jazzparser.data.corpora.temperley

1 """Corpus file handling for David Temperley's corpora. 2 3 David Temperley has various corpora that he has used in his books: 4 - The Cognition of Basic Musical Structures (2001) 5 - Music and Probability (2007) 6 7 He evaluates his own techniques and others' on this data, so it is 8 an important comparison for me. It's also an important source of 9 annotated data, aside for my own small corpus. 10 11 This module provides utilities for reading in the corpus files and 12 representing the data internally. 13 14 The data formats are described in Temperley's documentation for the 15 programs that make up Melisma. 16 17 @note: This implementation is not intended to be complete. It may not handle 18 all types of input that Temperley describes. I'm just implementing 19 things as I need them. 20 21 """ 22 """ 23 ============================== License ======================================== 24 Copyright (C) 2008, 2010-12 University of Edinburgh, Mark Granroth-Wilding 25 26 This file is part of The Jazz Parser. 27 28 The Jazz Parser is free software: you can redistribute it and/or modify 29 it under the terms of the GNU General Public License as published by 30 the Free Software Foundation, either version 3 of the License, or 31 (at your option) any later version. 32 33 The Jazz Parser is distributed in the hope that it will be useful, 34 but WITHOUT ANY WARRANTY; without even the implied warranty of 35 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 36 GNU General Public License for more details. 37 38 You should have received a copy of the GNU General Public License 39 along with The Jazz Parser. If not, see <http://www.gnu.org/licenses/>. 40 41 ============================ End license ====================================== 42 43 """ 44 __author__ = "Mark Granroth-Wilding <mark.granroth-wilding@ed.ac.uk>" 45 46 import os 47 from jazzparser.data.corpora import get_corpus_file 48 49 _EVENTS = {}

50 51 -class Event(object):

52 """ 53 Superclass of all events that occur in the data. 54 55 """ 56 EVENT_NAME = None 57 """The identifier that begins the line of an input file for this event.""" 58 59 @staticmethod

60 - def from_line(line):

61 """ 62 Creates an instance of the appropriate event type given a line 63 of an input file. 64 65 All subclasses should define a C{from_line} that creates an 66 instance of them given the list of string arguments from a line 67 of an input file. 68 69 """ 70 name = line.split()[0] 71 # Consult the registry for an event for this name 72 if name not in _EVENTS: 73 raise InputError, "unknown event type %s" % name 74 return _EVENTS[name].from_line(line.split()[1:])

75

76 - def to_line(self):

77 """ 78 Generates a string representation suitable for writing out to 79 a line of a file. 80 81 """ 82 raise NotImplementedError, "%s does not implement to_line()" % type(self).__name__

83

84 - def __str__(self):

85 return "<%s>" % self.to_line()

86

87 - class __metaclass__(type):

88 - def __init__(cls, name, bases, dict):

89 event_name = dict['EVENT_NAME'] 90 if event_name is None: 91 if name != "Event": 92 raise AttributeError, "event class %s does not define "\ 93 "an event name (%s.EVENT_NAME)" % \ 94 (name,name) 95 else: 96 if event_name in _EVENTS: 97 raise AttributeError, "multiply defined event name: %s" \ 98 % event_name 99 _EVENTS[event_name] = cls

100

101 - def _get_sort_key(self):

102 return 0

103

104 - def __cmp__(self, other):

105 if hasattr(other, "_get_sort_key"): 106 return cmp(self._get_sort_key(), other._get_sort_key()) 107 else: 108 return cmp(super(Event,self),other)

109

110 -class TPCNoteEvent(Event):

111 """ 112 Tonal pitch-class note, found in TPC files and chord files. 113 A tonal pitch-class is a note as notated - i.e. it distinguishes 114 all differently named notes. They are represented as a position on 115 the line of fifths: 116 117 ...-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 ... 118 ... Ab Eb Bb F C G D A E B F# C# G# D# A#... 119 120 They are also redundantly represented as a midi note value. 121 122 """ 123 EVENT_NAME = "TPCNote" 124

125 - def __init__(self, start, end, note, pitch_class):

126 self.start = start 127 self.end = end 128 self.note = note 129 self.pitch_class = pitch_class

130 131 @staticmethod

132 - def from_line(line):

133 debug_line = " ".join(line) 134 if len(line) != 4: 135 raise InputError, "TPCNote event needs 4 arguments. %d "\ 136 "found: %s" % (len(line), debug_line) 137 return TPCNoteEvent(int(line[0]), int(line[1]), int(line[2]), int(line[3]))

138

139 - def to_line(self):

140 return "%s %d %d %d %d" % (self.EVENT_NAME, self.start, 141 self.end, self.note, self.pitch_class)

142

143 - def _get_sort_key(self):

144 return self.start

145

146 -class BeatEvent(Event):

147 """ 148 A beat identifier, giving a time of occurrence and a beat level. 149 150 """ 151 EVENT_NAME = "Beat" 152

153 - def __init__(self, time, level):

154 self.time = time 155 self.level = level

156 157 @staticmethod

158 - def from_line(line):

159 debug_line = " ".join(line) 160 if len(line) != 2: 161 raise InputError, "Beat event needs 2 arguments. %d "\ 162 "found: %s" % (len(line), debug_line) 163 return BeatEvent(int(line[0]), int(line[1]))

164

165 - def to_line(self):

166 return "%s %d %d" % (self.EVENT_NAME, self.time, 167 self.level)

168

169 - def _get_sort_key(self):

170 return self.time

171

172 -class ChordEvent(Event):

173 """ 174 A chord, identified just by its start time, end time and root. The 175 root is a position on the line of fifths. 176 177 @see: L{TPCNoteEvent} 178 179 """ 180 EVENT_NAME = "Chord" 181

182 - def __init__(self, start, end, root):

183 self.start = start 184 self.end = end 185 self.root = root

186 187 @staticmethod

188 - def from_line(line):

189 debug_line = " ".join(line) 190 if len(line) != 3: 191 raise InputError, "Chord event needs 3 arguments. %d "\ 192 "found: %s" % (len(line), debug_line) 193 return ChordEvent(int(line[0]), int(line[1]), int(line[2]))

194

195 - def to_line(self):

196 return "%s %d %d %d" % (self.EVENT_NAME, self.start, 197 self.end, self.root)

198

199 - def _get_sort_key(self):

200 return self.start

201

202 -class DataSequence(object):

203 """ 204 Data structure to store a list of events read in from a corpus 205 file. 206 207 """

208 - def __init__(self):

209 self._events = [] 210 self._event_types = {} 211 self._ordered_types = []

212

213 - def add_event(self, event):

214 self._events.append(event) 215 self._event_types.setdefault(event.EVENT_NAME, []).append(event) 216 # Keep a list of the event types in the order they appeared in the input 217 if event.EVENT_NAME not in self._ordered_types: 218 self._ordered_types.append(event.EVENT_NAME)

219

220 - def __get_events(self):

221 return list(sorted(self._events))

222 events = property(__get_events) 223

224 - def get_events_by_type(self, type):

225 """ 226 Returns a list of all events of a specific type. 227 228 """ 229 if type in self._event_types: 230 return list(sorted(self._event_types[type])) 231 else: 232 return []

233

234 - def get_grouped_events(self):

235 """ 236 Returns a list of all events, grouped by their event type, 237 with the event types in the order they were found in the input. 238 239 This should provide a form suitable for outputing to the lines 240 of a file (using C{to_line()} on each event). Alternatively, 241 just use C{to_lines()}. 242 243 """ 244 return sum([self.get_events_by_type(typ) for typ in self._ordered_types], [])

245

246 - def to_lines(self):

247 """ 248 Returns a list of lines suitable for outputing to a file. 249 250 """ 251 return [ev.to_line() for ev in self.get_grouped_events()]

252 253 @staticmethod

254 - def from_file(infile):

255 """ 256 Creates a new DataSequence to represent the data in a file. 257 258 @type infile: str or open file object 259 @param infile: filename or file object. Filename may be the 260 path to the file or the path within the corpus. 261 262 """ 263 if type(infile) == str: 264 if not os.path.exists(infile): 265 # Try getting it from the corpus instead 266 infile = get_corpus_file('kp', infile) 267 infile = open(infile, 'r') 268 269 lines = infile.readlines() 270 dseq = DataSequence() 271 for line in lines: 272 dseq.add_event(Event.from_line(line)) 273 return dseq

274

275 - def __str__(self):

276 evs = self.events 277 if len(evs) > 10: 278 evs = evs[:9] + ["...", evs[-1]] 279 return "<Temperley: %s>" % ", ".join([str(ev) for ev in evs])

280

281 - def __iter__(self):

282 return iter(self.events)

283

284 - def to_midi(self):

285 """ 286 Constructs a L{MIDI EventStream<midi.EventStream>} from the 287 data in this stream. 288 This can then be output to a file to be played. 289 290 Note that TPCNotes will be output as normal MIDI notes. We 291 can't do anything of the clever tuning stuff that we can do 292 with tonal space coordinates, since we'd need to do a further 293 step of analysis to work out the fully specified TS point from 294 the pitch class. 295 296 """ 297 tempo = 120 298 299 from midi import EventStream, NoteOffEvent, NoteOnEvent, SetTempoEvent 300 mid = EventStream() 301 mid.add_track() 302 # Set the tempo first at the beginning 303 temp = SetTempoEvent() 304 temp.tempo = tempo 305 temp.tick = 0 306 mid.add_event(temp) 307 # Work out how many ticks there are in a millisecond 308 ticks_per_ms = float(mid.resolution) * tempo / 60000 309 # Create midi events for every event in our stream 310 for ev in self.events: 311 if isinstance(ev, TPCNoteEvent): 312 # Create note-on and note-off events 313 note = ev.note 314 315 noteon = NoteOnEvent() 316 noteon.pitch = note 317 noteon.tick = int(ev.start * ticks_per_ms) 318 noteon.velocity = 100 319 mid.add_event(noteon) 320 321 noteoff = NoteOffEvent() 322 noteoff.pitch = note 323 noteoff.tick = int(ev.end * ticks_per_ms) 324 noteoff.velocity = 100 325 mid.add_event(noteoff) 326 elif isinstance(ev, (ChordEvent,BeatEvent)): 327 # These events don't affect the midi data 328 continue 329 else: 330 raise TypeError, "event type %s not recognised by "\ 331 "MIDI converter." % type(ev).__name__ 332 return mid

333 334 read_file = DataSequence.from_file

335 336 -class InputError(Exception):

337 pass

338

Source Code for Module jazzparser.data.corpora.temperley