Continuous importation of bank statement

Before writing a proposal for a new feature, I would like to have feedback from others tryton users, to see if my problematic is shared and how it is managed somewhere else.

I would like to have a way to add periodically new bank statements in Tryton. For some bank accounts it would be daily based whereas some others could be weekly or monthly. Import via files (like OFX) is fine for low level of iteration, but as soon several banks are involved, or for a daily based import, it become annoying.

Does any user has such use of account.statement via automatic importing ?

Currently, I could see several ways to answer my need:

  • As french user, I could use web-oob project which provide python script/modules to retreive, among other things, statements from several french banks. So a tryton module for importing Statement thought web-oob is a possibility. But I am unsure about the reusability.

  • Alternatively, having a more generic module to import generated content from external script (instead of from file like currently) is a possibility. It would be having a predefined sets of scripts, and let the user runs them (or run via cron task): “retreive Bank A” (and behind run the configured script, captures the output, save it as file, and import the file). But does anybody else has scripts for generating CSV or OFX from online bank ? (again for french bank, web-oob could do it)

  • just use a completely external solution like a proteus based script, and run it via system cron.

Could you describe what is annoying for your side?

We import our statements (manually) on a non regular basis (normally every week, but sometimes it’s some more days) and we’ve found that it will be good to show the last operation date imported for a journal. WE normally want to import the data from the last imported day to yesterday. I imagine this will be also useful for automatic importing.

On a project I work on, I have more than twelve banks accounts to manage. It means a lot of files to download from the bank websites and to import.

I am unsure it will be easy to manage so many files.

yes, I agree it could help. It is a good information to have.

On the first though, this solution looks very nice but I have some reticence in integrating a scrapping library in Tryton code. There could be some security and stability concern.

I do not see really the point to have this machinery in Tryton if it is to run external scripts. As they would require to be declared in some configuration file, only a sysadmin could do that. So the UI user could not really configure it by himself.

Indeed I think it is the best solution but instead of using a proteus script, I think we should provide a User Application API which will run the wizard in background.
In order to be user-friendly, I think this API should work with idempotent key (like the Stripe API) which will prevent to post twice the same file (aka same idempotent key).
Such API could be used by external program like weboob.org to provide import to Tryton feature or by service providers like plaid.com or teller.io.

yes the external script would have to be installed by the local administrator. but after, it could be run by a user (and the user would not known script details neither the credentials used for example).

but it wouldn’t apply for automatic importing.