Célande Adrien authored Aug 14, 2018 and Romain Bignon committed Aug 18, 2018

When the user is already logged in and it tries to log anyway, the website response by a json page which says it.
But sometimes, this page may be catch by the Login2Page because of the URL.

So now, the page exist with a is_here to be sure.

Closes: 16413@sibi
Closes: 16408@sibi
Closes: 17916@sibi
Closes: 10663@sibi
Closes: 13746@sibi
Closes: 10622@sibi
Closes: 16474@sibi
Closes: 15877@sibi
Closes: 16473@sibi
Closes: 17884@sibi
Closes: 18531@sibi
Closes: 15851@sibi

Quentin Defenouillere authored Jul 03, 2018 and Romain Bignon committed Jul 29, 2018

Added a liquidity line for Compte Titres PEA and removed it from
Portefeuille Titres PEA.
Reversed code for bp and linebourse as required.
Removed balance update to fit website balances.

Closes: 5653@zendesk, 5765@zendesk, 5162@zendesk

Quentin Defenouillere authored Jul 02, 2018 and Romain Bignon committed Jul 29, 2018

The certificate chain is sometimes broken on the linebourse website but
not always ; it seems to depend on the customers connection. In order to
prevent bugs related to SSL errors, the VERIFY = False enables
linebourse navigation without issues.

Quentin Defenouillere authored Jul 02, 2018 and Romain Bignon committed Jul 29, 2018

Since the set_gocardless_transaction_details() method often causes a crash in
get_history() because of a wrong data form, and since we did not manage
to figure out why the params are incorrect yet, we decided to remove the
call to this method to prevent connection bugs.

Closes: 5879@zendesk

Romain Bignon authored Jul 01, 2018 and Romain Bignon committed Jul 29, 2018

This partially reverts commit c34fed0bc9e8b637575d8ba7f41cd7cacc010848.

Quentin Defenouillere authored Jun 28, 2018 and Romain Bignon committed Jul 29, 2018

The get_history method used a for loop a yielded transactions one by
one, which led to a crash if the gocardless information filling did not
work properly. I added a return in case we land on the ErrorPage.
This patch fills the transaction list while filling the gocardless information
for each gocardless transaction and eventually returns the transaction_list:
this way, the transaction list is always properly returned without crash due
to bugs in the set_gocardless_transaction_details() method.

Closes: 5879@zendesk

Quentin Defenouillere authored Jun 26, 2018 and Romain Bignon committed Jul 29, 2018

Some Gocardless transactions do not have any href, leading to crashes
during iter_transactions when the set_gocardless_transaction_details()
method is called. This patch first checks if there is a link before
calling the method, thus preventing crashes during iter_history.

Closes: 5887@zendesk, 5746@zendesk, 5744@zendesk

Baptiste Delpey authored Jun 07, 2018 and Romain Bignon committed Jun 09, 2018

page after getting back of transactiondetail seems to have a new url

Closes: 4173@sibi

Célande Adrien authored May 15, 2018 and Romain Bignon committed Jun 09, 2018

To get the references, one must open the detail page for one particular transaction.
To do so, the module get the parameters, load the page and return to the transaction page.
Knowing that to get the details, you must know and retrieve the transaction id which changes on each loading of the transaction page.

The references are stored in the raw field.

Closes: 5152@zendesk

Quentin Defenouillere authored May 15, 2018 and Romain Bignon committed Jun 09, 2018

Re-added the Investments class to lifebourse/pages.py
Modified banquepopulaire browser to scrap 'TIT' but also 'PRV' accounts
Modified banque postale browser to avoid scraping liquidities and avoid
redundancies.

The previous Banque Populaire code only scrapped accounts starting with
'TIT', this was modified since accounts starting with 'PRV' also appear
in Lifebourse with invests and liquidites.
An 'XX-liquidity' class was re-added to the pages files of Lifebourse,
since Banque Populaire customers wanted liquidities to appear. The
browser file from Banque Postale file was also modified to prevent
liquidities from appearing since it would be redundant with current
accounts.

Closes: 5162@zendesk

Baptiste Delpey authored Jan 03, 2018 and Romain Bignon committed Jan 14, 2018

We might land on a error page indicating that the browser is unavailable
when we can actually scrap things. This patch might seems ugly but I
don't think we should remove the raise on the on_load of the ErrorPage.

Etienne Lachere authored Nov 28, 2017 and Romain Bignon committed Dec 03, 2017

Fixed by checking the class of the page before calling "no_operations" method

Closes: 453874@redmine

[banquepopulaire] Changed condition of previous fix

Closes: 453874@redmine

[banquepopulaire] Added check of the page class before calling "no_operations" method

Closes: 453874@redmine

ntome authored Sep 27, 2017 and Romain Bignon committed Oct 08, 2017

Operating on life insurances from natixis sub-site did not work anymore.
What actually changed is unclear.
Parameters after "#" may not be fetched from URL though since requests
strips them from the request.

ntome authored Jul 25, 2017 and Romain Bignon committed Aug 30, 2017

The table @id changed

ntome authored Jul 25, 2017 and Romain Bignon committed Aug 30, 2017

Remove cssselect use. Fix bug in RetryLoginBrowser if a list is
returned. Avoid lxml getparent.

baptiste delpey authored Jul 24, 2017 and Romain Bignon committed Aug 30, 2017

prefrer this than browserunavailable since natixis can be broken a
pretty long time

ntome authored Jul 05, 2017 and Romain Bignon committed Jul 08, 2017

The weboob arg should now be passed as a kwarg, this avoids passing
multiple times the weboob as positional argument in case of multiple
AbstractBrowser.

Vincent Ardisson authored Jan 30, 2017 and Romain Bignon committed Feb 11, 2017

Support non-weboob iterators, and parameterize number of retries.

Vincent Ardisson authored Jan 20, 2017 and Romain Bignon committed Jan 28, 2017

Login pages raise an exception if not coming from do_login. The
exception is caught at browser-method level which retries the operation.