- Jan 11, 2024
-
-
'/ws_q44/api/v1/infoclient/donnesClient' becomes '/ws_q44/api/v1/infoclient/donneesClient'
-
- Jun 21, 2023
-
-
Banque Postale has changed its virtual keyboard, it is no longer an image but HTML elements filled with random numbers.
-
- Apr 11, 2023
-
-
- ensure flake8 compatibility - sort imports - drop test file - drop Py2 support
-
-
- Apr 08, 2023
-
-
- Mar 28, 2023
-
-
- Mar 05, 2023
-
-
Romain Bignon authored
-
- Feb 28, 2023
-
-
Romain Bignon authored
-
- Jan 26, 2023
-
-
Existing code base was messy, obsolete & unfunctional
-
Since a few days many different requests are timing out. Hard to know if increasing to 20 will be enough as the error appears ramdomly, may need to increase further.
-
- Dec 02, 2022
-
-
Both in DA and fallback it is possible to encounter a generic error during the login process. To handle this, we add a new page to retrieve this error and raise a BrowserUnavailable if necessary.
-
"COMPTE ATTENTE" are checking accounts. When fetching their history, they must be handled differently than other checking accounts. This distinction is made by looking at their label that was "COMPTE ATTENTE" until now. "Compte Attente" can now be found but for the moment, it is impossible to know if the full uppercase label can still be found so the code handles both cases.
-
-
Renamed "types" dictionary to "ACCOUNTS_TYPES" and put it outside the class to be able to use it for loans typing too. Added mortgage type.
-
Life insurance history has recently been modified on the website. We have to get some params on different pages before we can finally request the life insurance history URL. One of these parameters is an "identifiantContrat", which was seen, until now, formed of only one digit. For some connections that have a massive amount of accounts, this number can go above 9. The life insurance history URL had a regex allowing only one digit so these URL with an "identifiantContrat" bigger than 9 were not matched.
-
- Oct 07, 2022
-
-
We assumed life insurances could only have one product code. It appears that there are at least 4 different types of life insurances, each with its own product code. Plus, if there is more than one life insurance of the same type for one account, the life insurances coming after the first one have their product code incremented by one each time there is one more life insurance. We now fetch the "codeProduit" and "identifiantContrat" dynamically. Wrong URL we used caused errors 500.
-
modification of xpath and url in the iter_accounts methods and modification of the parsing of LifeInsuranceHistory following an evolution of la banque postale website
-
bp returns us a 500 status code on documents page if we request for documents dating of 2013. So we simply skip it during the for loop that iter documents.
-
Cachemire account investments were partially parsed among other types of accounts. Now, cachemire accounts have their own page with a specific JSON to parse.
-
We had an unavailable page on the transfer confirm page on a given transfer. Since both transfer confirm and transfer summary pages share some URL patterns, they both define an is_here, leading to the unavailable page being matched in neither, causing the on_load raising the BrowserUnavailable not to be called. This commit adds a separate unavailable page which is loaded for every URL, and for which the presence is checked with only an is_here checking for the unavailable message.
-
- Sep 05, 2022
-
-
fill_account is based on profile to fill account ownership. but the profile is not always available and this causes bugs.
-
- Aug 12, 2022
-
-
Used to retrieve account.ownership. This prevents whole iter_accounts to break when bug is situated on the get_profile method. Tested on accounts with/out ownership, Par and Pro. This is a follow up to 3a1a9cb059becfc5affdb9e7ae3d41d47ff47fa7 but make it more resilient.
-
- Aug 08, 2022
-
-
- Aug 05, 2022
-
-
Profile page is now a JSON page. Its URL and way of accessing it has changed. Some headers must be present for that specific request or we'll get a ClientError 400 response.
-
- Jun 27, 2022
-
-
ntome authored
-
- Jun 13, 2022
-
-
- May 05, 2022
-
-
SCA is mandatory on La Banque Postale website since 25/04/2022. They removed an URL that was used to skip the SCA, causing the module to crash since the XPath does not exist anymore.
-
- Apr 28, 2022
-
-
- Feb 28, 2022
-
-
ntome authored
basestring, range, long, url*, parse*, unicode, html_unescape also, remove some obsolete sys.version checks
-
- Feb 18, 2022
-
- Jan 26, 2022
-
- Dec 18, 2021
-
-
- Nov 24, 2021
-
-
- Oct 13, 2021
-
-
- Sep 13, 2021
-
-
- AccountList define a get_error method, but we are calling self.page.has_error() in the browser - RevolvingAttributesPage define an has_error method, but we are calling self.page.get_error() in the browser Fix the name definitions and set them to get_error for consistency
-
-
- Aug 27, 2021
-
-