GitLab

Avec python3, l'ordre d'itération sur les dictionnaires est aléatoire.
Lorsqu'un règlement concerne plusieurs factures, celles-ci sont donc
listées dans un ordre aléatoire, ce qui génère une écriture différente à
chaque lancement de himport.

Pour résoudre ce problème, on liste les factures dans l'ordre alphabétique
(ce qui correspond normalement à un ordre chronologique).

This caused Python3 (3.7.2 to be precise) to fail to print to stdout:

    Traceback (most recent call last):
      File "/home/zorun/.local/bin/himport", line 11, in <module>
        load_entry_point('himport', 'console_scripts', 'himport')()
      File "/home/zorun/tmp/himport/himport/cli.py", line 99, in main
        options = process_args()
      File "/home/zorun/tmp/himport/himport/cli.py", line 44, in process_args
        print(usage)
      File "/usr/lib/python3.7/codecs.py", line 378, in write
        self.stream.write(data)
    TypeError: write() argument must be str, not bytes

After removing the sys.stdout redefinition, python3 is now happy, and
python2 seems to still be able to print to stdout even with unicode
characters.

Dolibarr allows to group invoice lines under a title (actually, there can
be up to 9 levels of title/subtitle/etc).

Unfortunately, the title of each group appears in the database as a
legitimate invoice line, so it gets imported by himport even though it
does not make any sense.

This patch ignores these "fake" invoice lines by filtering on the
'special_code' column.

Fix #1

Read the README.md to use himport

table llx_accountingaccount -> llx_accounting_account

The multicurrency is not supported

Need to define the sqlalchemy according to the dolibarr version.
DolibarrSQLAlchemy embeds now the classes.

Releasing himport 0.1