/19 and you set date filter between 01/07/19 and 01/11/19
website tells you there is no document at all.
but give documents from last 6 months on the main documents page

this commit let the old code for the day they will fix their website since it doesn't crash

When you try to get transactions from the 10000th you have to reset to 0 and change the last_date to continue