# -*- coding: utf-8 -*- # ------------------------------------------------------------------------------- # Name: tinyNotation.py # Purpose: A simple notation input format. # # Authors: Michael Scott Asato Cuthbert # # Copyright: Copyright © 2009-2024 Michael Scott Asato Cuthbert # License: BSD, see license.txt # ------------------------------------------------------------------------------- ''' tinyNotation is a simple way of specifying single line melodies that uses a notation somewhat similar to Lilypond but with WAY fewer options. It was originally developed to notate trecento (medieval Italian) music, but it is pretty useful for a lot of short examples, so we have made it a generally supported music21 format. N.B.: TinyNotation is not meant to expand to cover every single case. Instead, it is meant to be subclassable to extend to the cases *your* project needs. Here are the most important rules by default: 1. Note names are: a,b,c,d,e,f,g and r for rest 2. Flats, sharps, and naturals are notated as #,- (not b), and (if needed) n. If the accidental is above the staff (i.e., editorial), enclose it in parentheses: (#), etc. Make sure that flats in the key signatures are explicitly specified. 3. Note octaves are specified as follows:: CC to BB = from C below bass clef to second-line B in bass clef C to B = from bass clef C to B below middle C. c to b = from middle C to the middle of treble clef c' to b' = from C in treble clef to B above treble clef Octaves below and above these are specified by further doublings of letter (CCC) or apostrophes (c'') -- this is one of the note name standards found in many music theory books. 4. After the note name, a number may be placed indicating the note length: 1 = whole note, 2 = half, 4 = quarter, 8 = eighth, 16 = sixteenth. etc. If the number is omitted then it is assumed to be the same as the previous note. I.e., c8 B c d is a string of eighth notes. 5. After the number, a ~ can be placed to show a tie to the next note. A "." indicates a dotted note. (If you are entering data via Excel or other spreadsheet, be sure that "capitalize the first letter of sentences" is turned off under "Tools->AutoCorrect," otherwise the next letter will be capitalized, and the octave will be screwed up.) 6. For triplets use this notation: `trip{c4 d8}` indicating that these two notes both have "3s" over them. For 4 in the place of 3, use `quad{c16 d e8}`. No other tuplets are supported. Here is an example of TinyNotation in action. >>> stream1 = converter.parse("tinyNotation: 3/4 E4 r f# g=lastG trip{b-8 a g} c4~ c") >>> stream1.show('text') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {3.0} {0.0} {1.0} {1.3333} {1.6667} {2.0} {6.0} {0.0} {1.0} >>> stream1.recurse().getElementById('lastG').step 'G' >>> stream1.flatten().notesAndRests[1].isRest True >>> stream1.flatten().notesAndRests[0].octave 3 >>> stream1.flatten().notes[-2].tie.type 'start' >>> stream1.flatten().notes[-1].tie.type 'stop' Changing time signatures are supported: >>> s1 = converter.parse('tinynotation: 3/4 C4 D E 2/4 F G A B 1/4 c') >>> s1.show('t') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {3.0} {0.0} {0.0} {1.0} {5.0} {0.0} {1.0} {7.0} {0.0} {0.0} {1.0} Here is an equivalent way of doing the example above, but using the lower level :class:`music21.tinyNotation.Converter` object: >>> tnc = tinyNotation.Converter('3/4 E4 r f# g=lastG trip{b-8 a g} c4~ c') >>> stream2 = tnc.parse().stream >>> len(stream1.recurse()) == len(stream2.recurse()) True This lower level is needed in case you want to add additional features. For instance, here we will set the "modifierStar" to change the color of notes: >>> class ColorModifier(tinyNotation.Modifier): ... def postParse(self, m21Obj): ... m21Obj.style.color = self.modifierData ... return m21Obj >>> tnc = tinyNotation.Converter('3/4 C4*pink* D4*green* E4*blue*') >>> tnc.modifierStar = ColorModifier >>> s = tnc.parse().stream >>> for n in s.recurse().getElementsByClass(note.Note): ... print(n.step, n.style.color) C pink D green E blue Or more usefully, and often desired: >>> class HarmonyModifier(tinyNotation.Modifier): ... def postParse(self, n): ... cs = harmony.ChordSymbol(n.pitch.name + self.modifierData) ... cs.duration = n.duration ... return cs >>> tnc = tinyNotation.Converter('4/4 C2_maj7 D4_m E-_sus4') >>> tnc.modifierUnderscore = HarmonyModifier >>> s = tnc.parse().stream >>> s.show('text') {0.0} {0.0} {0.0} {0.0} {2.0} {3.0} {4.0} >>> for cs in s.recurse().getElementsByClass(harmony.ChordSymbol): ... print([p.name for p in cs.pitches]) ['C', 'E', 'G', 'B'] ['D', 'F', 'A'] ['E-', 'A-', 'B-'] The supported modifiers are: * `=data` (`modifierEquals`, default action is to set `.id`) * `_data` (`modifierUnderscore`, default action is to set `.lyric`) * `[data]` (`modifierSquare`, no default action) * `` (`modifierAngle`, no default action) * `(data)` (`modifierParens`, no default action) * `*data*` (`modifierStar`, no default action) Another example: TinyNotation does not support key signatures -- well, no problem! Let's create a new Token type and add it to the tokenMap >>> class KeyToken(tinyNotation.Token): ... def parse(self, parent): ... keyName = self.token ... return key.Key(keyName) >>> keyMapping = (r'k(.*)', KeyToken) >>> tnc = tinyNotation.Converter('4/4 kE- G1 kf# A1') >>> tnc.tokenMap.append(keyMapping) >>> s = tnc.parse().stream >>> s.show('text') {0.0} {0.0} {0.0} {0.0} {0.0} {4.0} {0.0} {0.0} {4.0} TokenMap should be passed a string, representing a regular expression with exactly one group (which can be the entire expression), and a subclass of :class:`~music21.tinyNotation.Token` which will handle the parsing of the string. Tokens can take advantage of the `parent` variable, which is a reference to the `Converter` object, to use the `.stateDict` dictionary to store information about state. For instance, the `NoteOrRestToken` uses `parent.stateDict['lastDuration']` to get access to the last duration. There is also the concept of "State" which affects multiple tokens. The best way to create a new State is to define a subclass of the :class:`~music21.tinyNotation.State` and add it to `bracketStateMapping` of the converter. Here's one that a lot of people have asked for over the years: >>> class ChordState(tinyNotation.State): ... def affectTokenAfterParse(self, n): ... super().affectTokenAfterParse(n) ... return None # do not append Note object ... def end(self): ... ch = chord.Chord(self.affectedTokens) ... ch.duration = self.affectedTokens[0].duration ... return ch >>> tnc = tinyNotation.Converter("2/4 C4 chord{C4 e g'} F.4 chord{D8 F# A}") >>> tnc.bracketStateMapping['chord'] = ChordState >>> s = tnc.parse().stream >>> s.show('text') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {0.0} {1.5} {2.0} If you want to create a very different dialect, you can subclass tinyNotation.Converter and set it up once to use the mappings above. See :class:`~music21.alpha.trecento.notation.TrecentoTinyConverter` (especially the code) for details on how to do that. ''' from __future__ import annotations import collections import copy import fractions import re import typing # not importing as t, because of extensive preexisting use of `t` as token import unittest from music21 import note from music21 import duration from music21 import environment from music21 import exceptions21 from music21 import stream from music21 import tie from music21 import expressions from music21 import meter from music21 import pitch if typing.TYPE_CHECKING: from music21.base import Music21Object environLocal = environment.Environment('tinyNotation') class TinyNotationException(exceptions21.Music21Exception): pass class State: ''' State tokens apply something to every note found within it. State objects can have "autoExpires" set, which is False if it does not expire or an integer if it expires after a certain number of tokens have been processed. >>> tnc = tinyNotation.Converter() >>> ts = tinyNotation.TieState(tnc, '~') >>> isinstance(ts, tinyNotation.State) True >>> ts.autoExpires 2 ''' # expires after N tokens or never. autoExpires: typing.Literal[False]|int = False def __init__(self, parent: Converter|None = None, stateInfo: str|None = None): self.affectedTokens: list[Music21Object] = [] self.parent: Converter|None = parent self.stateInfo: str|None = stateInfo # print('Adding state', self, parent.activeStates) def start(self): ''' called when the state is initiated ''' pass def end(self) -> Music21Object|None: ''' called just after removing state ''' return None def affectTokenBeforeParse(self, tokenStr: str) -> str: ''' called to modify the string of a token. ''' return tokenStr def affectTokenAfterParseBeforeModifiers( self, m21Obj: Music21Object ) -> Music21Object: ''' called after the object has been acquired but before modifiers have been applied. ''' return m21Obj def affectTokenAfterParse( self, m21Obj: Music21Object ) -> Music21Object: ''' called to modify the tokenObj after parsing tokenObj may be None if another state has deleted it. ''' self.affectedTokens.append(m21Obj) if self.autoExpires is not False: if len(self.affectedTokens) == self.autoExpires: self.end() # this is a hack that should be done away with p = self.parent if p is None: return m21Obj # I thought this was potentially O(n^2) but it's actually # just O(n) on activeStates and on pop() operation. # most of the time, backCount will be -1 and pop(-1) # is O(1). in fact pop(-m) seems to be O(m) not O(n)! for i in range(len(p.activeStates)): backCount = -1 * (i + 1) if p.activeStates[backCount] is self: p.activeStates.pop(backCount) break return m21Obj class TieState(State): ''' A TieState is an auto-expiring state that applies a tie start to this note and a tie stop to the next note. ''' autoExpires = 2 def end(self): ''' end the tie state by applying tie ties to the appropriate notes ''' if self.affectedTokens[0].tie is None: self.affectedTokens[0].tie = tie.Tie('start') else: self.affectedTokens[0].tie.type = 'continue' if len(self.affectedTokens) > 1: # could be the end. self.affectedTokens[1].tie = tie.Tie('stop') class TupletState(State): ''' a tuplet state applies tuplets to notes while parsing and sets 'start' and 'stop' on the first and last note when end is called. ''' actual = 3 normal = 2 def end(self): ''' end a tuplet by putting start on the first note and stop on the last. ''' if not self.affectedTokens: return None self.affectedTokens[0].duration.tuplets[0].type = 'start' self.affectedTokens[-1].duration.tuplets[0].type = 'stop' return None def affectTokenAfterParse(self, n): ''' puts a tuplet on the note ''' super().affectTokenAfterParse(n) newTup = duration.Tuplet() newTup.durationActual = duration.durationTupleFromTypeDots(n.duration.type, 0) newTup.durationNormal = duration.durationTupleFromTypeDots(n.duration.type, 0) newTup.numberNotesActual = self.actual newTup.numberNotesNormal = self.normal n.duration.appendTuplet(newTup) return n class TripletState(TupletState): ''' a 3:2 tuplet ''' actual = 3 normal = 2 class QuadrupletState(TupletState): ''' a 4:3 tuplet ''' actual = 4 normal = 3 class Modifier: ''' a modifier is something that changes the current token, like setting the `.id` or Lyric. ''' def __init__(self, modifierData, modifierString, parent): self.modifierData = modifierData self.modifierString = modifierString self.parent = parent def preParse(self, tokenString): ''' called before the tokenString has been turned into an object ''' pass def postParse(self, m21Obj): ''' called after the tokenString has been turned into an m21Obj. m21Obj may be None Important: must return the m21Obj, or a different object! ''' return m21Obj class IdModifier(Modifier): ''' sets the .id of the m21Obj, called with = by default ''' def postParse(self, m21Obj): if hasattr(m21Obj, 'id'): m21Obj.id = self.modifierData return m21Obj class LyricModifier(Modifier): ''' sets the .lyric of the m21Obj, called with _ by default ''' def postParse(self, m21Obj): if hasattr(m21Obj, 'lyric'): m21Obj.lyric = self.modifierData return m21Obj class Token: ''' A single token made from the parser. Call .parse(parent) to make it work. ''' def __init__(self, token=''): self.token = token def parse(self, parent): ''' do NOT store parent -- probably too slow ''' return None class TimeSignatureToken(Token): ''' Represents a single time signature, like 1/4 ''' def parse(self, parent): tsObj = meter.TimeSignature(self.token) parent.stateDict['currentTimeSignature'] = tsObj return tsObj class NoteOrRestToken(Token): ''' represents a Note or Rest. Chords are represented by Note objects ''' def __init__(self, token=''): super().__init__(token) self.durationMap = [ (r'(\d+)', 'durationType'), (r'(\.+)', 'dots'), ] # tie will be dealt with later. self.durationFound = False def applyDuration(self, n, t, parent): ''' takes the information in the string `t` and creates a Duration object for the note or rest `n`. ''' for pm, method in self.durationMap: searchSuccess = re.search(pm, t) if searchSuccess: callFunc = getattr(self, method) t = callFunc(n, searchSuccess, pm, t, parent) if self.durationFound is False and hasattr(parent, 'stateDict'): n.duration.quarterLength = parent.stateDict['lastDuration'] # do this by quarterLength here, so that applied tuplets do not persist. if hasattr(parent, 'stateDict'): parent.stateDict['lastDuration'] = n.duration.quarterLength return t def durationType(self, element, search, pm, t, parent): ''' The result of a successful search for a duration type: puts a Duration in the right place. ''' self.durationFound = True typeNum = int(search.group(1)) if typeNum == 0: if parent.stateDict['currentTimeSignature'] is not None: element.duration = copy.deepcopy( parent.stateDict['currentTimeSignature'].barDuration ) element.expressions.append(expressions.Fermata()) else: try: element.duration.type = duration.typeFromNumDict[typeNum] except KeyError as ke: raise TinyNotationException( f'Cannot parse token with duration {typeNum}' ) from ke t = re.sub(pm, '', t) return t def dots(self, element, search, pm, t, parent): ''' adds the appropriate number of dots to the right place. Subclassed in TrecentoNotation where two dots has a different meaning. ''' element.duration.dots = len(search.group(1)) t = re.sub(pm, '', t) return t class RestToken(NoteOrRestToken): ''' A token starting with 'r', representing a rest. ''' def parse(self, parent: Converter|None = None): r = note.Rest() self.applyDuration(r, self.token, parent) return r class NoteToken(NoteOrRestToken): ''' A NoteToken represents a single Note with pitch >>> c3 = tinyNotation.NoteToken('C') >>> c3 >>> n = c3.parse() >>> n >>> n.nameWithOctave 'C3' >>> bFlat6 = tinyNotation.NoteToken("b''-") >>> bFlat6 >>> n = bFlat6.parse() >>> n >>> n.nameWithOctave 'B-6' ''' pitchMap = collections.OrderedDict([ ('lowOctave', r'([A-G]+)'), ('highOctave', r'([a-g])(\'*)'), ('editorialAccidental', r'\(([\#\-n]+)\)(.*)'), ('sharps', r'(\#+)'), ('flats', r'(\-+)'), ('natural', r'(n)'), ]) def __init__(self, token=''): super().__init__(token) self.isEditorial = False def parse(self, parent: Converter|None = None): ''' Extract the pitch from the note and then returns the Note. ''' t = self.token n = note.Note() t = self.processPitchMap(n, t) if parent: self.applyDuration(n, t, parent) return n def processPitchMap(self, n, t): ''' processes the pitchMap on the object. ''' for method, pm in self.pitchMap.items(): searchSuccess = re.search(pm, t) if searchSuccess: callFunc = getattr(self, method) t = callFunc(n, searchSuccess, pm, t) return t def editorialAccidental(self, n, search, pm, t): ''' indicates that the accidental is in parentheses, so set it up to be stored in ficta. ''' self.isEditorial = True t = search.group(1) + search.group(2) return t def _addAccidental(self, n, alter, pm, t): # noinspection PyShadowingNames r''' helper function for all accidental types. >>> nToken = tinyNotation.NoteToken('BB--') >>> n = note.Note('B') >>> n.octave = 2 >>> tPost = nToken._addAccidental(n, -2, r'(\-+)', 'BB--') >>> tPost 'BB' >>> n.pitch.accidental >>> nToken = tinyNotation.NoteToken('BB(--)') >>> nToken.isEditorial = True >>> n = note.Note('B') >>> n.octave = 2 >>> tPost = nToken._addAccidental(n, -2, r'(\-+)', 'BB--') >>> tPost 'BB' >>> n.editorial.ficta ''' acc = pitch.Accidental(alter) if self.isEditorial: n.editorial.ficta = acc else: n.pitch.accidental = acc t = re.sub(pm, '', t) return t def sharps(self, n, search, pm, t): # noinspection PyShadowingNames r''' called when one or more sharps have been found and adds the appropriate accidental to it. >>> import re >>> tStr = 'C##' >>> nToken = tinyNotation.NoteToken(tStr) >>> n = note.Note('C') >>> n.octave = 3 >>> searchResult = re.search(nToken.pitchMap['sharps'], tStr) >>> tPost = nToken.sharps(n, searchResult, nToken.pitchMap['sharps'], tStr) >>> tPost 'C' >>> n.pitch.accidental ''' alter = len(search.group(1)) return self._addAccidental(n, alter, pm, t) def flats(self, n, search, pm, t): # noinspection PyShadowingNames ''' called when one or more flats have been found and calls adds the appropriate accidental to it. >>> import re >>> tStr = 'BB--' >>> nToken = tinyNotation.NoteToken(tStr) >>> n = note.Note('B') >>> n.octave = 2 >>> searchResult = re.search(nToken.pitchMap['flats'], tStr) >>> tPost = nToken.flats(n, searchResult, nToken.pitchMap['flats'], tStr) >>> tPost 'BB' >>> n.pitch.accidental ''' alter = -1 * len(search.group(1)) return self._addAccidental(n, alter, pm, t) def natural(self, n, search, pm, t): # noinspection PyShadowingNames ''' called when an explicit natural has been found. All pitches are natural without being specified, so not needed. Adds a natural accidental to it. >>> import re >>> tStr = 'En' >>> nToken = tinyNotation.NoteToken(tStr) >>> n = note.Note('E') >>> n.octave = 3 >>> searchResult = re.search(nToken.pitchMap['natural'], tStr) >>> tPost = nToken.natural(n, searchResult, nToken.pitchMap['natural'], tStr) >>> tPost 'E' >>> n.pitch.accidental ''' return self._addAccidental(n, 0, pm, t) def lowOctave(self, n, search, pm, t): # noinspection PyShadowingNames ''' Called when a note of octave 3 or below is encountered. >>> import re >>> tStr = 'BBB' >>> nToken = tinyNotation.NoteToken(tStr) >>> n = note.Note('B') >>> searchResult = re.search(nToken.pitchMap['lowOctave'], tStr) >>> tPost = nToken.lowOctave(n, searchResult, nToken.pitchMap['lowOctave'], tStr) >>> tPost '' >>> n.octave 1 ''' stepName = search.group(1)[0].upper() octaveNum = 4 - len(search.group(1)) n.step = stepName n.octave = octaveNum t = re.sub(pm, '', t) return t def highOctave(self, n, search, pm, t): # noinspection PyShadowingNames ''' Called when a note of octave 4 or higher is encountered. >>> import re >>> tStr = "e''" >>> nToken = tinyNotation.NoteToken(tStr) >>> n = note.Note('E') >>> searchResult = re.search(nToken.pitchMap['highOctave'], tStr) >>> tPost = nToken.highOctave(n, searchResult, nToken.pitchMap['highOctave'], tStr) >>> tPost '' >>> n.octave 6 ''' stepName = search.group(1)[0].upper() octaveNum = 4 + len(search.group(2)) n.step = stepName n.octave = octaveNum t = re.sub(pm, '', t) return t def _getDefaultTokenMap() -> list[tuple[str, type[Token]]]: ''' Returns the default tokenMap for TinyNotation. Based on the following grammar (in Extended Backus-Naur form) (https://en.wikipedia.org/wiki/Extended_Backus%E2%80%93Naur_form) (* Items in parentheses are grouped *) (* Items in curly braces appear zero or more times *) (* Items in square brackets may appear exactly zero or one time *) (* Items in double quotes are literal strings *) (* Items between question marks should be interpreted as English *) (* Each rule is ended by a semicolon *) TINY-NOTATION = TOKEN, { WHITESPACE, TOKEN } ; WHITESPACE = ( " " | ? Carriage return ? ) , { " " | ? Carriage return ? } ; TOKEN = ( TIME-SIGNATURE | TUPLET | REST | NOTE ); TIME-SIGNATURE = INTEGER, "/", INTEGER ; INTEGER = DIGIT, { DIGIT } ; DIGIT = ( "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9" ) ; TUPLET = ( "trip" | "quad" | ALPHANUMERIC ), "{", [ WHITESPACE ], ( REST | NOTE ), { WHITESPACE, ( REST | NOTE ) }, [ WHITESPACE ], "}" ; REST = "r", [ DURATION ], [ MODIFIER ] ; DURATION = ( EVEN-NUMBER, { "." } | { "." }, EVEN-NUMBER | ".", { "." } ) ; EVEN-NUMBER = { INTEGER }, ( "0" | "2" | "4" | "6" | "8" ) ; NOTE = PITCH, [ DURATION ], [ TIE ], { MODIFIER } ; PITCH = ( ( LOW-A | LOW-B | LOW-C | LOW-D | LOW-E | LOW-F | LOW-G ), [ ACCIDENTAL ] | ( "a" | "b" | "c" | "d" | "e" | "f" | "g" ), [ ACCIDENTAL ], { "'" } | ( "a" | "b" | "c" | "d" | "e" | "f" | "g" ), { "'" }, [ ACCIDENTAL ] ) ; LOW-A = "A", { "A" } ; LOW-B = "B", { "B" } ; LOW-C = "C", { "C" } ; LOW-D = "D", { "D" } ; LOW-E = "E", { "E" } ; LOW-F = "F", { "F" } ; LOW-G = "G", { "G" } ; ACCIDENTAL = ( EDITORIAL | SHARPS | FLATS | NATURAL ) ; EDITORIAL = "(", ( SHARPS | FLATS | NATURAL ), ")" ; SHARPS = "#", { "#" } ; FLATS = "-", { "-" } ; NATURAL = "n" ; TIE = "~" ; MODIFIER = ( EQUALS-MODIFIER | UNDERSCORE-MODIFIER | SQUARE-MODIFIER | ANGLE-MODIFIER | PARENS-MODIFIER | STAR-MODIFIER ) ; EQUALS-MODIFIER = "=", EQUALS-DATA ; UNDERSCORE-MODIFIER = "_", UNDERSCORE-DATA ; SQUARE-MODIFIER = "[", ALPHANUMERIC, "]" ; ANGLE-MODIFIER = "<", ALPHANUMERIC, ">" ; PARENS-MODIFIER = "(", ALPHANUMERIC, ")" ; STAR-MODIFIER = "*", ALPHANUMERIC, "*" ; (* The following is just shorthand. *) ALPHANUMERIC = ? At least one alphanumeric character. So "a-z", "A-Z", or "0-9" ? ; EQUALS-DATA = ? At least one non-whitespace, non-"_" character. ? ; UNDERSCORE-DATA = ? At least one non-whitespace, non-"=" character. ? ; ''' sharpsFlatsOrNaturalRegex = r'#+|-+|n' editorialRegex = fr'\((?:{sharpsFlatsOrNaturalRegex})\)' accidentalRegex = fr'{editorialRegex}|(?:{sharpsFlatsOrNaturalRegex})' lowNoteRegex = fr'(?:A+|B+|C+|D+|E+|F+|G+)(?:{accidentalRegex})?' highNoteRegex = ( r'(?:a|b|c|d|e|f|g)' + fr"(?:(?:{accidentalRegex})?'*|'*(?:{accidentalRegex})?)" ) noteNameRegex = fr'{lowNoteRegex}|{highNoteRegex}' durationRegex = r'\d+\.*|\.*\d+|\.+' tieStateRegex = r'~' equalsRegex = r'=[^\s_]*' starRegex = r'\*.*?\*' angleRegex = r'<.*?>' parensRegex = r'\(.*?\)' squareRegex = r'\[.*?]' underscoreRegex = r'_[^\s=]' modifierRegex = ( fr'{equalsRegex}|{starRegex}|{angleRegex}|' + fr'{parensRegex}|{squareRegex}|{underscoreRegex}' ) return [ (r'^(\d+\/\d+)$', TimeSignatureToken), ( fr'^r((?:{durationRegex})?(?:{modifierRegex})*)$', RestToken ), ( ( fr'^((?:{noteNameRegex})(?:{durationRegex})?' + fr'(?:{tieStateRegex})?(?:{modifierRegex})*)$' ), NoteToken ), # last ] _stateDictDefault = { 'currentTimeSignature': None, 'lastDuration': 1.0 } class Converter: # noinspection GrazieInspection ''' Main conversion object for TinyNotation. Accepts keywords: * `makeNotation=False` to get "classic" TinyNotation formats without measures, Clefs, etc. * `raiseExceptions=True` to make errors become exceptions. Generally, users should just use the general music21 converter (lowercase c) parse function and do not need to be in this module at all: >>> part = converter.parse('4/4 C4 D4 E2 F1', format='tinyNotation') >>> part.show('text') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {4.0} {0.0} {4.0} But for advanced work, create a tinyNotation.Converter object as shown: >>> tnc = tinyNotation.Converter('4/4 C##4 D e-8 f~ f f# g4 trip{f8 e d} C2=hello') Run the parsing routine with `.parse()` >>> tnc.parse() And now the `.stream` attribute of the converter object has the Stream (generally Part) ready to send out: >>> tnc.stream And all normal Stream methods are available on it: >>> tnc.stream.show('text') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {2.5} {3.0} {3.5} {4.0} {0.0} {1.0} {1.3333} {1.6667} {2.0} {4.0} Or, breaking down what Parse does bit by bit: >>> tnc = tinyNotation.Converter('4/4 C##4 D e-8 f~ f f# g4 trip{f8 e d} C2=hello') >>> tnc.stream >>> tnc.makeNotation True >>> tnc.stringRep '4/4 C##4 D e-8 f~ f f# g4 trip{f8 e d} C2=hello' >>> tnc.activeStates [] >>> tnc.preTokens [] >>> tnc.splitPreTokens() >>> tnc.preTokens ['4/4', 'C##4', 'D', 'e-8', 'f~', 'f', 'f#', 'g4', 'trip{f8', 'e', 'd}', 'C2=hello'] >>> tnc.setupRegularExpressions() Then we parse the time signature. >>> tnc.parseOne(0, tnc.preTokens[0]) >>> tnc.stream.coreElementsChanged() >>> tnc.stream.show('text') {0.0} (Note that because we are calling `show()` after each note in these docs, but TinyNotation uses high efficiency `stream.core` routines, we need to set the stream to a stable-state by calling `coreElementsChanged` after each call. You would not need to do this in your own subclasses, since that would lose the `O(n)` efficiency when parsing) Then parse the first actual note: >>> tnc.parseOne(1, tnc.preTokens[1]) >>> tnc.stream.coreElementsChanged() >>> tnc.stream.show('text') {0.0} {0.0} The next notes to 'g4' are pretty similar: >>> for i in range(2, 8): ... tnc.parseOne(i, tnc.preTokens[i]) >>> tnc.stream.coreElementsChanged() >>> tnc.stream.show('text') {0.0} {0.0} {1.0} {2.0} {2.5} {3.0} {3.5} {4.0} The next note starts a "State" since it has a triplet: >>> tnc.preTokens[8] 'trip{f8' >>> tnc.parseOne(8, tnc.preTokens[8]) >>> tnc.activeStates [] >>> tnc.activeStates[0].affectedTokens [] The state is still active for the next token: >>> tnc.preTokens[9] 'e' >>> tnc.parseOne(9, tnc.preTokens[9]) >>> tnc.activeStates [] >>> tnc.activeStates[0].affectedTokens [, ] The :class:`~music21.tinyNotation.TripletState` state adds tuplets along the way, but does not set their type. >>> f_starts_tuplet = tnc.activeStates[0].affectedTokens[0] >>> f_starts_tuplet.duration.tuplets (,) >>> f_starts_tuplet.duration.tuplets[0].type is None True But the next token closes the state: >>> tnc.preTokens[10] 'd}' >>> tnc.parseOne(10, tnc.preTokens[10]) >>> tnc.activeStates [] And this sets the tupet types for the F and D properly: >>> f_starts_tuplet.duration.tuplets[0].type 'start' >>> tnc.stream.coreElementsChanged() >>> tnc.stream.show('text') {0.0} ... {4.0} {5.0} {5.3333} {5.6667} The last token has a modifier, which is an IdModifier: >>> tnc.preTokens[11] 'C2=hello' >>> tnc.parseOne(11, tnc.preTokens[11]) >>> tnc.stream.coreElementsChanged() >>> tnc.stream.show('text') {0.0} ... {5.6667} {6.0} >>> tnc.stream[-1].id 'hello' At this point the flattened stream is ready, if Converter.makeNotation is False, then not much more needs to happen, but it is True by default. So, then calling tnc.postParse() runs the makeNotation: >>> tnc.postParse() >>> tnc.stream.show('text') {0.0} {0.0} {0.0} {0.0} {1.0} {2.0} {2.5} {3.0} {3.5} {4.0} {0.0} {1.0} {1.3333} {1.6667} {2.0} {4.0} Let's look at edge cases. Normally, invalid notes or other bad tokens pass freely by dropping the unparseable token. Here `3` is not a valid duration, so `d3` will be dropped. >>> x = converter.parse('tinyNotation: 4/4 c2 d3 e2') >>> x.show('text') {0.0} {0.0} {0.0} {0.0} {2.0} {4.0} But with the keyword 'raiseExceptions=True' a `TinyNotationException` is raised if any token cannot be parsed. >>> x = converter.parse('tinyNotation: 4/4 c2 d3 e2', raiseExceptions=True) Traceback (most recent call last): music21.tinyNotation.TinyNotationException: Could not parse token: 'd3' ''' bracketStateMapping = { 'trip': TripletState, 'quad': QuadrupletState, } _modifierEqualsRe = re.compile(r'=([A-Za-z0-9]*)') _modifierStarRe = re.compile(r'\*(.*?)\*') _modifierAngleRe = re.compile(r'<(.*?)>') _modifierParensRe = re.compile(r'\((.*?)\)') _modifierSquareRe = re.compile(r'\[(.*?)]') _modifierUnderscoreRe = re.compile(r'_(.*)') def __init__( self, stringRep: str = '', *, makeNotation: bool = True, raiseExceptions: bool = False, **_keywords ): self.stream = stream.Part() self.stateDict = copy.copy(_stateDictDefault) self.stringRep = stringRep self.activeStates: list[State] = [] self.preTokens: list[str] = [] self.generalBracketStateRe = re.compile(r'(\w+){') self.tieStateRe = re.compile(r'~') self.tokenMap = _getDefaultTokenMap() self.modifierEquals = IdModifier self.modifierStar = None self.modifierAngle = None self.modifierParens = None self.modifierSquare = None self.modifierUnderscore = LyricModifier self.makeNotation = makeNotation self.raiseExceptions = raiseExceptions # will be filled by self.setupRegularExpressions() self._tokenMapRe: list[tuple[typing.Pattern, type]] = [] def load(self, stringRep: str) -> None: ''' Loads a stringRepresentation into `.stringRep` and resets the parsing state. >>> tnc = tinyNotation.Converter() >>> tnc.load('4/4 c2 d e f') >>> s = tnc.parse().stream >>> tnc.load('4/4 f e d c') >>> s2 = tnc.parse().stream >>> ns2 = s2.flatten().notes Check that the duration of 2.0 from the first load did not carry over. >>> ns2[0].duration.quarterLength 1.0 >>> len(ns2) 4 ''' # NOTE(msc): any changes here have to be reflected in the constructor # which does not call this routine. self.stream = stream.Part() self.stateDict = copy.copy(_stateDictDefault) self.activeStates = [] self.preTokens = [] self.stringRep = stringRep def splitPreTokens(self): ''' splits the string into textual preTokens. Right now just splits on spaces, but might be smarter to ignore spaces in quotes, etc. later. ''' self.preTokens = self.stringRep.split() # do something better alter. def setupRegularExpressions(self): ''' Regular expressions get compiled for faster usage. This is called automatically by .parse(), but can be called separately for testing. It is also important that it is not called in __init__ since subclasses should override the tokenMap, etc. for a class. ''' self._tokenMapRe = [] for rePre, classCall in self.tokenMap: try: self._tokenMapRe.append((re.compile(rePre), classCall)) except re.error as e: raise TinyNotationException( f'Error in compiling token, {rePre}: {e}' ) from e def parse(self): ''' splitPreTokens, setupRegularExpressions, then runs through each preToken, and runs postParse. ''' if not self.preTokens and self.stringRep != '': self.splitPreTokens() if not self._tokenMapRe: self.setupRegularExpressions() for i, t in enumerate(self.preTokens): self.parseOne(i, t) self.postParse() return self def parseOne(self, i: int, t: str): ''' parse a single token at position i, with text t, possibly adding it to the stream. Checks for state changes, modifiers, tokens, and end-state brackets. ''' t_orig = t t = self.parseStartStates(t) t, numberOfStatesToEnd = self.parseEndStates(t) t, activeModifiers = self.parseModifiers(t) # this copy is done so that an activeState can # remove itself from this list: for stateObj in self.activeStates[:]: t = stateObj.affectTokenBeforeParse(t) m21Obj = None tokenObj = None # parse token with state: hasMatch = False for tokenRe, tokenClass in self._tokenMapRe: matchSuccess = tokenRe.match(t) if matchSuccess is None: continue hasMatch = True tokenData = matchSuccess.group(1) tokenObj = tokenClass(tokenData) try: m21Obj = tokenObj.parse(self) if m21Obj is not None: # can only match one. break except TinyNotationException as excep: if self.raiseExceptions: raise TinyNotationException(f'Could not parse token: {t_orig!r}') from excep if not hasMatch and self.raiseExceptions: raise TinyNotationException(f'Could not parse token: {t_orig!r}') if m21Obj is not None: for stateObj in self.activeStates[:]: # iterate over copy so we can remove. m21Obj = stateObj.affectTokenAfterParseBeforeModifiers(m21Obj) if m21Obj is not None: for modObj in activeModifiers: m21Obj = modObj.postParse(m21Obj) if m21Obj is not None: for stateObj in self.activeStates[:]: # iterate over copy so we can remove. m21Obj = stateObj.affectTokenAfterParse(m21Obj) if m21Obj is not None: self.stream.coreAppend(m21Obj) if numberOfStatesToEnd > len(self.activeStates): if self.raiseExceptions: raise TinyNotationException(f'Token {t_orig!r} closes more states than are open.') numberOfStatesToEnd = len(self.activeStates) for i in range(numberOfStatesToEnd): stateToRemove = self.activeStates.pop() possibleObj = stateToRemove.end() if possibleObj is not None: self.stream.coreAppend(possibleObj) def parseStartStates(self, t: str) -> str: # noinspection PyShadowingNames,GrazieInspection ''' Changes the states in self.activeStates, and starts the state given the current data. Returns a newly processed token. A contrived example: >>> tnc = tinyNotation.Converter() >>> tnc.setupRegularExpressions() >>> len(tnc.activeStates) 0 >>> tIn = 'trip{quad{f8~' >>> tOut = tnc.parseStartStates(tIn) >>> tOut 'f8' >>> len(tnc.activeStates) 3 >>> tripState = tnc.activeStates[0] >>> tripState >>> quadState = tnc.activeStates[1] >>> quadState >>> tieState = tnc.activeStates[2] >>> tieState >>> tieState.parent is tnc True >>> tieState.stateInfo '~' >>> quadState.stateInfo 'quad{' Note that the affected tokens haven't yet been added: >>> tripState.affectedTokens [] Unknown state gives a warning or if `.raisesException=True` raises a TinyNotationException >>> tnc.raiseExceptions = True >>> tIn = 'blah{f8~' >>> tOut = tnc.parseStartStates(tIn) Traceback (most recent call last): music21.tinyNotation.TinyNotationException: Incorrect bracket state: 'blah' ''' bracketMatchSuccess = self.generalBracketStateRe.search(t) while bracketMatchSuccess: stateData = bracketMatchSuccess.group(0) bracketType = bracketMatchSuccess.group(1) t = self.generalBracketStateRe.sub('', t, count=1) bracketMatchSuccess = self.generalBracketStateRe.search(t) if bracketType not in self.bracketStateMapping: msg = f'Incorrect bracket state: {bracketType!r}' if self.raiseExceptions: raise TinyNotationException(msg) # else # pragma: no cover environLocal.warn(msg) continue stateObj = self.bracketStateMapping[bracketType](self, stateData) stateObj.start() self.activeStates.append(stateObj) tieMatchSuccess = self.tieStateRe.search(t) if tieMatchSuccess: stateData = tieMatchSuccess.group(0) t = self.tieStateRe.sub('', t) tieState = TieState(self, stateData) tieState.start() self.activeStates.append(tieState) return t def parseEndStates(self, t: str) -> tuple[str, int]: # noinspection GrazieInspection ''' Trims the endState token (`}`) from the t string and then returns a two-tuple of the new token and number of states to remove: >>> tnc = tinyNotation.Converter() >>> tnc.parseEndStates('C4') ('C4', 0) >>> tnc.parseEndStates('C4}}') ('C4', 2) ''' endBrackets = t.count('}') t = t.replace('}', '') return t, endBrackets def parseModifiers(self, t): ''' Parses `modifierEquals`, `modifierUnderscore`, `modifierStar`, etc. for a given token and returns the modified token and a (possibly empty) list of activeModifiers. Modifiers affect only the current token. To affect multiple tokens, use a :class:`~music21.tinyNotation.State` object. ''' activeModifiers = [] for modifierName in ('Equals', 'Star', 'Angle', 'Parens', 'Square', 'Underscore'): modifierClass = getattr(self, 'modifier' + modifierName, None) if modifierClass is None: continue modifierRe = getattr(self, '_modifier' + modifierName + 'Re', None) foundIt = modifierRe.search(t) if foundIt is not None: # is not None is necessary modifierData = foundIt.group(1) t = modifierRe.sub('', t) modifierObject = modifierClass(modifierData, t, self) activeModifiers.append(modifierObject) for modObj in activeModifiers: modObj.preParse(t) return t, activeModifiers def postParse(self): ''' Called after all the tokens have been run. It currently runs `.makeMeasures()` on `.stream` unless `.makeNotation` is `False`. ''' if self.makeNotation is not False: self.stream.makeMeasures(inPlace=True) class Test(unittest.TestCase): parseTest = '1/4 trip{C8~ C~_hello C=mine} F~ F~ 2/8 F F# quad{g--16 a## FF(n) g#} g16 F0' def testOne(self) -> None: c = Converter(self.parseTest) c.parse() s = c.stream sfn = s.flatten().getElementsByClass(note.Note) t0 = sfn[0].tie t1 = sfn[1].tie t2 = sfn[2].tie if typing.TYPE_CHECKING: assert t0 is not None assert t1 is not None assert t2 is not None self.assertIsInstance(t0, tie.Tie) self.assertIsInstance(t1, tie.Tie) self.assertIsInstance(t2, tie.Tie) self.assertEqual(t0.type, 'start') self.assertEqual(t1.type, 'continue') self.assertEqual(t2.type, 'stop') self.assertEqual(sfn[0].step, 'C') self.assertEqual(sfn[0].octave, 3) self.assertEqual(sfn[1].lyric, 'hello') self.assertEqual(sfn[2].id, 'mine') acc6 = sfn[6].pitch.accidental acc7 = sfn[7].pitch.accidental if typing.TYPE_CHECKING: assert acc6 is not None assert acc7 is not None self.assertEqual(acc6.alter, 1) self.assertEqual(acc7.alter, -2) self.assertEqual(sfn[9].editorial.ficta.alter, 0) self.assertEqual(sfn[12].duration.quarterLength, 1.0) self.assertEqual(sfn[12].expressions[0].classes, expressions.Fermata().classes) def testRaiseExceptions(self) -> None: error_states = [ { 'string': 'h', 'reason': 'h is not a valid note', }, { 'string': 'a;', 'reason': 'a semicolon is not a valid character or modifier', }, { 'string': 'r;', 'reason': 'a semicolon is not a valid character or modifier', }, { 'string': '4/4;', 'reason': 'a semicolon is not a valid character or modifier', }, { 'string': 'ABC', 'reason': ( 'only the same upper-cased letter may be repeated to ' + 'indicate lower octaves' ), }, { 'string': 'aaa', 'reason': ( 'the same lower-cased letter may not be repeated to ' + 'indicate higher octaves. Instead use apostrophes.' ), }, ] for error_state in error_states: with self.assertRaises(TinyNotationException, msg=( 'Should have raised a TinyNotationException for input ' + f"'{error_state['string']}' because {error_state['reason']}." )): converter = Converter(error_state['string'], raiseExceptions=True) converter.parse() def testGetDefaultTokenMap(self) -> None: defaultTokenMap = _getDefaultTokenMap() self.assertEqual( len(defaultTokenMap), 3, ( 'There should be three valid token types by default: Time ' + 'signatures, Notes, and Rests' ) ) validTokenTypeCounts = { NoteToken: 0, RestToken: 0, TimeSignatureToken: 0, } for regex, tokenType in defaultTokenMap: self.assertIn( tokenType, validTokenTypeCounts, ( 'Found unexpected token type in default token map:' + f'{tokenType.__class__.__name__}.' ) ) # noinspection PyTypeChecker validTokenTypeCounts[tokenType] += 1 self.assertGreater( len(regex), 0, ( 'Should provide a non-empty string for the regular ' + 'expression in the default token map for tokens of type ' + f'{tokenType.__class__.__name__}.' ) ) for tokenType in validTokenTypeCounts: self.assertEqual( validTokenTypeCounts[tokenType], 1, ( 'Should have found each valid token type exactly once in ' + 'the default token map.' ) ) def test_too_many_states(self): c = Converter('2/4 trip{c8 d e}} f4', makeNotation=False) c.parse() s = c.stream self.assertEqual(s.notes[-2].duration.quarterLength, fractions.Fraction(1, 3)) self.assertEqual(s.notes.last().duration.quarterLength, 1.0) c = Converter('2/4 trip{c8 d e}} f4', makeNotation=False) c.raiseExceptions = True with self.assertRaisesRegex( TinyNotationException, "Token 'e}}' closes more states than are open" ): c.parse() class TestExternal(unittest.TestCase): show = True def testOne(self): c = Converter(Test.parseTest) c.parse() if self.show: c.stream.show('musicxml.png') # TODO: Chords # ------------------------------------------------------------------------------ # define presented order in documentation _DOC_ORDER = [Converter, Token, State, Modifier] if __name__ == '__main__': import music21 music21.mainTest(Test)