Wrong capitalized letter in Portuguese names

[kelvins]

First of all, congrats for the great project.

I have found a small issue related to Portuguese names. By running the following code:

from nameparser import HumanName
name = HumanName('joao da silva do amaral de souza')
name.capitalize()
str(name)

I get the following result:

'Joao da Silva Do Amaral de Souza'

when it should be:

'Joao da Silva do Amaral de Souza'

The d from do should be lowercase.

[derek73]

Thanks for the ticket. I can add "do" to the prefixes which will fix the capitalization. What should the correct parsing of the name be? This is what it does currently. Is that correct?

$ python tests.py "Joao da Silva do Amaral de Souza"
<HumanName : [
	title: '' 
	first: 'Joao' 
	middle: 'da Silva do Amaral de' 
	last: 'Souza' 
	suffix: ''
	nickname: ''
]>

[derek73]

If that's not the correct way to parse Portuguese names, just open another ticket and I'll fix it.

[kelvins]

Hello @derek73, sorry for the delay. Yes, it is correct. Thank you.

[derek73]

hey @kelvins, I'm working on fixing a different bug and my fix results in a new parsing of your example name where everything ends up in the last name:

$ python tests.py "Joao da Silva do Amaral de Souza"
<HumanName : [
	title: '' 
	first: 'Joao' 
	middle: '' 
	last: 'da Silva do Amaral de Souza' 
	suffix: ''
	nickname: ''
]>

My understanding of Portuguese last names is limited, but it seems like it could be more technically correct to include them all as last names since they appear to be joined by conjunctions. How do you feel about that? If it's obviously not correct to you, I'd like to know the reasons why it's not correct so I can figure out how to change my fix for #60.

[kelvins]

Hello @derek73, to be honest, in Brazilian Portuguese we do not use (only) the last name too much. We usually use the entire surname, for example:

First name: João
Middle name: da Silva do Amaral
Last name: de Souza
"Entire Surname": da Silva do Amaral de Souza

Personally, I don't think it is correct to say da Silva do Amaral de Souza is the last name, but, as I have said, we do not often use only the last name, so I don't think it is a big problem.

If you're interested, here are some other examples of Brazilian names:

First Name	Middle Name	Last Name
Kelvin	Salton	do Prado
João	da Silva	Siqueira
Pedro	Albuquerque	da Rosa
Joana	Silveira	Pinto
Arlindo	Luz	da Conceição
Jonas	Andriolli	da Fonseca
Maria	Pereira	de Queiroz
Tereza	Jesus	de Oliveira
Ana	Gonçalves	dos Santos
Luiza	Lima	Ferreira
Carla	Rodrigues	Almeida
José	do Nascimento	Carvalho
Taís	Araújo	Ribeiro
Daiane		dos Santos
Flávio		Prado

I hope it helps.

[derek73]

Thanks, that's what I needed. The table is super helpful. Here's my interpretation, let me know if I got anything wrong.

do/da/de/dos/das/des are always connected to the piece that follows it, so we can always treat those as one name piece. Then, the attribute bucket that they belong to (first, middle, etc) is determined by the position of the pair (e.g. da + following_piece) in the name, just like if the pair was a single name piece. As long as those articles are connected to the thing that follows them, we should be able to parse out the position just like any other name if we treat them like one piece.

I think I can do that. I think it just means that middle names can have prefixes too and right now I only allow last names.

Also, I noticed I only have the plural of one of those articles right now. I don't have "des" and "das" as prefixes. Are those possible? I only see a "dos" in your example names.

[kelvins]

Yes, that's exactly it. You can treat the pair (do/da/de/dos + following_piece) as a single name piece.

In Portuguese, I have never seen "des" or "das" as prefixes. If it actually exists, it should be very rare.

Here is a List of most common surnames in South America.

Note that are variations in surnames, for example:

João de Oliveira - João Oliveira
Flávio do Prado - Flávio Prado
Pedro da Silva - Pedro Silva

[derek73]

I just released v1.0 which I think handles prefixes as I described. This is the output now for your test name:

$ python tests.py "Joao da Silva do Amaral de Souza"
<HumanName : [
	title: '' 
	first: 'Joao' 
	middle: 'da Silva do Amaral' 
	last: 'de Souza' 
	suffix: ''
	nickname: ''
]>

Probably best then to leave "des" and "das" out of the constants if they are rare, it will more likely cause errors than be the desired output.

Thanks again for your help on this. I like it when it works right. :) Let me know if I misinterpreted anything.

[kelvins]

Perfect, thank you so much @derek73.

[fabiuz]

In portugues, 'do' is a contraction of preposition 'de' plus the article 'o'.
'Do' before of masculine substantive.
'Da' before of feminine substantive.
'Dos', is plural of 'do'.
'Das' is plural of 'da'.
'De' is preposition.
'Des' not exists into portugueses.

'do' is a contraction of preposition 'de' plus the article 'o', but, never in portuguese you see:
[Incorret] O github de o Fábio.
But yes:
[Corret] O github do Fábio.

In portugueses, there not are this tradition of 'Title', 'first', 'middleName', 'lastname'.

I speak portuguese of Brazil.