moparisthebest/filebot
1
0
Fork 0
mirror of https://github.com/mitb-archive/filebot synced 2024-08-13 17:03:45 -04:00
Code Issues Releases Wiki Activity

* reduce index size by reducing entries with redundant substring sequences

Browse Source
This commit is contained in:
Reinhard Pointner 2014-09-15 19:04:25 +00:00
parent 1ce479e858
commit 56253dfb7a
1 changed files with 10 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
BuildData.groovy
View File

@ -64,20 +64,22 @@ def isValidMovieName(s) {
}
def getNamePermutations(names) {
def fn1 = { s -> s.replaceAll(/^(?i)(The|A)\s/, '') }
def normalize = { s -> s.toLowerCase().normalizePunctuation() }.memoize()
def fn1 = { s -> s.replaceAll(/(?i)(^(The|A)\s)|([,]\s(The|A)$)/, '') }
def fn2 = { s -> s.replaceAll(/\s&\s/, ' and ') }
def fn3 = { s -> s.replaceAll(/\([^\)]*\)$/, '') }
def out = names*.trim().unique().collectMany{ original ->
def simplified = original
[fn1, fn2, fn3].each{ fn -> simplified = fn(simplified).trim() }
return [original, simplified]
}.unique().toList()
def simplified = original
[fn1, fn2, fn3].each{ fn -> simplified = fn(simplified).trim() }
return [original, simplified]
}.unique{ normalize(it) }.findAll{ it.length() > 0 }
out = out.findAll{ it.length() >= 2 && !(it ==~ /[1][0-9][1-9]/) && !(it =~ /^[a-z]/) && it =~ /^[@.\p{L}\p{Digit}]/ } // MUST START WITH UNICODE LETTER
out = out.findAll{ !MediaDetection.releaseInfo.structureRootPattern.matcher(it).matches() } // IGNORE NAMES THAT OVERLAP WITH MEDIA FOLDER NAMES
return out.unique{ it.toLowerCase().normalizePunctuation() }.findAll{ it.length() > 0 }
out = out.findAll{ a -> names.take(1).contains(a) || out.findAll{ b -> normalize(a).startsWith(normalize(b) + ' ') }.size() == 0 } // TRY TO EXCLUDE REDUNDANT SUBSTRING DUPLICATES
return out
}
def treeSort(list, keyFunction) {