def GScsv2RDF(infilename, outfilename, format="xml", withdescriptions=False):
if withdescriptions:
commonchords = ConjunctiveGraph()
commonchords.load("CommonChords.rdf")
extrachords = ConjunctiveGraph()
lines = open(infilename).readlines()
#
# Initial model bits
#
mi = mopy.MusicInfo()
homepage = mopy.foaf.Document("http://sourceforge.net/projects/motools")
mi.add(homepage)
program = mopy.foaf.Agent()
program.name = "GScsv2RDF.py"
program.homepage = homepage
mi.add(program)
tl = TimeLine("#tl")
tl.label = "Timeline derived from "+infilename
tl.maker = program
mi.add(tl)
[artistStr, titleStr] = [f.strip() for f in lines[0].split("\t")]
# Add artist & title metadata
signal = Signal()
signal.time = sig_int = Interval()
sig_int.label="Whole signal interval";
sig_int.beginsAtDuration = secondsToXSDDuration(0);
sig_int.onTimeLine = tl
signal.published_as = track = Track()
artist = MusicArtist()
artist.made = track
artist.name = artistStr
track.title = titleStr
mi.add(sig_int)
mi.add(signal)
mi.add(track)
mi.add(artist)
lineNum = 1
segmentNum = 0
thisSegment_i = None
chordSymbol=''
t_secs = 0.0
for line in lines[1:]:
# print "parsing line "+str(lineNum)
try:
lastChordSymbol = chordSymbol
t_secs = getTimestamp(line)
chordSymbol = getChordSymbol(line)
if chordSymbol != lastChordSymbol:
# print " handling new chord symbol"
segmentNum += 1
lastSegment_i = thisSegment_i
thisSegment_i = Interval("#i_"+str(segmentNum))
thisSegment_i.beginsAtDuration = secondsToXSDDuration(t_secs)
if lastSegment_i != None:
# print " terminating last interval"
lastSegment_i.endsAtDuration = secondsToXSDDuration(t_secs)
thisSegment_i.intervalAfter = lastSegment_i
lastSegment_i.intervalBefore = thisSegment_i
mi.add(thisSegment_i)
chordURI = "http://purl.org/ontology/chord/symbol/"+chordSymbol.replace("#","s").replace(",","%2C")
if withdescriptions and \
len(list(commonchords.predicate_objects(URIRef(chordURI)))) == 0 and \
len(list(extrachords.predicate_objects(URIRef(chordURI)))) == 0:
# Deref to grab chord info
print "loading <"+chordURI+">..."
extrachords.load(chordURI)
c = Chord(chordURI)
c_event = ChordEvent("#ce_"+str(segmentNum))
c_event.chord = c
c_event.time = thisSegment_i
c_event.label = chordSymbol
mi.add(c); mi.add(c_event);
# print " added new chord event for "+chordURI
except Exception, e:
print("ERROR : Problem parsing input file at line "+str(lineNum)+" !\n")
raise
lineNum+=1
def fpFile(self, filename):
"""Looks up the MusicDNS PUID for the given filename using fingerprinting
and stores the resulting info in the given graph"""
global genpuidbin, MusicDNSKey
track = Track()
audiofile = AudioFile(urlencode(os.path.basename(filename)))
track.available_as = audiofile
signal = Signal()
signal.published_as = track
mi = MusicInfo([track, audiofile, signal])
filename = clean(filename)
# TODO : If file isn't a WAV or MP3, use suitable decoder, and then pass the resulting wav to genpuid.
res_xml = os.popen(genpuidbin + " " + MusicDNSKey + " -rmd=2 -xml -noanalysis \""+filename+"\"").readlines()
retry_count=0
while ("".join(res_xml).find("unanalyzable") > 0) and (retry_count < 5):
warning("MusicDNS reports file is unanalyzable. Trying again...") # This can be caused by server hiccups
retry_count+=1
res_xml = os.popen(genpuidbin + " " + MusicDNSKey + " -rmd=2 -xml -noanalysis \""+filename+"\"").readlines()
# parse results
try:
if (res_xml[0] == res_xml[1]):
res_xml=res_xml[1:] # oddly, we see "<genpuid songs="1">\n" twice when the file is "unanalyzable"
clean_xml = "".join(res_xml).replace("mip:","") # strip out unknown prefix so minidom can parse
dom = xml.dom.minidom.parseString(clean_xml)
root = dom.getElementsByTagName("genpuid")[0]
if (root.hasAttribute("songs") == False) or (int(root.getAttribute("songs")) == 0):
return MusicInfo()
trackelem = root.getElementsByTagName("track")[0]
if (trackelem.childNodes[0].nodeName=="#text") and (trackelem.childNodes[0].data == "unavailable"):
info(" No PUID available for track : "+str(trackelem.childNodes[0].data))
return MusicInfo()
titles = trackelem.getElementsByTagName("title")
if len(titles)>0:
track.title = titles[0].childNodes[0].wholeText
artists = trackelem.getElementsByTagName("artist")
if len(artists)>0:
artistobj=MusicArtist()
artistobj.name = artists[0].getElementsByTagName("name")[0].childNodes[0].wholeText
mi.add(artistobj)
track.maker = artistobj
puid_list = trackelem.getElementsByTagName("puid-list")
puid_nodes = puid_list[0].getElementsByTagName("puid")
puids = []
for puid_node in puid_nodes:
puids.append(puid_node.getAttribute("id"))
if len(puids) == 0:
info(" No PUID available for track : "+str(trackelem.childNodes[0].data))
return MusicInfo()
elif len(puids) > 1:
warning("Multiple PUIDs returned for track !")
puid = puids[0]
signal.puid = puid
# FIXME
# years = trackelem.getElementsByTagName("first-release-date")
# if len(years)>0:
# year = years[0].childNodes[0].wholeText
# else:
# year = None
# FIXME
# release_dates = trackelem.getElementsByTagName("first-release_date")
# if len(release_dates)>0:
# release_date = release_dates[0]
# else:
# release_date = None
# FIXME
# genres = []
# genre_lists = trackelem.getElementsByTagName("genre-list")
# if len(genre_lists)>0:
# for genre_list in genre_lists:
# genre_nodes = genre_list.getElementsByTagName("genre")
# for genre_node in genre_nodes:
# genres.append(genre_node.getElementsByTagName("name")[0].childNodes[0].wholeText)
except Exception, e:
error("Failure while parsing results !")
debug("xml :\n"+"".join(res_xml))
error(str(e))
return MusicInfo()