# basic_mp3_renamer.py ^ 2016.01

# -*- coding: utf-8 -*-
# basic_mp3_renamer.py

import re
import os

tk=0
try:
    from Tkinter import Tk
    root=Tk()
    tk=1
except: print "Tkinter not found...\n\n"



aaa=u'''
DJ 99___ - It'll be The Music___ Ft The A.B.C. and Artisté___ ..mp3
05-dj_99-it'll_be_the_music...._ft_the_a.b.c._&_artisté-no.ix_rmx.mp3
05  dj 99 Ft. The A-B-C & Artisté ---- it'll be the music - remix 1.mp3
5.) dj 99 featuring The A.B.C. and Artisté - it'll be the music - rmx 2.mp3
#5 - dj 99 (Feat. The A.B.C. and L-Artisté)(No.Space Remix) - it'll be the music (~playin' loud).mp3
[5] dj 99 presents The A.B.C. and Lil' L-Artisté - it'll be the music - instrumental.mp3
{5} DJ JR. 99 - ITLL BE THE MUSIC - REMIX [ALL_VOCALS.COM].MP3.MP3
[05] dj Jr. 99 - itll be the music(~playin' loud) - remix - (((allvocals.org))).mp3
05: dj 99 - itll be the music - uk remix email@allvocals.uk ..mp3
100 - dj 99 - it'll be the music - - - (((year 2000 remix))) (free download).mp3
dj 99 - 05 - it'll be the music - original remix (lyrics in description).mp3
dj 99 - 05 - it'll be the music (the original mix) produced by kapytulz .mp3
【RnB】 dj 99 - 09 - Åññõýâñçë !!! - no.ix remix prod. by KAPYTULZ.mp3
【RnB】 dj 99 - 09 - Åññõýâñçë !!! - no.ix remix prod. by KAPYTULZ(1).mp3
dj%2099%20-%2009%20-%20Åññõýâñçë%20!!!%20【Year%202000%20RNB%20Remix】.mp3
'''
aaa=aaa.split('\n')

symbols=(
    u"абвгдеёжзийклмнопрстуфхцчшщъыьэюяАБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯÀÁÂÃÄÅÇÈÉÊËḔÌÍÎÏÏÑǸÒÓÔÕÖØŠÙÚÛÜÝàáâãäåçèéêëƒìíîïðñǹòóóôõöøšùúûüýÿı–|`【】[]{}",
    u"abvgdeejzijklmnoprstufhzcss_y_euaABVGDEEJZIJKLMNOPRSTUFHZCSS_Y_EUAAAAAAACEEEEEIIIIINNOOOOOOSUUUUYaaaaaaceeeefiiiionnooooooosuuuuyy_--'()()()")
tr={ord(a):ord(b) for a, b in zip(*symbols)}
br="[0-9a-zA-Z, '~#.$-]*"

# basic corrections
corr='''
aint=Ain't
arent=Aren't
cant=Can't
couldve=Could've
couldnt=Couldn't
couldntve=Couldn't've
didnt=Didn't
doesnt=Doesn't
dont=Don't
hadnt=Hadn't
hadntve=Hadn't've
hasnt=Hasn't
havent=Haven't
hed=He'd
hedve=He'd've
hes=He's
Id=I'd
Idve=I'd've
Im=I'm
Ive=I've
isnt=Isn't
itd=It'd
itdve=It'd've
itll=It'll
maam=Ma'am
mightve=Might've
mightnt=Mightn't
mightntve=Mightn't've
mustve=Must've
mustnt=Mustn't
mustntve=Mustn't've
oclock=O'clock
oughtnt=Oughtn't
oughtntve=Oughtn't've
shedve=She'd've
shellve=She'll've
shes=She's
shouldve=Should've
shouldnt=Shouldn't
shouldntve=Shouldn't've
thats=That's
thered=There'd
theres=There's
theyd=They'd
theyll=They'll
theyllve=They'll've
theyre=They're
theyve=They've
wasnt=Wasn't
wellve=We'll've
weve=We've
werent=Weren't
whatll=What'll
whatllve=What'll've
whatre=What're
whats=What's
whatve=What've
whens=When's
whenve=When've
whered=Where'd
wheres=Where's
whereve=Where've
wholl=Who'll
whollve=Who'll've
whos=Who's
whove=Who've
willve=Will've
wont=Won't
wontve=Won't've
wouldve=Would've
wouldnt=Wouldn't
wouldntve=Wouldn't've
yall=Y'all
yalldve=Y'all'd've
yallre=Y'all're
yallve=Y'all've
youd=You'd
youdve=You'd've
youll=You'll
youllve=You'll've
youre=You're
youve=You've
RMX=Remix
randb=RnB
r&b=R&B
Original Remix=Remix
Original Mix=Mix
produced=Prod.
Instrum\w+=(Instr)
Jr\.=Jr
Sr\.=Sr
Mr\.=Mr
Mrs\.=Mrs
Ms\.=Ms
Dr\.=Dr
Lil'=Li'l
vs\.=vs
&=and
featuring=(feat.
ft=(feat.
pres=(feat.
presenting=(feat.
presents=(feat.
'''
corr=corr.split('\n')
corr=[i.split('=') for i in corr if i]

# to reconfirm
corr2='''
cuz='Cuz
hell=He'll
hows=How's
Ill=I'll
its=It's
lets=Let's
shed=She'd
shell=She'll
whys=Why's
were=We're
wed=We'd
well=We'll
verses=vs
a\.k\.a.=aka
'''
corr2=corr2.split('\n')
corr2=[i.split('=') for i in corr2 if i]

# exclude from title formatting
excl=u'''
3D
DJ
TV
and
USA
UK
vs
RnB
(of
(aka.
(feat.
II
III
IV
VI
VII
VIII
IX
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
XIX
XX
XXI
XXII
XXIII
XXIV
XXV
XXX
'''
for z in '().':
    excl=excl.replace(z,'\\'+z)
excl=excl.split('\n')[1:-1]

years='(?: |\()?(?:19|20)\d\d(?:\)| |.)'

wo='''
free
download
official|original
karaoke
version
hd|hq
high
definition|quality
music
lyrics
audio|video
click
the
link
below
in desc(\.|ription)?
no (tags?|shouts?)?
'''.split('\n')
wo=')? ('.join(wo[1:-1])
wo='('+wo+')'

ahead,behind='(?i)(?<=\W)','(?=\W)'
bbb=[]

rootDir=r'C:\download\MP3s2chk'
#aaa=[f for f in os.listdir(rootDir) if f.lower().endswith('.mp3')]
tk=0
if tk:
    aaa=root.selection_get(selection="CLIPBOARD").split('\n')
#
fix = "&amp;:&,%20: ,_: ,\(+:(,\)+:),-+:-,'\(\d+\):,([\(\)-])\1+':'\1"
fix = [z.split(':') for z in fix.split(',')]
www = "com|net|tv|fm|ly|ws|me|cc|co|org|biz|mobi|name|int|tel|info|edu|uk|nl|eu|ru|asia|coop|de|cn"
v='[#|\(]?\d{1,2}[\.|:|\]|\)|\}| -]+'
version = 'Remix Mix Remake Dub (Instr) Remastered'.split()

def self():
    a=''

print "Processing data -- Might take a very long while to complete...\n\n"
def fixmp3s():
    def fmb(c1,c2=None):
        if not c2: c2=c1
        a=self.a
        r=re.findall(r'(?i)[\W]{1}'+c1+'[\W]{1}', a)
        if r:
            r=r[0]
            try: self.a=a.replace(r,r[0]+c2+r[-1])
            except: pass
    for a in aaa:
        if a:
            prev=a
            a, ext=os.path.splitext(a)
            a=a.translate(tr)
            print a
            a=re.sub(r'(?i)[ \(]+\S+\.(?:'+www+').+', '', a)
            for zzz,z in fix:
                a=re.sub(zzz,z,a)
            a=re.sub('''[^0-9a-zA-Z, '()-~#&.$@]''', '', a)
            if ' - ' not in a: a=a.replace('-',' - ')
            a=re.sub('^.?'+v+'[ ]?', '', a)
            a=re.sub(' - '+v, ' - ', a)

            ### To include first and last index for the following --
            a='$'+a+'$'

            ### To have both [ A.B.C. ] and [ It'll ]
            if a in [a.lower(),a.upper()]:
                a=a.title()
            else:
                for z in a.split():
                    if z is not z.upper(): a=a.replace(z,z.title())
            a="'".join([z[0].lower()+z[1:] for z in a.split("'")])
            self.a=a
            for c1,c2 in corr: fmb(c1,c2)
            for z in excl: fmb(z)
            a=self.a
            x=[]
            b=a.replace(' - ',':').replace('$','').replace('(',':(').replace(')','):').split(":")
            for z in ['(~','(-','feat']+version:
                for zzz in b:
                    if z in zzz:
                        b.remove(zzz)
                        if z in version:
                            zzz=zzz.replace(z,z+':').split(':')[0]
                        x.append('('+zzz+')')
                        break
            bbb=b[:]
            for z in bbb:
                if '(' in z:
                    x.append('('+z+')')
                    b.remove(z)
            b=' - '.join([z for z in b if len(z)>1][:2])
            x=' '.join(x)

            a=('$'+b+' '+x+'$').replace('\\','')
            for z in '()&-': a=a.replace(z*2,z)
            a=' '.join(a.split())
            a=a.replace('-)',')')
            a=a.replace('( ','(')
            a=a.replace(' )',')')
            a=a.replace(') (',')(')
            a=a.replace('$ -','')
            a=a.replace('- $','')
            a=a.replace('$','')
            a=a.replace(' .','')
            while '- -' in a: a=a.replace('- -','-')
            while '....' in a: a=a.replace('....','...')
            a=a.replace('- (',' (')
            for zzz,z in fix: # twice
                a=re.sub(zzz,z,a)
            a=' '.join(a.split())
            print a
            print

            a=a.strip()+ext.lower()
            z=''
            i=1
            while 1:
                if z+a+'\n' not in bbb:
                    bbb.append(z+a+'\n')
                    #bbb.append(prev+'\n\n')
                    break
                z='(%s) ' % i
                i+=1
#

web="""<!DOCTYPE html>
<html>
<head></head>
<body><p>@</p></body>
</html>"""

fixmp3s()
ccc='<br>'.join(bbb)
bbb=''.join(bbb)
#print bbb

import tempfile
import webbrowser

srcfilename=tempfile.mktemp(".html", "demo_")
temp=open(srcfilename, 'w')
temp.write(web.replace('@',ccc))

### webbrowser.open_new_tab(srcfilename)
### temp.write(bbb)
### os.system("notepad.exe "+srcfilename) ### char io.write limit ???

temp.close()

#print srcfilename