Repository: commons-text Updated Branches: refs/heads/master 6fd10f89a -> 7de03831b
TEXT-20: Update salutations regex as per comments in pull request Project: http://git-wip-us.apache.org/repos/asf/commons-text/repo Commit: http://git-wip-us.apache.org/repos/asf/commons-text/commit/7de03831 Tree: http://git-wip-us.apache.org/repos/asf/commons-text/tree/7de03831 Diff: http://git-wip-us.apache.org/repos/asf/commons-text/diff/7de03831 Branch: refs/heads/master Commit: 7de03831b805f795041ce4f13d53a6330cd64904 Parents: 6fd10f8 Author: Bruno P. Kinoshita <brunodepau...@yahoo.com.br> Authored: Sun Nov 20 20:43:25 2016 +1300 Committer: Bruno P. Kinoshita <brunodepau...@yahoo.com.br> Committed: Sun Nov 20 20:43:25 2016 +1300 ---------------------------------------------------------------------- src/main/java/org/apache/commons/text/names/HumanNameParser.java | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/commons-text/blob/7de03831/src/main/java/org/apache/commons/text/names/HumanNameParser.java ---------------------------------------------------------------------- diff --git a/src/main/java/org/apache/commons/text/names/HumanNameParser.java b/src/main/java/org/apache/commons/text/names/HumanNameParser.java index 1348313..5f4fa20 100644 --- a/src/main/java/org/apache/commons/text/names/HumanNameParser.java +++ b/src/main/java/org/apache/commons/text/names/HumanNameParser.java @@ -146,7 +146,7 @@ public final class HumanNameParser { // TODO compile regexes only once when the parser is created final String suffixes = StringUtils.join(this.suffixes, "\\.*|") + "\\.*"; final String prefixes = StringUtils.join(this.prefixes, " |") + " "; - final String salutations = StringUtils.join(this.salutations, " |") + " "; + final String salutations = StringUtils.join(this.salutations, " |"); // The regex use is a bit tricky. *Everything* matched by the regex will be replaced, // but you can select a particular parenthesized submatch to be returned. @@ -155,7 +155,7 @@ public final class HumanNameParser { final String nicknamesRegex = "(?i) ('|\\\"|\\(\\\"*'*)(.+?)('|\\\"|\\\"*'*\\)) "; final String suffixRegex = "(?i),* *((" + suffixes + ")$)"; final String lastRegex = "(?i)(?!^)\\b([^ ]+ y |" + prefixes + ")*[^ ]+$"; - final String salutationRegex = "^(?i)(("+salutations+")\\.)"; + final String salutationRegex = "^(?i)(("+salutations+")(\\.|\\s))"; // note the lookahead, which isn't returned or replaced final String leadingInitRegex = "(?i)(^(.\\.*)(?= \\p{L}{2}))"; final String firstRegex = "(?i)^([^ ]+)";
