pref_csv_separator is ignored with save statement for csv files

[Yapper689]

Describe the bug
List is not saved correctly in the CSV file, with each element written to a separate cell instead of all elements in one cell.

To Reproduce
GAMA model

model bug

global {
	
	list<int> closed_roads <- [1,2,3,4,5];
	int max_aqi <- 1;
	int mean_aqi <- 1;
	
	reflex save_results when: ( time > 0 ) {
		save [closed_roads, max_aqi, mean_aqi] to: "bug.csv" format: "csv" rewrite: false header: true  ;
	}
	 
}
	
experiment 'exp' {}

[lesquoyb]

To be clearer here is an example of the files produced by this model:

closed_roads,max_aqi,mean_aqi
[1,2,3,4,5],1,1
[1,2,3,4,5],1,1

The problem is that the delimiter for the array and the columns are the same ( , ), and to my knowledge there's no way of setting the delimiter when writing csv.
The file is still understandable and could be parsed by an external script, but it can't be loaded correctly by gama operators, for example this:

file f <- csv_file("bug.csv", true);
		loop el over: f {
            write el;
        }

Raises an outofboundsexception.
And not even the csv viewer of gama can read it correctly:

[AlexisDrogoul]

I have to say that there is no such thing as "saving a list inside a list" in the CSV format, at least no shared protocol and format to do so (contrary to other formats like JSON, which are based on nested data structures). This means (1) what you obtain is not a bug, just the normal result of saving a list inside a list; (2) you have to come up with your own way of encoding/decoding this list so that you can read it again. If the list is of fixed length, it is super easy, otherwise it might be a bit more complex (but not more complicated than what you would have to do in other languages).

[lesquoyb]

Yes the resulting file is not bugged, it's just very inconvenient and there's no way to produce a different csv with the current save statement.
Can't we add a delimiter facet so we can open the csv file directly in gaml using the csv_file operator instead of having to write a complete parser ?

[lesquoyb]

I tried a workaround suggested by @AlexisDrogoul and found out that the csv_parameter in gama preferences is ignored:

model bug

global {
	
	list<int> closed_roads <- [1,2,3,4,5];
	int max_aqi <- 1;
	int mean_aqi <- 1;
	
	 
}
	
experiment 'exp' {
	
	init {
		gama.pref_csv_separator <- "*";
	}
	
	reflex save_results  {
		save [closed_roads, max_aqi, mean_aqi] to: "bug.csv" format: "csv" rewrite: false header: true  ;
		file f <- csv_file("bug.csv","*" ,true);
		loop el over: f {
            write el;
        }
	}
	
}

The generated file still uses , as a separator. I tried different separator values without any difference. I also try to change it from the preference window with the same results.

[lesquoyb]

Fixed with commit bfa98eb0a0a10c599d76ed8724f9e99cf6dda7b4