Ignore initial unprotected uppercase in translated titles

[ryan-carpenter]

Some databases store translated titles in the title field, surrounded by square brackets like this.

title = {[Entry with a translated title]}

Because the title phrase begins at the second position of the title string JabRef warns that the initial uppercase letter needs to be protected. To avoid matching this pattern, the regex could include something like [A-Z](?<!^\[.) or (?<^!\[)[A-Z].

[ryan-carpenter]

jabref/src/main/java/org/jabref/logic/integrity/TitleChecker.java

Line 14 in ca00d64

private static final Pattern INSIDE_CURLY_BRAKETS = Pattern.compile("\\{[^}\\{]*\\}");

jabref/src/main/java/org/jabref/logic/integrity/TitleChecker.java

Line 15 in ca00d64

private static final Pattern DELIMITERS = Pattern.compile("\\.|\\!|\\?|\\;|\\:");

jabref/src/main/java/org/jabref/logic/integrity/TitleChecker.java

Line 16 in ca00d64

private static final Predicate<String> HAS_CAPITAL_LETTERS = Pattern.compile("[\\p{Lu}\\p{Lt}]").asPredicate();

Possibly one of these changes would work in line 15, but I am not sure if the patterns only match if there is white space following the delimiters.
Pattern.compile("\\.|\\!|\\?|\\;|\\:"|\\A\\[);
Pattern.compile("\\.|\\!|\\?|\\;|\\:"|^\\[);

[Siedlerchr]

I added this to the project board, I think it's a valid issue.
Additional requirement: Write/Enhance Unit test

[michalfarago]

Hello, I'd like to inquire about this issue.

If I've understood it correctly, you are talking about the following code:

jabref/src/main/java/org/jabref/logic/integrity/TitleChecker.java

Lines 54 to 63 in ca00d64

	for (String subTitle : splitTitle) {
	subTitle = subTitle.trim();
	if (!subTitle.isEmpty()) {
	subTitle = subTitle.substring(1);
	if (HAS_CAPITAL_LETTERS.test(subTitle)) {
	return Optional.of(Localization.lang("capital letters are not masked using curly brackets {}"));
	}
	}
	}

You'd like to handle this scenario by modifying the pattern and create unit tests. Correct?

[Siedlerchr]

FIrst create a test case for the behavior.
Then adjust the pattern/code etc to take this into account
The test could also be converted to a parameterized test

[ManaMatrix]

Hey team I'm new to open-source contributing, I'd love to work on this issue if it is still free to take. Thank you

[michalfarago]

Hello, I've already started working on this. Could you please assign this to me instead?

[ThiloteE]

@ManaMatrix please check out our other good first issues.

[michalfarago]

Hello, I'd like to double-check the following:

Titles like: title = {Folding kinetics of a polymer [corrigendum]}, are supported and don't raise a warning.

From the issue description, it's clear that the following should also not raise a warning: title = {[Translated title]},

My question is whether the following should raise a warning: title = {[Translated title] Original title},

I've checked the resource folder, but didn't find any entry of that format, so I suspect that it should raise a warning. Let me know if I'm mistaken.

Thank you.

[Siedlerchr]

@michalfarago Good question, I think it should still raise a warning. It's at least unusual do have both in the same entry
Maybe @ryan-carpenter has a preference here as well?

[Siedlerchr]

@ryan-carpenter The issue is now fixed in the latest development version