124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
|
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
|
+
+
+
+
+
+
+
+
+
|
else:
aRes = [ (sMorph, self.lexicographer.readableMorph(sMorph)) for sMorph in lMorph ]
if aRes:
lWordAndMorph.append((sElem, aRes))
return lWordAndMorph
def readableMorph (self, sMorph):
"returns a human readable meaning of tags of <sMorph>"
if not self.lexicographer:
return ""
return self.lexicographer.readableMorph(sMorph)
def setLabelsOnToken (self, dToken):
"""on <dToken>,
adds:
- lMorph: list of morphologies
- aLabels: list of labels (human readable meaning of tags)
for WORD tokens:
- bValidToken: True if the token is valid for the spellchecker
- lSubTokens for each parts of the split token
"""
if not self.lexicographer:
return
if "lMorph" not in dToken:
dToken["lMorph"] = self.getMorph(dToken["sValue"])
if dToken["sType"] == "WORD":
dToken["bValidToken"] = self.isValidToken(dToken["sValue"])
sPrefix, sStem, sSuffix = self.lexicographer.split(dToken["sValue"])
|