X-Git-Url: http://iramuteq.org/git?p=iramuteq;a=blobdiff_plain;f=corpus.py;h=080b980ff7bbe13c4ff0d665db751a2af92fbfd7;hp=0d661d6d7ad2c6a8e6185dae4300778a95b3b4fb;hb=80f4bfad30ece8835cb1f91349b1dda36439e4ca;hpb=065a856b87fe4fe3258b1406d6ace509e9da6e69
diff --git a/corpus.py b/corpus.py
old mode 100644
new mode 100755
index 0d661d6..080b980
--- a/corpus.py
+++ b/corpus.py
@@ -8,7 +8,7 @@ _ = gettext.gettext
import locale
import sys
from time import time
-from functions import decoupercharact, ReadDicoAsDico, DoConf, ReadLexique
+from functions import decoupercharact, ReadDicoAsDico, DoConf, ReadLexique, progressbar
import re
import sqlite3
import itertools
@@ -16,7 +16,7 @@ import logging
from operator import itemgetter
from uuid import uuid4
from chemins import PathOut
-from dialog import CorpusPref, SubTextFromMetaDial
+from dialog import CorpusPref, SubTextFromMetaDial, MergeClusterFrame
from copy import copy
from colors import colors
import datetime
@@ -36,14 +36,15 @@ def copycorpus(corpus) :
def CopyUce(uce) :
return Uce(uce.ident, uce.para, uce.uci)
-
+
def CopyUci(uci):
nuci = Uci(uci.ident, '')
nuci.etoiles = copy(uci.etoiles)
nuci.uces = [CopyUce(uce) for uce in uci.uces]
+ nuci.paras = copy(uci.paras)
return nuci
-
+
class Corpus :
@@ -53,7 +54,7 @@ class Corpus :
def __init__(self, parent, parametres = {}, read = False) :
self.parent = parent
self.parametres = parametres
- self.cformes = None
+ self.cformes = None
self.connformes = None
self.connuces = None
self.conncorpus = None
@@ -93,7 +94,7 @@ class Corpus :
lem = word
self.formes[word] = Word(word, gramtype, len(self.formes), lem)
self.idformesuces[self.formes[word].ident] = {self.ucis[-1].uces[-1].ident : 1}
-
+
def add_word_from_forme(self, word, stident):
if word.forme in self.formes :
self.formes[word.forme].freq += 1
@@ -106,7 +107,7 @@ class Corpus :
self.idformesuces[self.formes[word.forme].ident] = {stident: 1}
else :
self.formes[word.forme] = Word(word.forme, word.gram, len(self.formes), word.lem)
- self.idformesuces[self.formes[word.forme].ident] = {stident : 1}
+ self.idformesuces[self.formes[word.forme].ident] = {stident : 1}
def conn_all(self):
"""connect corpus to db"""
@@ -142,13 +143,13 @@ class Corpus :
res = self.ccorpus.execute('SELECT * FROM formes;')
self.formes = dict([[forme[1], Word(forme[1], forme[3], forme[0], lem = forme[2], freq = forme[4])] for forme in res])
self.ccorpus.close()
-
+
def getworduces(self, wordid) :
if isinstance(wordid, basestring) :
wordid = self.formes[wordid].ident
res = self.cformes.execute('SELECT uces FROM uces where id=? ORDER BY id;', (`wordid`,))
return list(itertools.chain(*[[int(val) for val in row[0].split()] if not isinstance(row[0], int) else [row[0]] for row in res]))
-
+
def getworducis(self, wordid) :
res = self.getworduces(wordid)
return list(set([self.getucefromid(uce).uci for uce in res]))
@@ -164,7 +165,7 @@ class Corpus :
formeuceeff = {}
for i, uce in enumerate(uces) :
formeuceeff[uce] = formeuceeff.get(uce, 0) + eff[i]
- return formeuceeff
+ return formeuceeff
def getlemuces(self, lem) :
formesid = ', '.join([`val` for val in self.lems[lem].formes])
@@ -172,6 +173,34 @@ class Corpus :
res = self.cformes.execute(query)
return list(set(list(itertools.chain(*[[int(val) for val in row[0].split()] if not isinstance(row[0], int) else [row[0]] for row in res]))))
+ def gettgenst(self, tgen):
+ formesid = []
+ for lem in tgen :
+ if lem in self.lems :
+ formesid += self.lems[lem].formes
+ else :
+ print 'abscent : %s' % lem
+ query = 'SELECT uces FROM uces where id IN %s ORDER BY id' % str(tuple(formesid))
+ res = self.cformes.execute(query)
+ return list(set(list(itertools.chain(*[[int(val) for val in row[0].split()] if not isinstance(row[0], int) else [row[0]] for row in res]))))
+
+ def gettgenstprof(self, tgen, classe, i, clnb):
+ tgenst = []
+ for lem in tgen :
+ if lem in self.lems :
+ lemst = self.getlemuces(lem)
+ tgenst += lemst
+ if not lem in self.tgenlem :
+ self.tgenlem[lem] = [0] * clnb
+ self.tgenlem[lem][i] = len(set(lemst).intersection(classe))
+ else :
+ print 'abscent: ',lem
+ return list(set(tgenst))
+
+ def gettgentxt(self, tgen):
+ sts = self.gettgenst(tgen)
+ return list(set([self.getucefromid(val).uci for val in sts]))
+
def getlemucis(self, lem) :
uces = self.getlemuces(lem)
return list(set([self.getucefromid(val).uci for val in uces]))
@@ -217,19 +246,24 @@ class Corpus :
def getucisize(self) :
ucesize = self.getucesize()
return [sum(ucesize[uci.uces[0].ident:(uci.uces[-1].ident + 1)]) for uci in self.ucis]
-
+
def getucesize(self) :
res = self.getalluces()
return [len(uce[1].split()) for uce in res]
def getconcorde(self, uces) :
return self.cuces.execute('select * from uces where id IN (%s) ORDER BY id;' % ', '.join([`i` for i in uces]))
-
+
def getuciconcorde(self, ucis) :
uces = [[val,[uce.ident for uce in self.ucis[val].uces]] for val in ucis]
uces = [[val[0], '\n'.join([row[1] for row in self.getconcorde(val[1])])] for val in uces]
return uces
+ def getuciconcorde_uces(self, uciid, uceid) :
+ uces = [uce.ident for uce in self.ucis[uciid].uces]
+ uces = [row for row in self.getconcorde(uces)]
+ return uces
+
def getwordconcorde(self, word) :
return self.getconcorde(self.getworduces(word))
@@ -238,11 +272,11 @@ class Corpus :
def getalluces(self) :
return self.cuces.execute('SELECT * FROM uces')
-
+
def getallucis(self):
uces = [row[1] for row in self.getalluces()]
return [[uci.ident, '\n'.join([uces[uce.ident] for uce in uci.uces])] for uci in self.ucis]
-
+
def getucesfrometoile(self, etoile) :
return [uce.ident for uci in self.ucis for uce in uci.uces if etoile in uci.etoiles]
@@ -267,7 +301,7 @@ class Corpus :
else :
idpara += 1
return etoileuces
-
+
def getetoileucis(self):
etoileuces = {}
for uci in self.ucis :
@@ -314,7 +348,7 @@ class Corpus :
self.lems[self.formes[forme].lem] = Lem(self, self.formes[forme])
else :
self.lems = dict([[forme, Lem(self, self.formes[forme])] for forme in self.formes])
-
+
def make_lems_from_dict(self, dictionnaire, dolem = True) :
log.info('make lems from dict')
self.lems = {}
@@ -338,7 +372,7 @@ class Corpus :
self.lems[self.formes[forme].lem] = Lem(self, self.formes[forme])
else :
self.lems[forme] = Lem(self, self.formes[forme])
-
+
def make_idformes(self) :
self.idformes = dict([[self.formes[forme].ident, self.formes[forme]] for forme in self.formes])
@@ -347,6 +381,7 @@ class Corpus :
self.iduces = dict([[uce.ident, uce] for uci in self.ucis for uce in uci.uces])
def make_lexitable(self, mineff, etoiles, gram = 0) :
+ log.info('making lexical table...')
if gram == 0 :
grams = {1:'', 2:''}
else :
@@ -364,10 +399,12 @@ class Corpus :
for lem in tokeep :
deff = self.getlemuceseff(lem)
ucesk = deff.keys()
- tab.append([lem] + [sum([deff[uce] for uce in et.intersection(ucesk)]) for et in etuces])
+ line = [lem] + [sum([deff[uce] for uce in et.intersection(ucesk)]) for et in etuces]
+ if sum(line[1:]) >= mineff :
+ tab.append(line)
tab.insert(0, [''] + etoiles)
return tab
-
+
def make_tgen_table(self, tgen, etoiles, tot = None):
lclasses = [self.getucesfrometoile(etoile) for etoile in etoiles]
sets = [set(cl) for cl in lclasses]
@@ -391,6 +428,29 @@ class Corpus :
tgenoccurrences[t][etoiles[i]] += sum([lemuceeff[uce] for uce in concern])
return tgenoccurrences, totoccurrences
+ def make_tgen_profile(self, tgen, ucecl, uci = False) :
+ log.info('tgen/classes')
+ self.tgenlem = {}
+ clnb = len(ucecl)
+ if uci :
+ #FIXME : NE MARCHE PLUS CHANGER CA
+ tab = [[lem] + [len(set(self.gettgentxt(tgen[lem])).intersection(classe)) for i, classe in enumerate(ucecl)] for lem in tgen]
+ else :
+ tab = [[lem] + [len(set(self.gettgenstprof(tgen[lem], classe, i, clnb)).intersection(classe)) for i, classe in enumerate(ucecl)] for lem in tgen]
+ tab = [[line[0]] + [val for val in line[1:]] for line in tab if sum(line[1:]) >= 3]
+ return tab
+ #i = 0
+ #nam = 'total'
+ #while nam + `i` in tgen :
+ # i += 1
+ #nam = nam + `i`
+ #last = [nam] + [`len(classe)` for classe in ucecl]
+ #tab += [last]
+ #line0 = ['tgen'] + ['_'.join(['cluster', `i+1`]) for i in range(len(ucecl))]
+ #tab = [line0] + tab
+ #with open(fileout, 'w') as f :
+ # f.write('\n'.join(['\t'.join(line) for line in tab]).encode(self.parametres['syscoding']))
+
def make_efftype_from_etoiles(self, etoiles) :
dtype = {}
etuces = [[] for et in etoiles]
@@ -509,6 +569,14 @@ class Corpus :
ident += 1
f.write('\n'.join([self.ucis[self.iduces[uce[0]].uci].paras[ident].encode(self.parametres['syscoding']), uce[1].encode(self.parametres['syscoding'])]) + '\n')
+ def export_meta_table(self, outf) :
+ metas = [[`i`] + text.etoiles[1:] for i, text in enumerate(self.ucis)]
+ longueur_max = max([len(val) for val in metas])
+ first = ['column_%i' % i for i in range(longueur_max)]
+ metas.insert(0, first)
+ with open(outf, 'w') as f :
+ f.write('\n'.join(['\t'.join(line) for line in metas]).encode(self.parametres['syscoding']))
+
def export_corpus_classes(self, outf, alc = True, lem = False, uci = False) :
ucecl = {}
for i, lc in enumerate(self.lc) :
@@ -538,7 +606,7 @@ class Corpus :
f.write(guce.encode(self.parametres['syscoding']) + '\n\n')
def export_classe(self, outf, classe, lem = False, uci = False) :
- sts = self.lc[classe - 1]
+ sts = self.lc[classe - 1]
if not uci :
res = self.getconcorde(sts)
self.make_iduces()
@@ -596,7 +664,7 @@ class Corpus :
nbl += 1
f.write(''.join([' '.join([`uce+1`, `i+1`,`1`]),'\n']))
f.seek(0)
- with open(outfile, 'w') as ffin :
+ with open(outfile, 'w') as ffin :
ffin.write("%%%%MatrixMarket matrix coordinate integer general\n%i %i %i\n" % (self.getucenb(), len(actives), nbl))
for line in f :
ffin.write(line)
@@ -614,7 +682,7 @@ class Corpus :
nbl += 1
f.write(''.join([' '.join([`uci+1`, `i+1`,`1`]),'\n']))
f.seek(0)
- with open(outfile, 'w') as ffin :
+ with open(outfile, 'w') as ffin :
ffin.write("%%%%MatrixMarket matrix coordinate integer general\n%i %i %i\n" % (self.getucinb(), len(actives), nbl))
for line in f :
ffin.write(line)
@@ -622,7 +690,7 @@ class Corpus :
if listuci :
with open(listuci, 'w') as f :
f.write('\n'.join(['uci;uc'] + [';'.join([`i`,`i`]) for i in range(0, self.getucinb())]))
-
+
def make_and_write_sparse_matrix_from_classe(self, actives, uces, outfile) :
log.info('make_and_write_sparse_matrix_from_classe %s' % outfile)
nbl = 0
@@ -633,12 +701,12 @@ class Corpus :
for uce in uces_ok :
f.write(''.join([' '.join([`duces[uce]+1`,`i+1`,`1`]),'\n']))
f.seek(0)
- with open(outfile, 'w') as ffin :
- ffin.write("%%%%MatrixMarket matrix coordinate integer general\n%i %i %i\n" % (self.getucenb(), len(actives), nbl))
+ with open(outfile, 'w') as ffin :
+ ffin.write("%%%%MatrixMarket matrix coordinate integer general\n%i %i %i\n" % (len(uces), len(actives), nbl))
for line in f :
ffin.write(line)
os.remove(outfile + '~')
-
+
def make_table_with_classe(self, uces, list_act, uci = False) :
table_uce = [[0 for val in list_act] for line in range(0,len(uces))]
uces = dict([[uce, i] for i, uce in enumerate(uces)])
@@ -651,8 +719,8 @@ class Corpus :
for uce in lemuces :
table_uce[uces[uce]][i] = 1
table_uce.insert(0, list_act)
- return table_uce
-
+ return table_uce
+
def make_pondtable_with_classe(self, uces, list_act) :
table_uce = [[0 for val in list_act] for line in range(0,len(uces))]
uces = dict([[uce, i] for i, uce in enumerate(uces)])
@@ -662,7 +730,7 @@ class Corpus :
for uce in lemuces :
table_uce[uces[uce]][i] = uceseff[uce]
table_uce.insert(0, list_act)
- return table_uce
+ return table_uce
def parse_active(self, gramact, gramsup = None) :
log.info('parse actives')
@@ -683,7 +751,7 @@ class Corpus :
if self.idformes is None :
self.make_idformes()
return [lem for lem in self.lems if self.getlemeff(lem) >= limit and self.lems[lem].act == key]
-
+
def make_actives_nb(self, nbmax, key) :
log.info('make_actives_nb : %i - %i' % (nbmax,key))
if self.idformes is None :
@@ -711,7 +779,7 @@ class Corpus :
stop = nbmax - 1
lim = effs[stop]
log.info('nb actives = %i - eff min = %i ' % (stop + 1, lim))
- return [val[1] for val in allactives[0:stop + 1]], lim
+ return [val[1] for val in allactives[0:stop]], lim
def make_and_write_profile(self, actives, ucecl, fileout, uci = False) :
log.info('formes/classes')
@@ -722,6 +790,7 @@ class Corpus :
tab = [[line[0]] + [`val` for val in line[1:]] for line in tab if sum(line[1:]) >= 3]
with open(fileout, 'w') as f :
f.write('\n'.join([';'.join(line) for line in tab]).encode(self.parametres['syscoding']))
+ f.write('\n')
def make_etoiles(self) :
etoiles = set([])
@@ -756,6 +825,28 @@ class Corpus :
except IndexError :
det[et[0]] = 1
return det
+
+ def make_theme_dict(self):
+ themes = [val for uci in self.ucis for val in uci.paras]
+ det = {}
+ for theme in themes :
+ th = theme.split('_')
+ if th[0] in det :
+ try :
+ endth = '_'.join(th[1:])
+ if theme in det[th[0]] :
+ det[th[0]][theme] += 1
+ else :
+ det[th[0]][theme] = 1
+ except IndexError :
+ det[th[0]] += 1
+ else :
+ try :
+ endth = '_'.join(th[1:])
+ det[th[0]] = {theme:1}
+ except IndexError :
+ det[th[0]] = 1
+ return det
def make_etline(self, listet) :
etuces = [[] for et in listet]
@@ -773,8 +864,13 @@ class Corpus :
etoileuces = self.getetoileuces()
else :
etoileuces = self.getetoileucis()
- etoileuces = dict([[et, etoileuces[et]] for et in etoileuces if len(etoileuces[et]) > 1])
+ print 'etoilesuces ok'
+ etoileuces = dict([[et, etoileuces[et]] for et in etoileuces if
+ len(etoileuces[et]) > 1 ]) #and not et.startswith(u'*reference_')
+ print len(etoileuces)
+ print 'etoilesuces ok2'
with open(fileout, 'w') as f :
+ print 'write...'
f.write('\n'.join([';'.join([et] + [`len(set(etoileuces[et]).intersection(classe))` for classe in ucecl]) for et in etoileuces]).encode(self.parametres['syscoding']))
#etoiles = self.make_etoiles()
#with open(fileout, 'w') as f :
@@ -783,11 +879,11 @@ class Corpus :
def make_colored_corpus(self, uci = False) :
ucecl = {}
for i, lc in enumerate(self.lc) :
- for uce in lc :
+ for uce in lc :
ucecl[uce] = i + 1
for uce in self.lc0 :
ucecl[uce] = 0
- color = ['black'] + colors[len(self.lc) - 1]
+ color = ['black'] + colors[len(self.lc) - 1]
txt = '''
@@ -816,6 +912,32 @@ class Corpus :
txt += '' % (color[ucecl[uce[0]]]) + uce[1] + '
'
return txt + '\n'
+ def make_cut_corpus(self, uci = False) :
+ txt = u''
+ if not uci :
+ res = self.getalluces()
+ self.make_iduces()
+ actuci = ''
+ actpara = False
+ for uce in res :
+ if self.iduces[uce[0]].uci != actuci :
+ actuci = self.iduces[uce[0]].uci
+ txt += u'\n' + ' '.join(self.ucis[self.iduces[uce[0]].uci].etoiles) + u'\n'
+ txt += ''.join([u'\n',uce[1],u'\n'])
+ else :
+ txt += ''.join([u'\n',uce[1],u'\n'])
+ else :
+ res = self.getallucis()
+ actuci = ''
+ for uce in res :
+ if self.ucis[uce[0]].ident != actuci :
+ actuci = self.ucis[uce[0]].ident
+ txt += u'\n' + ' '.join(self.ucis[self.ucis[uce[0]].ident].etoiles) + u'\n'
+ txt += ''.join([u'\n',uce[1],u'\n'])
+ else :
+ txt += ''.join([u'\n',uce[1],u'\n'])
+ return txt
+
def count_from_list(self, l, d) :
for val in l :
if val in d :
@@ -847,7 +969,7 @@ class Corpus :
def find_segments_in_classe(self, list_uce, taille_segment, taille_limite, uci = False):
d={}
- if not uci :
+ if not uci :
concorde = self.getconcorde
else :
concorde = self.getuciconcorde
@@ -860,7 +982,7 @@ class Corpus :
if len(l) > taille_limite :
l = l[-taille_limite:]
return l
-
+
def make_segments_profile(self, fileout, lenmin = 3, lenmax = 10, effmin = 50, lem = False) :
d = {}
for b, classe in enumerate(self.lc) :
@@ -873,7 +995,7 @@ class Corpus :
result = [[seg] + [str(val) for val in d[seg]] for seg in d if sum(d[seg]) >= effmin]
with open(fileout, 'w') as f :
f.write('\n'.join([';'.join(line) for line in result]))
-
+
def make_proftype(self, outf) :
res = {}
for lem in self.lems :
@@ -904,7 +1026,7 @@ class Corpus :
self.lc = [[uce[0] for uce in self.lc if uce[1] == i] for i in range(clnb+1)]
self.lc0 = self.lc.pop(0)
#return ucecl
-
+
def get_stat_by_cluster(self, outf, lclasses = None) :
log.info('get_stat_by_cluster')
if lclasses is None :
@@ -924,7 +1046,7 @@ class Corpus :
formescl[i+1] += 1
if self.formes[forme].freq == 1 :
hapaxcl[i+1] += 1
- log.info('%f' % (time() - t1))
+ log.info('%f' % (time() - t1))
if outf is not None :
toprint = '\n'.join([';'.join([`i`, `occurrences[i]`, `formescl[i]`, `hapaxcl[i]`, `lenclasses[i]`, `float(hapaxcl[i])/float(formescl[i])`]) for i in occurrences])
with open(outf, 'w') as f :
@@ -1007,7 +1129,7 @@ class Corpus :
listlem.sort()
with open(fileout, 'w') as f :
f.write('\n'.join(['\t'.join(lem) for lem in listlem]).encode(syscoding))
-
+
class MakeUciStat :
@@ -1015,7 +1137,7 @@ class MakeUciStat :
ucinb = corpus.getucinb()
ucisize = corpus.getucisize()
ucimean = float(sum(ucisize))/float(ucinb)
- detoile = corpus.make_etoiles_dict()
+ detoile = corpus.make_etoiles_dict()
class Uci :
def __init__(self, iduci, line, paraset = None) :
@@ -1066,12 +1188,12 @@ def decouperlist(chaine, longueur, longueurOptimale) :
dsep = dict([[val[0],val[1]] for val in separateurs])
trouve = False # si on a trouvé un bon séparateur
iDecoupe = 0 # indice du caractere ou il faut decouper
-
+
longueur = min(longueur, len(chaine) - 1)
chaineTravail = chaine[:longueur + 1]
nbCar = longueur
meilleur = ['', 0, 0] # type, poids et position du meilleur separateur
-
+
try :
indice = chaineTravail.index(u'$')
trouve = True
@@ -1165,14 +1287,14 @@ class BuildCorpus :
self.ucesize = self.corpus.parametres.get('ucesize', 240)
log.info('method uce : %s' % method)
- def dobuild(self) :
+ def dobuild(self) :
t1 = time()
try :
self.read_corpus(self.infile)
except Warning, args :
log.info('pas kool %s' % args)
raise Warning
- else :
+ else :
self.indexdb()
self.corpus.parametres['ira'] = self.corpus.pathout['Corpus.cira']
self.time = time() - t1
@@ -1243,11 +1365,13 @@ class BuildCorpus :
self.cleans.append(self.dotiret)
def make_expression(self,txt) :
- for expression in self.expressions:
+ exp = self.expressions.keys()
+ exp.sort(reverse=True)
+ for expression in exp :
if expression in txt :
txt = txt.replace(expression, self.expressions[expression][0])
return txt
-
+
def dolower(self, txt) :
return txt.lower()
@@ -1255,7 +1379,7 @@ class BuildCorpus :
#rule = u"^a-zA-Z0-9à ÃâÃäÃáÃéÃèÃêÃëÃìÃîÃïÃòÃôÃöÃùÃûÃüÃçÃÃÅÅâñ.:,;!?*'_-"
list_keep = u"[" + self.rule + "]+"
return re.sub(list_keep, ' ', txt)
-
+
def doapos(self, txt) :
return txt.replace(u'\'', u' ')
@@ -1279,7 +1403,7 @@ class BuildCorpus :
toeff = [(`forme`, ' '.join([`val` for val in self.corpus.idformesuces[forme].values()])) for forme in self.corpus.idformesuces]
self.cf.executemany('INSERT INTO uces VALUES (?,?);', touce)
self.cf.executemany('INSERT INTO eff VALUES (?,?);', toeff)
- self.corpus.idformesuces = {}
+ self.corpus.idformesuces = {}
self.count = 1
def backup_corpus(self) :
@@ -1287,7 +1411,7 @@ class BuildCorpus :
t = time()
for uci in self.corpus.ucis :
self.ccorpus.execute('INSERT INTO etoiles VALUES (?,?,?);' ,(uci.ident,' '.join(uci.etoiles), ' '.join(uci.paras,)))
- for uce in uci.uces :
+ for uce in uci.uces :
self.ccorpus.execute('INSERT INTO luces VALUES (?,?,?);',(`uci.ident`,`uce.para`,`uce.ident`,))
for forme in self.corpus.formes :
self.ccorpus.execute('INSERT INTO formes VALUES (?,?,?,?,?);', (`self.corpus.formes[forme].ident`, forme, self.corpus.formes[forme].lem, self.corpus.formes[forme].gram, `self.corpus.formes[forme].freq`,))
@@ -1315,6 +1439,7 @@ class BuildSubCorpus(BuildCorpus):
self.infile = None
self.corpus = Corpus(self, {'type' : 'corpus', 'originalpath' : corpus.parametres['originalpath'], 'encoding' : corpus.parametres['encoding']})
self.last = 0
+ self.parametres = parametres
self.encoding = corpus.parametres['encoding']
self.corpus.parametres['corpus_name'] = parametres['corpus_name']
self.corpus.pathout = PathOut(filename = corpus.parametres['originalpath'], dirout = parametres['pathout'])
@@ -1343,40 +1468,43 @@ class BuildSubCorpus(BuildCorpus):
nuci.paras = newpara
self.corpus.ucis.append(nuci)
else :
- idpara += 1
- elif parametres.get('fromcluster', False) :
- pass
- elif parametres.get('fromuceids', False) :
- print 'fromuceids'
- dictucekeep = dict(zip(parametres['uceids'], parametres['uceids']))
- idpara = 0
- for uci in self.ori.ucis :
- if uci.paras == [] :
- keepuces = [CopyUce(uce) for uce in uci.uces if uce.ident in dictucekeep]
- if keepuces != [] :
- nuci = CopyUci(uci)
- nuci.uces = keepuces
- self.corpus.ucis.append(nuci)
idpara += 1
- else :
- newuces = []
- newpara = []
- for et in uci.paras :
- keepuces = [CopyUce(uce) for uce in uci.uces if uce.para == idpara]
- idpara += 1
- if keepuces != [] :
- newuces += keepuces
- newpara.append(et)
- if newuces != [] :
- nuci = CopyUci(uci)
- nuci.uces = newuces
- nuci.paras = newpara
- self.corpus.ucis.append(nuci)
-
+ elif parametres.get('fromclusters', False) :
+ self.parametres['uceids'] = [st for i in self.parametres['meta'] for st in self.parametres['lc'][i]]
+ self.fromuceids()
+ elif parametres.get('fromuceids', False) :
+ self.fromuceids()
#create database
self.connect()
self.dobuild()
-
+
+ def fromuceids(self):
+ print 'fromuceids'
+ dictucekeep = dict(zip(self.parametres['uceids'], self.parametres['uceids']))
+ idpara = 0
+ for uci in self.ori.ucis :
+ if uci.paras == [] :
+ keepuces = [CopyUce(uce) for uce in uci.uces if uce.ident in dictucekeep]
+ if keepuces != [] :
+ nuci = CopyUci(uci)
+ nuci.uces = keepuces
+ self.corpus.ucis.append(nuci)
+ idpara += 1
+ else :
+ newuces = []
+ newpara = []
+ for et in uci.paras :
+ keepuces = [CopyUce(uce) for uce in uci.uces if uce.ident in dictucekeep and uce.para == idpara]
+ idpara += 1
+ if keepuces != [] :
+ newuces += keepuces
+ newpara.append(et)
+ if newuces != [] :
+ nuci = CopyUci(uci)
+ nuci.uces = newuces
+ nuci.paras = newpara
+ self.corpus.ucis.append(nuci)
+
def read_corpus(self, infile = None):
self.olduceid = [uce.ident for uci in self.corpus.ucis for uce in uci.uces]
ident_uci = 0
@@ -1464,7 +1592,7 @@ class BuildFromAlceste(BuildCorpus) :
raise Exception('EmptyText %i' % linenb)
if iduci != -1 and iduce != -1:
self.backup_uce()
- else :
+ else :
log.info(_(u"No Text in corpus. Are you sure of the formatting ?"))
raise Exception('TextBeforeTextMark %i' % linenb)
except UnicodeDecodeError :
@@ -1510,7 +1638,7 @@ class BuildFromAlceste(BuildCorpus) :
out.append(uce)
reste, texte_uce, suite = self.decouper(suite, self.ucesize + 15, self.ucesize)
uce = ' '.join([val for val in texte_uce if val not in self.ponctuation_espace])
- if uce != '' :
+ if uce != '' :
out.append(uce)
return out
else :
@@ -1525,6 +1653,7 @@ class Builder :
def __init__(self, parent, dlg = None) :
self.parent = parent
self.dlg = dlg
+
parametres = DoConf(os.path.join(self.parent.UserConfigPath,'corpus.cfg')).getoptions('corpus')
parametres['pathout'] = PathOut(parent.filename, 'corpus').mkdirout()
parametres['corpus_name'] = os.path.split(parametres['pathout'])[1]
@@ -1533,20 +1662,29 @@ class Builder :
dial.txtpath.SetLabel(parent.filename)
#dial.repout_choices.SetValue(parametres['pathout'])
self.res = dial.ShowModal()
+ if self.dlg is not None :
+ self.dlg = progressbar(self.parent, self.dlg)
if self.res == 5100 :
parametres = dial.doparametres()
parametres['originalpath'] = parent.filename
PathOut().createdir(parametres['pathout'])
- ReadLexique(self.parent, lang = parametres['lang'])
+ if parametres.get('dictionary', False) :
+ filein = parametres['dictionary']
+ else :
+ filein = None
+ if dial.corpusname.GetValue() != '' :
+ parametres['corpus_name'] = dial.corpusname.GetValue()
+ dial.Destroy()
+ ReadLexique(self.parent, lang = parametres['lang'], filein = filein)
if parametres['lang'] != 'other' and os.path.exists(self.parent.DictPath.get(parametres['lang']+'_exp', 'french_exp')):
self.parent.expressions = ReadDicoAsDico(self.parent.DictPath.get(parametres['lang']+'_exp', 'french_exp'))
else :
self.parent.expressions = {}
self.parametres = parametres
else :
+ dial.Destroy()
if self.dlg is not None :
self.dlg.Destroy()
- dial.Destroy()
def doanalyse(self) :
return BuildFromAlceste(self.parent.filename, self.parametres, self.parent.lexique, self.parent.expressions, dlg = self.dlg).corpus
@@ -1565,9 +1703,12 @@ class SubBuilder :
parametres['meta'] = corpus.make_etoiles()
elif parametres.get('fromtheme', False) :
parametres['meta'] = corpus.make_themes()
+ elif parametres.get('fromclusters', False) :
+ parametres['meta'] = [' '.join(['classe', `i`]) for i in range(1,parametres['clnb'] + 1)]
else :
parametres['meta'] = []
- parametres['meta'].sort()
+ if 'fromclusters' not in parametres :
+ parametres['meta'].sort()
if dlg is not None :
del busy
dial = SubTextFromMetaDial(parent, parametres)
@@ -1585,17 +1726,193 @@ class SubBuilder :
i += 1
parametres['pathout'] = pathout + '_%i' % i
meta = dial.m_listBox1.GetSelections()
- parametres['meta'] = [parametres['meta'][val] for val in meta]
+ if not 'fromclusters' in parametres :
+ parametres['meta'] = [parametres['meta'][val] for val in meta]
+ else :
+ parametres['meta'] = meta
self.parametres = parametres
dial.Destroy()
else :
dial.Destroy()
-
+
def doanalyse(self):
return BuildSubCorpus(self.ori, parametres = self.parametres, dlg = self.dlg).corpus
-if __name__ == '__main__' :
- t1 = time()
- parametres = {'formesdb':'formes.db', 'ucesdb': 'uces.db', 'corpusdb' : 'corpus.db', 'syscoding' : 'utf-8', 'encoding' : 'utf8'}
- intro = BuildCorpus(infile, parametres)#, tar_in, tar_infouce)#, tar_formes)
- print time() - t1
+class BuildMergeFromClusters(BuildCorpus):
+ def __init__(self, analyses, parametres, dlg = None) :
+ log.info('begin subcorpus...')
+ self.dlg = dlg
+ self.infile = None
+ self.corpus = Corpus(self, {'type' : 'corpus', 'originalpath' : 'MergeFromClusters', 'encoding' : 'merge'})
+ self.last = 0
+ self.analyses = analyses
+ self.lcl = []
+ self.parametres = parametres
+ #self.encoding = corpus.parametres['encoding']
+ self.corpus.parametres['corpus_name'] = parametres['corpus_name']
+ self.corpus.pathout = PathOut(filename = 'MFC', dirout = parametres['pathout'])
+ self.corpus.pathout.createdir(parametres['pathout'])
+ self.corpus.parametres['pathout'] = parametres['pathout']
+ self.corpus.parametres['meta'] = parametres.get('meta', False)
+ self.corpus.parametres['uuid'] = str(uuid4())
+ for i, analyse in enumerate(analyses) :
+ self.lcl.append([])
+ self.analyseid = i
+ corpus_uuid = analyse['corpus']
+ #if corpus_uuid not in self.parent.history.openedcorpus :
+ irapath = parametres['corpusira'][i]
+ corpus = Corpus(self, parametres = DoConf(irapath).getoptions('corpus'), read = True)
+ ucepath = os.path.join(analyse['pathout'], 'uce.csv')
+ corpus.make_ucecl_from_R(ucepath)
+ self.ori = corpus
+ for j, cl in enumerate(parametres['clusters'][i]) :
+ #print cl, self.ori.lc[cl-1]
+ self.parametres['uceids'] = self.ori.lc[cl-1]#[st for st in self.ori['lc'][cl-1]]
+ self.lcl[i] += self.ori.lc[cl-1]
+ self.et = parametres['newet'][i][j]
+ self.fromuceids()
+ #create database
+ self.connect()
+ self.dobuild()
+
+ def fromuceids(self):
+ print 'fromuceids'
+ dictucekeep = dict(zip(self.parametres['uceids'], self.parametres['uceids']))
+ idpara = 0
+ for uci in self.ori.ucis :
+ if uci.paras == [] :
+ keepuces = [CopyUce(uce) for uce in uci.uces if uce.ident in dictucekeep]
+ if keepuces != [] :
+ nuci = CopyUci(uci)
+ nuci.uces = keepuces
+ nuci.etoiles.append(self.et)
+ nuci.analyseid = self.analyseid
+ self.corpus.ucis.append(nuci)
+ idpara += 1
+ else :
+ newuces = []
+ newpara = []
+ for et in uci.paras :
+ keepuces = [CopyUce(uce) for uce in uci.uces if uce.ident in dictucekeep]
+ idpara += 1
+ if keepuces != [] :
+ newuces += keepuces
+ newpara.append(et)
+ if newuces != [] :
+ nuci = CopyUci(uci)
+ nuci.uces = newuces
+ nuci.paras = newpara
+ nuci.etoiles.append(self.et)
+ nuci.analyseid = self.analyseid
+ self.corpus.ucis.append(nuci)
+ #print nuci.etoiles, nuci.ident, nuci.uces
+
+ def read_corpus(self, infile = None):
+ #self.olduceid = [uce.ident for uci in self.corpus.ucis for uce in uci.uces]
+ ident_uci = 0
+ ident_uce = 0
+ ident_para = -1
+ lastpara = -1
+ newuceident = {}
+ print 'redo text, para and st ident'
+ for uci in self.corpus.ucis :
+ #print uci.ident, ident_uci, [uce.ident for uce in uci.uces], uci.etoiles
+ uci.ident = ident_uci
+ ident_uci += 1
+ for uce in uci.uces :
+ uce.uci = uci.ident
+ if uce.para != lastpara :
+ ident_para += 1
+ lastpara = uce.para
+ uce.para = ident_para
+ else :
+ uce.para = ident_para
+ newuceident['%i-%i' %(uci.analyseid, uce.ident)] = ident_uce
+ uce.ident = ident_uce
+ #print uce.ident
+ ident_uce += 1
+ print 'backup st text and forms'
+ rowid = 0
+ for i, analyse in enumerate(self.analyses) :
+ #print analyse, self.parametres['corpusira']
+ irapath = self.parametres['corpusira'][i]
+ old = Corpus(self, parametres = DoConf(irapath).getoptions('corpus'), read = True)
+ for row in old.getconcorde(self.lcl[i]) :
+ self.c.execute('INSERT INTO uces VALUES(?,?);', (newuceident['%i-%i' % (i,row[0])], row[1]))
+ for word in row[1].split() :
+ self.corpus.add_word_from_forme(old.formes[word], newuceident['%i-%i' % (i,row[0])])
+ rowid += 1
+ self.backup_uce()
+ print 'done'
+
+
+class MergeClusters :
+ def __init__(self, parent, parametres = None, dlg = None):
+ self.parent = parent
+ #self.ori = corpus
+ self.dlg = dlg
+ corpus_name = 'MergeFromClusters'
+ if dlg is not None :
+ busy = wx.BusyInfo(_("Please wait...").decode('utf8'), self)
+ wx.SafeYield()
+ parametres['corpus_name'] = corpus_name
+ if dlg is not None :
+ del busy
+ dial = MergeClusterFrame(parent)
+ dial.m_textCtrl4.SetValue(corpus_name)
+ self.res = dial.ShowModal()
+ if self.res == 5100 :
+ self.analyses = {}
+ self.clusters = {}
+ self.newet = {}
+ self.corpusira = {}
+ if dial.m_textCtrl4.GetValue() != '' :
+ corpus_name = ''.join([l for l in dial.m_textCtrl4.GetValue() if l.isalnum() or l in ['_']])
+ if corpus_name != '' :
+ parametres['corpus_name'] = corpus_name
+ else :
+ parametres['corpus_name'] = 'MergeFromClusters'
+ for cl in dial.selected :
+ corpus_uuid = cl[1]
+ #if corpus_uuid not in self.parent.history.openedcorpus :
+ irapath = self.parent.history.corpus[corpus_uuid]['ira']
+ #corpus = Corpus(self.parent, parametres = DoConf(irapath).getoptions('corpus'), read = True)
+ #self.parent.history.openedcorpus[corpus_uuid] = corpus
+ if cl[0] not in self.analyses :
+ analyse = DoConf(dial.irapath[cl[0]]).getoptions()
+ #ucepath = os.path.join(os.path.dirname(dial.irapath[cl[0]]), 'uce.csv')
+ #corpus = copycorpus(self.parent.history.openedcorpus[corpus_uuid])
+ #corpus.make_ucecl_from_R(ucepath)
+ self.analyses[cl[0]] = analyse
+ self.clusters[cl[0]] = [cl[2]]
+ self.newet[cl[0]] = [dial.selected[cl]]
+ self.corpusira[cl[0]] = irapath
+ else :
+ self.clusters[cl[0]].append(cl[2])
+ self.newet[cl[0]].append(dial.selected[cl])
+
+
+ analyses = [val for val in self.clusters]
+ clusters = [self.clusters[val] for val in analyses]
+ self.newet = [self.newet[val] for val in analyses]
+ corpusira = [self.corpusira[val] for val in analyses]
+ analyses = [self.analyses[val] for val in analyses]
+ pathout = os.path.dirname(os.path.dirname(analyses[0]['pathout']))
+ self.analyses = analyses
+
+ pathout = os.path.join(pathout, parametres['corpus_name'])
+ i = 1
+ while os.path.exists(pathout + '_%i' % i) :
+ i += 1
+ parametres['pathout'] = pathout + '_%i' % i
+ self.parametres = parametres
+ self.parametres['clusters'] = clusters
+ self.parametres['newet'] = self.newet
+ self.parametres['corpusira'] = corpusira
+ dial.Destroy()
+ else :
+ dial.Destroy()
+
+ def doanalyse(self):
+ return BuildMergeFromClusters(self.analyses, parametres = self.parametres, dlg = self.dlg).corpus
+