adding Luxembourg administrative sanctions

[bgmello]

Fixes opensanctions/crawler-planning#157

[bgmello]

There are some sanctions where the title doesn't mention the company, such as "Sanction administrative prononcée à l’encontre d’un réviseur d’entreprises agréé" (Administrative sanction pronounced against an approved company auditor). We are currently ignoring those.

There also appears to be a bug where the lookup doesn't recognize 2 titles:

• Sanction administrative prononcée à l’encontre de l’établissement de crédit BEMO EUROPE – Banque Privée S.A.
• Sanctions et mesures administratives prononcées à l’encontre de quatorze professionnels agréés ou enregistrés soumis à la surveillance de la Commission de Surveillance du Secteur Financier (la « CSSF ») en matière de lutte contre le blanchiment et le financement du terrorisme

Thus they are raising warnings even though they are already in the YAML

[jbothma]

Thanks! Some potential improvements

[jbothma]

instead of maintaining all items in the yaml, I think we can come up with an algorithm that's fairly safe against accidental nonsense names:

1. strip r"(Sanction|amende) administrative prononcée à l’encontre d[eu] (gestionnaire de fonds d’investissement)?" from the start
2. if the rest starts with upper case, that looks like a name, so use it
3. else if there's a match in lookups, use the values from that
4. else use the rest as a name and warn to check

[bgmello]

I added more rules to the REGEX and I was able to get more names, but there was still some that required lookups

[jbothma]

yeah totally fine - it's just good to automatically handle the common cases so we don't literally have to add each and every record.

I realise now, the instances where they don't mention a name, we'll have to manually pull the name from PDF. How about the following logic to do a second lookup:

if not res.names:
    pdf_res = lookups(detail_url)
    if pdf_res is None:
        #warn
    names = names_res.names
else:
    names = res.names
for name in names:
    ...

the fist lookup handles if we can't parse a name from the subtitle
the second lookup happens if the first lookup gave us an empty list, then we use the detail url to look up a name someone found in the PDF and added - for cases like these https://github.com/opensanctions/opensanctions/pull/849/files#diff-31dfffcd8054c60130382e3389c526fa63b3e350581059737aaf4f8ab5c12834R55

[jbothma]

I don't want to assume that we can give the name AIFM Law for Sanction administrative prononcée à l’encontre d’un gestionnaire de fonds d’investissement - it's totally conceivable that they'd use the same string again for a different sanction, right?

[bgmello]

I checked all links and could only find 3 with a company name in the PDF. From what I understood, AIFM Law is not a name of a company, it's the Alternative Investment Fund Managers Law.

[jbothma]

lol shows how well I read that. sorry. But this, looks good!

[jbothma]

Thanks! looking good! Just a couple of changes then we're there