- Jan 08, 2020
-
-
-
-
-
when two (or more) credit cards are linked to a same parent account, as card numbers are close to each others it can mess'up the fetch of history providing the transactions from the other card. fixing in a row a side effect of the same issue preventing to revert CB payment record from parent account to card account to perform monthly reset. fixing also an issue in the history which was missing some transactions: if history is being called when no new transaction has been recorded since last credit card payement then history tab will have only last period and before last period shown. But currently hsbc module assume that first tab always contains "coming" entries (which are due in the future) so it is always skipped for fetching history. So past entries of the last period are not fetched from history and neither form coming as they are filtered out due to past date. Add new mapping for transaction types related to Global Transferts (with own foreign accounts).
-
-
-
Currently in python 2 if we try to print a BrowserQuestion that has a unicode message, it crash with the error UnicodeEncodeError or UnicodeDecodeError. This can be fixed by using the __unicode__ from StrConv.
-
Boursorama implemented a payment in 3 months called CLIC, the url was not handled, most of the page is similar to other loans but most of the Xpaths are not persistent. For example some Xpaths are there but empty while other lack of accents. So I added the URL, changed most of the Xpaths to handle all loans, and since it does not have a rate, I put a default value for this attribute.
-
The ListElement's fill_env added in 077105e8e2f89049d301cde804062360af3c3ceb was not implemented properly and crashes the module. Moreover, it is useless because it is only important that the DateGuesser object be the same between the Transactions.
-
-
-
-
We set the default value of selenium logger to ERROR to avoid spamming logs with useless information. Also, the data we send to the browser using selenium (with send_keys) can be displayed clearly in the log, if the log level is set to DEBUG.
-
Get rid of NoCopyLinearDateGuesser because it was only created to change the behaviour of LinearDateGuesser's __deepcopy__ to avoid the copy of the object when copying *Elements' env. Now that you can override this behaviour with fill_env this modification is no longer needed. See : e1027be50a506baa6fe525a61896e3cfb9c25e6c
-
Added fill_env method to AbstractElement so that it can be overriden in pages methods. It is useful to chose how Env behaves
-
Some children of creditmutuel do not use the same xpath where we search for the deferred date.
-
The xpath to get the loan's name changed on the website.
-
Waiting for an account to develop it, and because a consent is cross browser, we raise an ActionNeeded to make the user performing the SCA in its side.
-
Irriga, Generali & Delfea were not typed correctly. After typing, the iter_investments did not work because Euro funds have no ISIN code and the codeIsin key is absent from the JSON.
-
The CloudCard uses biometric recognition to authenticate the user. So it is not Supported. Otherwise, if an exception is not raised, it generates a 400 error with the code 102 and the message 'Unauthorized to execute fallback'
-
The is_here on ProAccountsList caused the errors to not be catched. New ActionNeeded error catched in ProAccountsList. New BrowserUnavailable error catched in the RedirectPage.
-
Somme patterns for matching rdates from transactions labels weren't strict enough, for exemple : ^(?P<category>PRELEVEMENT) (?P<text>.*) (?P<dd>\d{2})\s(?P<mm>\d{2})\s(?P<yy>\d{4}) .* matched PRELEVEMENT TEXT TEXT (YYYY MM DD YYYY MM DD) ^(?P<category>PRELEVEMENT) (?P<text>.*) (?P<dd>\d{2})-(?P<mm>\d{2})$ matched PRELEVEMENT TEXT TEXT MM-YY Both were made stricter as to avoid these unwanted matches that would crash the date parser.
-
Otherwise iter_history crashes for accounts that have no pid.
-
-
The investment details page contains the performance history of the investment. Handling these pages correctly also avoids logouts and BrowserHTTPSDowngrade errors.
-
-
-
-
The history_history parsing was completely broken. Now, all transaction have an 'id' key = 0 so we cannot retrieve proper transaction IDs anymore. The code to get the amount was wrong since there was a (self) missing for Field('raw'), we were evulating a Regexp() on the Field class itself. We do not scrape transaction.amount if it is in a different currency than the account; in these cases we only scrape the original_amoutn and the original_currency.
-
Field('code') was never evaluated, my bad!
-
The module didn't respect the data model for transations concerning orgininal_currency and original_amount. The acounts are in Euros but the website doesn't provide conversion so foreign amount and currencies had to be handled.
-
A few code tweaks too
-
-
-
The account balance is not available in the JSON, so we need to sum the investment valuations to get the correct value. However, it seems that investments in JPY are multiplied by 100, so we must divide this value accordingly otherwise the invest AND the account abalnce will be wrong.
-
Transactions date are displayed in the JSON in yyyy-mm-dd format but where retrieved using day_first=True. Transactions are currently lacking an id. When they get one, we can merge this MR and the transactions in the backend will be corrected.
-
The JSON after login slightly changed, now we need to use the path 'data/intAccount' instead of 'intAccount'.
-
Transactions were lacking an id. Patch related to !2547 in order to correct wrong dates in ther backend. We first add an id to map the backend, and then when this patch is live we patch with the !2547. Needs backend!1286 first.
-
-
Balance : The account balance and the total amount of the investments in that same account was not the same because of the wrong dictionnary key. original_valuation : added parse function to handle correclty investments in foreign currency liquidity : replaced the investments that are cash in the iter_investments list by liquidities.
-