Should we have a new validate() function?

[torbengb]

As originally suggested here perhaps main() should call a new validate() function where a variety of validation checks could live?

1. "Configuration file not found [...]"
2. "Specified directory not found, attempting to find Downloads folder."

[torbengb]

Note: Item 1 above already exists as closed ticket #20. If we decide to build that new function, the #20 code could be moved there.

[nocalla]

I don't think this is a good idea, on further consideration. It goes against the try-except flow we're using. The CLI output at the moment is superfluous, I agree, but I don't think this is the right approach as it will end up duplicating code. I think a better approach would be to tell the user what folder is being used once the file list is found. I'll play with this a bit and get back to you.

[nocalla]

Please review the code over in #58 and see what you think of the output.

[torbengb]

goes against the try-except flow we're using

You're right. I like the idea of being consistent. I just tested #58 as you asked, using an invalid path, and it acted odd to me: first it complained about the directory not found, and then it processed a file, then it complained again, and processed the other file. I understand, technically, that it just does a fallback to the default which is where 2 files happen to be, but that doesn't rhyme with the error messages then. Thoughts?

[nocalla]

It's 2 different sections that are being used. So it makes sense to try the path each time.

…

On 25 Oct 2017 21:48, "Torben Gundtofte-Bruun" ***@***.***> wrote: goes against the try-except flow we're using You're right. I like the idea of being consistent. I just tested #58 <#58> as you asked, using an invalid path, and it acted odd to me: first it complained about the directory not found, and then it processed a file, then it complained again, and processed the other file. I understand, technically, that it just does a fallback to the default which is where 2 files happen to be, but that doesn't rhyme with the error messages then. Thoughts? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APeoH6MwIXDvqh713Dov4Yk86IHntiDgks5sv56ugaJpZM4QCcm_> .

[torbengb]

Do you feel this looks like what you expected:

Your specified download directory was not found: /home/torben/Downloadsx
Parsing file: elbi_umsaetze_20171024191306.csv
Using format: AT Raiffeisen Bank
Writing file: fixed_elbi_umsaetze_20171024191306.csv
Done!
Your specified download directory was not found: /home/torben/Downloadsx
Parsing file: finstatus_20171024191324.csv
Using format: AT Raiffeisen VISA
Writing file: fixed_finstatus_20171024191324.csv
Done!
2 files processed.

[nocalla]

Yeah, that's what I would've expected. Maybe it would be better to specify the format first, but that makes sense to me.

[torbengb]

I'm sorry but that output doesn't make sense to me as a user...

• "First it says it can't find the download, and then it still processes something? How?" Maybe it should then also state what directory it defaults to. That would clear up some confusion.
• "It has already told me once that it couldn't find my directory; small wonder that it still can't find it half a second later, eh?" I understand that it happens because of where it is in the code, but it doesn't look too smart.

I hope I didn't offend, it isn't meant to!

[nocalla]

Ok, the first point makes sense. I think the second point would be cleared up by stating the format beforehand also. I think it's necessary to state it for each section in case of custom directories being set in the individual formats.

…

On 25 October 2017 at 22:36, Torben Gundtofte-Bruun < ***@***.***> wrote: I'm sorry but that output doesn't make sense to me as a user... - "First it says it can't find the download, and then it still processes something? How?" Maybe it should then also state what directory it defaults to. That would clear up some confusion. - "It has already told me once that it couldn't find my directory; small wonder that it still can't find it half a second later, eh?" I understand that it happens because of where it is in the code, but it doesn't look too smart. I hope I didn't offend, it isn't meant to! — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APeoH_6LCkvl0fiZur84GtZQ3fTlI4G0ks5sv6nwgaJpZM4QCcm_> .

[torbengb]

necessary to state it for each section in case of custom directories being set in the individual formats.

That's a valid point!

[nocalla]

Changed things around a bit. Example output below.

python bank2ynab.py
Format: AT Raiffeisen Bank
Can't find: c:\users\example-username\Downloads
Trying: D:\Users\nocalla\Downloads
Parsing file: elbi_umsaetze_ (2).csv
Writing file: fixed_elbi_umsaetze_ (2).csv
Removing file: elbi_umsaetze_ (2).csv
Done!
Format: IE Bank of Ireland
Can't find: c:\users\example-username\Downloads
Trying: D:\Users\nocalla\Downloads
Parsing file: test_BOI_TransactionExport.csv
Writing file: fixed_test_BOI_TransactionExport.csv
Removing file: test_BOI_TransactionExport.csv
Done!
2 files processed.

[nocalla]

This should be resolved appropriately in master now.