(QIF) mapping ubs-ch-fr: better normalization and other fixes

[sphakka]

• qif: Correct handling of date format.
• csvtrim: much improved with CL options, detection of EOL and handling of partially broken CSV exports.

[sphakka]

As for the failing test "3.13, macos-latest", I guess that's due to some missing locales. Needed: fr_CH. Any advice?

[jaraco]

Overall, this looks fine to me.

I'm not very familiar with csvtrim, and I sort-of feel like it should be moved to another project (or replaced with something like csvkit or miller). csvtrim appears to be bash-specific and thus minimally portable. That said, I can see that you've contributed to this utility before, and it only applies to this one mapping, so getting this code uncommented and fixed is useful. Let's proceed.

[jaraco]

I fixed the ruff formatting error, but there are still failures in the tests for the mapping. Please fix.

[sphakka]

I fixed the ruff formatting error, but there are still failures in the tests for the mapping. Please fix.

Hi, thanks for all this. Actually, the failing test is on mapping 'n26', not my 'ubs-ch-fr'. I see it also on my local clone:

=================================================================================== FAILURES ===================================================================================
____________________________________________________________________ test_sample[opts30-n26-fr.csv-n26.ofx] ____________________________________________________________________

opts = ['-m', 'n26'], in_filename = 'n26-fr.csv', out_filename = 'n26.ofx', capsys = <_pytest.capture.CaptureFixture object at 0x7fcaa4db11d0>
monkeypatch = <_pytest.monkeypatch.MonkeyPatch object at 0x7fcaa50030e0>

    @pytest.mark.parametrize(['opts', 'in_filename', 'out_filename'], samples)
    @freezegun.freeze_time("2016-10-31 11:29:08")
    def test_sample(opts, in_filename, out_filename, capsys, monkeypatch):
        monkeypatch.setattr(csv2ofx.main, '_time_from_file', lambda path: time.time())
        arguments = [str(data / 'test' / in_filename)]
        command = list(itertools.chain(['csv2ofx'], flatten_opts(opts), arguments))
        with pytest.raises(SystemExit) as exc:
            csv2ofx.main.run(command[1:])
        # Success - exit code 0
        assert exc.value.code == 0
    
        expected = data.joinpath("converted", out_filename).read_text(encoding='utf-8')
>       assert capsys.readouterr().out == expected, (
            f"Unexpected output from {subprocess.list2cmdline(command)}"
        )
E       AssertionError: Unexpected output from csv2ofx -m n26 data/test/n26-fr.csv
E       assert 'DATA:OFXSGML...V1>\n</OFX>\n' == 'DATA:OFXSGML...V1>\n</OFX>\n'
E         
E         Skipping 638 identical leading characters in diff, use -v to show
E         Skipping 514 identical trailing characters in diff, use -v to show
E         - TEND>20250406</DTEND>
E         ?           ^^^
E         + TEND>20250729</DTEND>
E         ?           ^^^

tests/test_cli.py:148: AssertionError
=========================================================================== short test summary info ============================================================================
FAILED tests/test_cli.py::test_sample[opts30-n26-fr.csv-n26.ofx] - AssertionError: Unexpected output from csv2ofx -m n26 data/test/n26-fr.csv

I haven't touched m26 code and, sorry, have no idea about how to fix it -- my branch is ahead of upstream...

As for 'csvtrim' (I'm the author), @reubano though of integrating it (#144) through 'meza' which is fine for me. Otherwise, it may be moved out of the code base.