commit 1af10c4da687379d5ea75882afeac0bde963c081
parent 322ee37350f05fcd7edda86a31ad555ddd64dbfd
Author: Antoine Amarilli <a3nm@a3nm.net>
Date: Thu, 27 Aug 2015 02:23:15 +0200
continue
Diffstat:
10 files changed, 175 insertions(+), 56 deletions(-)
diff --git a/.gitignore b/.gitignore
@@ -6,3 +6,10 @@ lexique.txt
lexique3
lexique_full
raw_graph
+test.raw
+test.ogg
+song.xml
+3chatsrep.txt
+lyrics
+song
+output
diff --git a/3chats.mid b/3chats.mid
Binary files differ.
diff --git a/3chats.txt b/3chats.txt
@@ -0,0 +1,24 @@
+0.625 C6
+0.625 D6
+1.25 Ds6
+0.625 D6
+0.625 Ds6
+1.25 C6
+0.625 Ds6
+0.625 F6
+1.25 G6
+1.25 G6
+1.25 G6
+0.625 G6
+0.625 Gs6
+0.625 G6
+0.625 F6
+0.625 F6
+0.625 G6
+0.625 F6
+0.625 Ds6
+0.625 Ds6
+0.625 F6
+1.25 Ds6
+1.25 D6
+1.25 C6
diff --git a/arrange.py b/arrange.py
@@ -0,0 +1,55 @@
+#!/usr/bin/python
+
+import sys
+
+notes = open(sys.argv[1], "r")
+bpm = sys.argv[2]
+
+def transp(n):
+ if n[-1] == "4":
+ n = n[:-1] + "3"
+ if n[-1] == "5":
+ n = n[:-1] + "4"
+ if n[-1] == "6":
+ n = n[:-1] + "5"
+ if n[-1] == "7":
+ n = n[:-1] + "6"
+ return n
+
+print """<?xml version="1.0"?>
+<!DOCTYPE SINGING PUBLIC "-//SINGING//DTD SINGING mark up//EN"
+ "Singing.v0_1.dtd"
+[]>"""
+print '<SINGING BPM="%s">' % bpm
+
+while True:
+ note = notes.readline()
+ note = note.strip().split(" ")
+ try:
+ if note[1][0] == "S":
+ print ("<REST BEATS=\"%s\"></REST>" % note[0])
+ continue
+ except Exception:
+ break
+ l = sys.stdin.readline()
+ if not l:
+ break
+ l = l.strip()
+ nh = len(l.split("-"))
+ l = "".join(l.split("-"))
+ n1 = [note[0]]
+ n2 = [transp(note[1])]
+ for i in range(nh-1):
+ note = notes.readline()
+ note = note.strip().split(" ")
+ n1.append(note[0])
+ n2.append(transp(note[1]))
+ notes1 = ",".join(n1)
+ notes2 = ",".join(n2)
+ print ("<DURATION BEATS=\"%s\"><PITCH NOTE=\"%s\">"
+ "%s</PITCH></DURATION>" % (notes1, notes2, l))
+
+
+notes.close()
+
+print "</SINGING>"
diff --git a/graph2.py b/graph2.py
@@ -4,7 +4,7 @@ import sys
from common import vowels, semivowels, fem, start
from collections import defaultdict
-g = defaultdict(lambda: defaultdict(lambda: set()))
+g = defaultdict(lambda: defaultdict(lambda: []))
def explore(graph, fr):
seen = set()
@@ -41,21 +41,25 @@ for l in sys.stdin.readlines():
# we are between first vowel included and second vowel excluded
fr = (not fems, allp[:i+1])
to = (fems, p[-1])
+ if allp.startswith(to[1]):
+ # loop (e.g., trinitrine), not very pretty
+ continue
#if (w.startswith('mandib')):
#print(w, fr, to)
- todel = set()
- toadd = True
- for (ww, syl) in g[fr][to]:
- if ww.startswith(w) and w != ww:
- # we will kill something longer
- todel.add(ww)
- elif w.startswith(ww):
- # we already have something shorter
- toadd = False
- break
- g[fr][to] -= todel
- if toadd:
- g[fr][to].add((w, syl))
+ g[fr][to].append((w, syl))
+# todel = set()
+# toadd = True
+# for (ww, syl) in g[fr][to]:
+# if ww.startswith(w) and w != ww:
+# # we will kill something longer
+# todel.add(ww)
+# elif w.startswith(ww):
+# # we already have something shorter
+# toadd = False
+# break
+# g[fr][to] -= todel
+# if toadd:
+# g[fr][to].add((w, syl))
# compute reverse graph
rg = defaultdict(lambda: defaultdict(lambda: set()))
@@ -75,7 +79,8 @@ for f in g.keys():
#print ("HOHOHO", t, start, f in useful, t in useful)
if f not in useful or t not in useful:
continue
- for w in g[f][t]:
+ # only print the most frequent word, for now
+ for w in g[f][t][:1]:
print ("%s\t%s\t%s\t%s\t%s\t%s" % ('f' if f[0] else 'm', f[1], 'f' if t[0] else 'm', t[1], w[0], w[1]))
diff --git a/infersplit.py b/infersplit.py
@@ -0,0 +1,30 @@
+# OLD CODE
+
+def print_word(w, prn):
+ s = ""
+ vowels = False
+ for i in range(len(w)):
+ p = -(i+1)
+ s += w[p]
+ x = strip_accents(w[p])[0]
+ if not vowels:
+ if x in vowels_script and (i > 0 or w[-1] != 'e') and (i > 1 or w[-1] !=
+ 's' or w[-2] != 'e') and (i > 2 or (not w.endswith('gue'))
+ and
+ not (w.endswith('que'))) and (i > 3 or (not
+ w.endswith('gues') and (not w.endswith('ques')))):
+ vowels = True
+ else:
+ if x not in vowels_script and x != '-':
+ if prn[0] in semivowels:
+ s = s[:-1]
+ break
+ # exception
+ if w[p] == 'h':
+ s += w[-(i+2)]
+ break
+ s = s[::-1]
+ print ("## %s, %s, %s, %s, %s" % (w, w, w, s, s))
+
+print_word(w[0], l[i+1][1])
+
diff --git a/print.sh b/print.sh
@@ -0,0 +1,15 @@
+#!/bin/bash
+
+./printcycle.py output 0 < graph > song
+./printcycle.py output 1 < graph | tr -d ',' | sed '$d' > lyrics
+
+N=$(wc -l < lyrics)
+echo > 3chatsrep.txt
+for i in `seq $(($N/2))`
+do
+ cat 3chats.txt >> 3chatsrep.txt
+done
+cat lyrics | tr ' ' '\n' | grep -v '^ *$' |
+ ./arrange.py <(grep -v '^$' 3chatsrep.txt) 120 |
+ sed 's/Fs/Gb/g;s/As\([1-9]\)/Bb\1/g;s/Cs/Db/g;s/Ds/Eb/g' |
+ sed 's/"Gs/"Ab/' > song.xml
diff --git a/printcycle.py b/printcycle.py
@@ -8,32 +8,6 @@ g = defaultdict(lambda: {})
visited = set()
-def print_word(w, prn):
- s = ""
- vowels = False
- for i in range(len(w)):
- p = -(i+1)
- s += w[p]
- x = strip_accents(w[p])[0]
- if not vowels:
- if x in vowels_script and (i > 0 or w[-1] != 'e') and (i > 1 or w[-1] !=
- 's' or w[-2] != 'e') and (i > 2 or (not w.endswith('gue'))
- and
- not (w.endswith('que'))) and (i > 3 or (not
- w.endswith('gues') and (not w.endswith('ques')))):
- vowels = True
- else:
- if x not in vowels_script and x != '-':
- if prn[0] in semivowels:
- s = s[:-1]
- break
- # exception
- if w[p] == 'h':
- s += w[-(i+2)]
- break
- s = s[::-1]
- print ("## %s, %s, %s, %s, %s" % (w, w, w, s, s))
-
def print_list(l):
global out
if not start[0]:
@@ -43,23 +17,23 @@ def print_list(l):
w = g[l[i]][l[i+1]]
#print(w[0])
fs = w[1].split('-')
- if len(w[1]) == 0 or len(fs) < 3:
- print_word(w[0], l[i+1][1])
+ # remove spaces in syll for singing
+ for i in range(len(fs)):
+ fs[i] = ''.join(fs[i].split(' '))
+ end = ''.join(fs[2:])
+ ssep = ' ' if songmode else ''
+ spl = ssep.join(fs)
+ if w[0]:
+ spl2 = fs[0] + ssep + fs[1] + ssep + ''.join(fs[2:])
else:
- # remove spaces in syll for singing
- for i in range(len(fs)):
- fs[i] = ''.join(fs[i].split(' '))
- end = ''.join(fs[2:])
- spl = ' '.join(fs)
- if w[0]:
- spl2 = fs[0] + ' ' + fs[1] + ' ' + ''.join(fs[2:])
- else:
- spl2 = spl
- print (" %s, %s, %s, %s, %s"
- % (spl, spl, spl2, end, end))
+ spl2 = spl
+ sep = '' if songmode else ','
+ print ("%s%s %s%s %s%s %s%s %s"
+ % (spl, sep, spl, sep, spl2, sep, end, sep, end))
fname = sys.argv[1]
f = open(fname)
+songmode = sys.argv[2] == '1'
for l in sys.stdin.readlines():
l = l.strip().split('\t')
diff --git a/script.sh b/script.sh
@@ -5,10 +5,13 @@
#cat Lexique371/Bases+Scripts/Lexique3.txt |
# ./lexique_fix.sh| cut -f1,2,4 | grep NOM | uniq > lexique.txt
-cat ~/documents/lexique/lexique | cut -f1,2,4,23,24,28 | grep NOM | grep '\s3\s' | cut -f1,4,6 | rev | uniq -f 2 | rev > lexique_full
+cat - | sort -k 14,14n -k8,8nr |
+ grep -vi "zanimaux" |
+ cut -f1,2,4,23,24,28 | grep NOM |
+ grep '\s3\s' | cut -f1,4,6 | awk '!a[$1]++' > lexique_full
pv lexique_full | ./only3.py > lexique3
cat lexique3 | ./graph2.py > graph
-cat graph | ./cycle.py
+cat graph | ./cycle.py output
# cat <(echo "digraph G {") <(sed 's/\([^ ]*\) \([^ ]*\) \(.*\)/\1 -> \2
# [label="\3"]/g' graph) <(echo "}") | sed 's/#/_/g;s/8/W/g;s/\$/Y/g;s/)/X/g' >
diff --git a/sing.sh b/sing.sh
@@ -0,0 +1,6 @@
+#!/bin/bash
+
+echo "(tts \"song.xml\" 'singing)" | festival
+# while playing, do cp /tmp/audio* test.raw
+# then convert with audacity
+
