219 lines
6.8 KiB
Groovy
219 lines
6.8 KiB
Groovy
import org.tukaani.xz.*
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
def sortRegexList(path) {
|
||
def set = new TreeSet(String.CASE_INSENSITIVE_ORDER)
|
||
new File(path).eachLine('UTF-8'){
|
||
// check if regex compiles
|
||
set += java.util.regex.Pattern.compile(it.trim()).pattern()
|
||
}
|
||
def out = set.join('\n').saveAs(path)
|
||
println "$out\n$out.text\n"
|
||
}
|
||
|
||
|
||
// sort and check shared regex collections
|
||
sortRegexList("website/data/release-groups.txt")
|
||
sortRegexList("website/data/query-blacklist.txt")
|
||
sortRegexList("website/data/exclude-blacklist.txt")
|
||
sortRegexList("website/data/series-mappings.txt")
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
def reviews = []
|
||
new File('reviews.csv').eachLine('UTF-8'){
|
||
def s = it.split(';', 3)
|
||
reviews << [user: s[0], date: s[1], text: s[2].replaceAll(/^\"|\"$/, '').replaceAll(/["]{2}/, '"') ]
|
||
}
|
||
reviews = reviews.sort{ it.date }
|
||
|
||
def json = new groovy.json.JsonBuilder()
|
||
json.call(reviews as List)
|
||
json.toPrettyString().saveAs('website/reviews.json')
|
||
println "Reviews: " + reviews.size()
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
def moviedb_out = new File("website/data/moviedb.txt")
|
||
def thetvdb_out = new File("website/data/thetvdb.txt")
|
||
def anidb_out = new File("website/data/anidb.txt")
|
||
|
||
def pack(file, lines) {
|
||
new File(file.parentFile, file.name + '.xz').withOutputStream{ out ->
|
||
new XZOutputStream(out, new LZMA2Options(LZMA2Options.PRESET_DEFAULT)).withWriter('UTF-8'){ writer ->
|
||
lines.each{ writer.append(it).append('\n') }
|
||
}
|
||
}
|
||
}
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
// BUILD moviedb index
|
||
def isValidMovieName(s) {
|
||
return s=~ /^[A-Z0-9]/ && s =~ /[\p{Alpha}]{3}/
|
||
}
|
||
|
||
def treeSort(list, keyFunction) {
|
||
def sorter = new TreeMap(String.CASE_INSENSITIVE_ORDER)
|
||
list.each{
|
||
sorter.put(keyFunction(it), it)
|
||
}
|
||
return sorter.values()
|
||
}
|
||
|
||
|
||
def omdb = []
|
||
new File('omdb.txt').eachLine('Windows-1252'){
|
||
def line = it.split(/\t/)
|
||
if (line.length > 11 && line[0] ==~ /\d+/) {
|
||
def imdbid = line[1].substring(2).toInteger()
|
||
def name = line[2].replaceAll(/\s+/, ' ').trim()
|
||
def year = line[3].toInteger()
|
||
def runtime = line[5]
|
||
def rating = tryQuietly{ line[11].toFloat() } ?: 0
|
||
def votes = tryQuietly{ line[12].replaceAll(/\D/, '').toInteger() } ?: 0
|
||
|
||
if ((year >= 1970 && (runtime =~ /h/ || votes >= 200) && rating >= 1 && votes >= 50) || (year >= 1950 && votes >= 5000)) {
|
||
omdb << [imdbid.pad(7), name, year]
|
||
}
|
||
}
|
||
}
|
||
omdb = omdb.findAll{ (it[0] as int) <= 9999999 && isValidMovieName(it[1]) }
|
||
|
||
|
||
def tmdb_txt = new File('tmdb.txt')
|
||
def tmdb_index = csv(tmdb_txt, '\t', 1, [0..-1])
|
||
|
||
def tmdb = omdb.findResults{ m ->
|
||
if (tmdb_index.containsKey(m[0])) {
|
||
return tmdb_index[m[0]]
|
||
}
|
||
|
||
def sync = System.currentTimeMillis()
|
||
def row = [sync, m[0].pad(7), 0, m[2], m[1]]
|
||
try {
|
||
def info = net.sourceforge.filebot.WebServices.TMDb.getMovieInfo("tt${m[0]}", Locale.ENGLISH, false)
|
||
def names = [info.name, info.originalName, m[1]]
|
||
row = [sync, m[0].pad(7), info.id.pad(7), info.released?.year ?: m[2]] + names.findResults{ it ?: '' }
|
||
} catch(FileNotFoundException e) {
|
||
}
|
||
|
||
println row
|
||
tmdb_txt << row.join('\t') << '\n'
|
||
return row
|
||
}
|
||
tmdb*.join('\t').join('\n').saveAs(tmdb_txt)
|
||
|
||
movies = tmdb.findResults{
|
||
def ity = it[1..3] // imdb id, tmdb id, year
|
||
def names = it[4..-1].findAll{ isValidMovieName(it) }.unique{ it.toLowerCase().normalizePunctuation() }
|
||
if (ity[1].toInteger() > 0 && names.size() > 0)
|
||
return ity + names
|
||
else
|
||
return null
|
||
}
|
||
movies = treeSort(movies, { it[3, 2].join(' ') })
|
||
|
||
pack(moviedb_out, movies.findResults{ it.join('\t') })
|
||
println "Movie Count: " + movies.size()
|
||
|
||
// sanity check
|
||
if (movies.size() < 50000) { throw new Exception('Movie index sanity failed') }
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
// BUILD tvdb index
|
||
def tvdb = new HashMap()
|
||
def tvdb_txt = new File('tvdb.txt')
|
||
new File('tvdb.txt').eachLine{
|
||
def line = it.split('\t', 5).toList()
|
||
tvdb.put(line[0] as Integer, [line[0] as Integer, line[1], line[2], line[3], line[4] as Integer])
|
||
}
|
||
|
||
def tvdb_updates = new File('updates_all.xml').text.xml.'**'.Series.findResults{ s -> tryQuietly{ [id:s.id.text() as Integer, time:s.time.text() as Integer] } }
|
||
tvdb_updates.each{ update ->
|
||
if (tvdb[update.id] == null || update.time > tvdb[update.id][4]) {
|
||
try {
|
||
retry(2, 500) {
|
||
def xml = new URL("http://thetvdb.com/api/BA864DEE427E384A/series/${update.id}/en.xml").fetch().text.xml
|
||
def imdbid = xml.'**'.IMDB_ID.text()
|
||
def tvdb_name = xml.'**'.SeriesName.text()
|
||
def imdb_name = _guarded{
|
||
if (imdbid =~ /tt(\d+)/) {
|
||
def dom = IMDb.parsePage(IMDb.getMoviePageLink(imdbid.match(/tt(\d+)/) as int).toURL())
|
||
return net.sourceforge.tuned.XPathUtilities.selectString("//META[@property='og:title']/@content", dom)
|
||
}
|
||
}
|
||
def data = [update.id, imdbid ?: '', tvdb_name ?: '', imdb_name ?: '', update.time]
|
||
tvdb.put(update.id, data)
|
||
println "Update $update => $data"
|
||
}
|
||
}
|
||
catch(Throwable e) {
|
||
def data = [update.id, '', '', '', update.time]
|
||
tvdb.put(update.id, data)
|
||
println "Update $update => $data"
|
||
}
|
||
}
|
||
}
|
||
|
||
tvdb.values().findResults{ it.join('\t') }.join('\n').saveAs(tvdb_txt)
|
||
|
||
|
||
def thetvdb_index = []
|
||
tvdb.values().each{
|
||
def n1 = it[2].trim()
|
||
def n2 = it[3].replaceAll(/^(?i)(The|A)\s/, '').replaceAll(/\s&\s/, ' and ').replaceAll(/\([^\)]*\)$/, '').trim()
|
||
if (similarity(n1,n2) < 1) {
|
||
thetvdb_index << [it[0], n1]
|
||
thetvdb_index << [it[0], n2]
|
||
} else {
|
||
thetvdb_index << [it[0], n1]
|
||
}
|
||
}
|
||
|
||
|
||
thetvdb_index = thetvdb_index.findResults{ [it[0] as Integer, it[1].replaceAll(/\s+/, ' ').trim()] }.findAll{ !(it[1] =~ /(?i:duplicate)/ || it[1] =~ /\d{6,}/ || it[1].startsWith('*') || it[1].endsWith('*') || it[1].length() < 2) }
|
||
thetvdb_index = thetvdb_index.sort(new Comparator() { int compare(a, b) { a[0] <=> b[0] } })
|
||
|
||
// join and sort
|
||
def thetvdb_txt = thetvdb_index.groupBy{ it[0] }.findResults{ k, v -> ([k.pad(6)] + v*.getAt(1).unique()).join('\t') }
|
||
pack(thetvdb_out, thetvdb_txt)
|
||
println "TheTVDB Index: " + thetvdb_txt.size()
|
||
|
||
// sanity check
|
||
if (thetvdb_txt.size() < 30000) { throw new Exception('TheTVDB index sanity failed') }
|
||
|
||
|
||
/* ------------------------------------------------------------------------- */
|
||
|
||
|
||
// BUILD anidb index
|
||
def anidb = new net.sourceforge.filebot.web.AnidbClient(null, 0).getAnimeTitles()
|
||
|
||
def anidb_index = anidb.findResults{
|
||
def row = []
|
||
row += it.getAnimeId().pad(5)
|
||
row += it.effectiveNames*.replaceAll(/\s+/, ' ')*.replaceAll(/['`´‘’ʻ]+/, /'/)*.trim().unique()
|
||
return row
|
||
}
|
||
|
||
// join and sort
|
||
def anidb_txt = anidb_index.findResults{ row -> row.join('\t') }.sort().unique()
|
||
pack(anidb_out, anidb_txt)
|
||
println "AniDB Index: " + anidb_txt.size()
|
||
|
||
// sanity check
|
||
if (anidb_txt.size() < 5000) { throw new Exception('AniDB index sanity failed') }
|