Update modules/text_analysis/semantic_analysis.py
Browse files
modules/text_analysis/semantic_analysis.py
CHANGED
@@ -71,7 +71,100 @@ ENTITY_LABELS = {
|
|
71 |
}
|
72 |
}
|
73 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
74 |
##############################################################################################################
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
75 |
def perform_semantic_analysis(text, nlp, lang_code):
|
76 |
"""
|
77 |
Realiza el análisis semántico completo del texto.
|
@@ -125,25 +218,53 @@ def fig_to_html(fig):
|
|
125 |
|
126 |
|
127 |
|
128 |
-
def identify_key_concepts(doc):
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
133 |
|
134 |
|
135 |
def create_concept_graph(doc, key_concepts):
|
|
|
|
|
|
|
136 |
G = nx.Graph()
|
137 |
-
for concept
|
138 |
-
|
139 |
for sent in doc.sents:
|
140 |
-
|
141 |
-
for
|
142 |
-
|
|
|
|
|
|
|
|
|
|
|
143 |
if G.has_edge(concept1, concept2):
|
144 |
G[concept1][concept2]['weight'] += 1
|
145 |
else:
|
146 |
G.add_edge(concept1, concept2, weight=1)
|
|
|
147 |
return G
|
148 |
|
149 |
def visualize_concept_graph(G, lang_code):
|
|
|
71 |
}
|
72 |
}
|
73 |
|
74 |
+
CUSTOM_STOPWORDS = {
|
75 |
+
'es': {
|
76 |
+
# Artículos
|
77 |
+
'el', 'la', 'los', 'las', 'un', 'una', 'unos', 'unas',
|
78 |
+
# Preposiciones comunes
|
79 |
+
'a', 'ante', 'bajo', 'con', 'contra', 'de', 'desde', 'en',
|
80 |
+
'entre', 'hacia', 'hasta', 'para', 'por', 'según', 'sin',
|
81 |
+
'sobre', 'tras', 'durante', 'mediante',
|
82 |
+
# Conjunciones
|
83 |
+
'y', 'e', 'ni', 'o', 'u', 'pero', 'sino', 'porque',
|
84 |
+
# Pronombres
|
85 |
+
'yo', 'tú', 'él', 'ella', 'nosotros', 'vosotros', 'ellos',
|
86 |
+
'ellas', 'este', 'esta', 'ese', 'esa', 'aquel', 'aquella',
|
87 |
+
# Verbos auxiliares comunes
|
88 |
+
'ser', 'estar', 'haber', 'tener',
|
89 |
+
# Palabras comunes en textos académicos
|
90 |
+
'además', 'también', 'asimismo', 'sin embargo', 'no obstante',
|
91 |
+
'por lo tanto', 'entonces', 'así', 'luego', 'pues',
|
92 |
+
# Números escritos
|
93 |
+
'uno', 'dos', 'tres', 'primer', 'primera', 'segundo', 'segunda',
|
94 |
+
# Otras palabras comunes
|
95 |
+
'cada', 'todo', 'toda', 'todos', 'todas', 'otro', 'otra',
|
96 |
+
'donde', 'cuando', 'como', 'que', 'cual', 'quien',
|
97 |
+
'cuyo', 'cuya', 'hay', 'solo', 'ver', 'si', 'no',
|
98 |
+
# Símbolos y caracteres especiales
|
99 |
+
'#', '@', '/', '*', '+', '-', '=', '$', '%'
|
100 |
+
},
|
101 |
+
'en': {
|
102 |
+
# Articles
|
103 |
+
'the', 'a', 'an',
|
104 |
+
# Common prepositions
|
105 |
+
'in', 'on', 'at', 'by', 'for', 'with', 'about', 'against',
|
106 |
+
'between', 'into', 'through', 'during', 'before', 'after',
|
107 |
+
'above', 'below', 'to', 'from', 'up', 'down', 'of',
|
108 |
+
# Conjunctions
|
109 |
+
'and', 'or', 'but', 'nor', 'so', 'for', 'yet',
|
110 |
+
# Pronouns
|
111 |
+
'i', 'you', 'he', 'she', 'it', 'we', 'they', 'this',
|
112 |
+
'that', 'these', 'those', 'my', 'your', 'his', 'her',
|
113 |
+
# Auxiliary verbs
|
114 |
+
'be', 'am', 'is', 'are', 'was', 'were', 'been', 'have',
|
115 |
+
'has', 'had', 'do', 'does', 'did',
|
116 |
+
# Common academic words
|
117 |
+
'therefore', 'however', 'thus', 'hence', 'moreover',
|
118 |
+
'furthermore', 'nevertheless',
|
119 |
+
# Numbers written
|
120 |
+
'one', 'two', 'three', 'first', 'second', 'third',
|
121 |
+
# Other common words
|
122 |
+
'where', 'when', 'how', 'what', 'which', 'who',
|
123 |
+
'whom', 'whose', 'there', 'here', 'just', 'only',
|
124 |
+
# Symbols and special characters
|
125 |
+
'#', '@', '/', '*', '+', '-', '=', '$', '%'
|
126 |
+
],
|
127 |
+
'fr': {
|
128 |
+
# Articles
|
129 |
+
'le', 'la', 'les', 'un', 'une', 'des',
|
130 |
+
# Prepositions
|
131 |
+
'à', 'de', 'dans', 'sur', 'en', 'par', 'pour', 'avec',
|
132 |
+
'sans', 'sous', 'entre', 'derrière', 'chez', 'avant',
|
133 |
+
# Conjunctions
|
134 |
+
'et', 'ou', 'mais', 'donc', 'car', 'ni', 'or',
|
135 |
+
# Pronouns
|
136 |
+
'je', 'tu', 'il', 'elle', 'nous', 'vous', 'ils',
|
137 |
+
'elles', 'ce', 'cette', 'ces', 'celui', 'celle',
|
138 |
+
# Auxiliary verbs
|
139 |
+
'être', 'avoir', 'faire',
|
140 |
+
# Academic words
|
141 |
+
'donc', 'cependant', 'néanmoins', 'ainsi', 'toutefois',
|
142 |
+
'pourtant', 'alors',
|
143 |
+
# Numbers
|
144 |
+
'un', 'deux', 'trois', 'premier', 'première', 'second',
|
145 |
+
# Other common words
|
146 |
+
'où', 'quand', 'comment', 'que', 'qui', 'quoi',
|
147 |
+
'quel', 'quelle', 'plus', 'moins',
|
148 |
+
# Symbols
|
149 |
+
'#', '@', '/', '*', '+', '-', '=', '$', '%'
|
150 |
+
}
|
151 |
+
}
|
152 |
+
|
153 |
##############################################################################################################
|
154 |
+
def get_stopwords(lang_code):
|
155 |
+
"""
|
156 |
+
Obtiene el conjunto de stopwords para un idioma específico.
|
157 |
+
Combina las stopwords de spaCy con las personalizadas.
|
158 |
+
"""
|
159 |
+
try:
|
160 |
+
nlp = spacy.load(f'{lang_code}_core_news_sm')
|
161 |
+
spacy_stopwords = nlp.Defaults.stop_words
|
162 |
+
custom_stopwords = CUSTOM_STOPWORDS.get(lang_code, set())
|
163 |
+
return spacy_stopwords.union(custom_stopwords)
|
164 |
+
except:
|
165 |
+
return CUSTOM_STOPWORDS.get(lang_code, set())
|
166 |
+
|
167 |
+
|
168 |
def perform_semantic_analysis(text, nlp, lang_code):
|
169 |
"""
|
170 |
Realiza el análisis semántico completo del texto.
|
|
|
218 |
|
219 |
|
220 |
|
221 |
+
def identify_key_concepts(doc, min_freq=2, min_length=3):
|
222 |
+
"""
|
223 |
+
Identifica conceptos clave en el texto, excluyendo stopwords
|
224 |
+
y aplicando criterios de frecuencia y longitud.
|
225 |
+
"""
|
226 |
+
stopwords = get_stopwords(doc.lang_)
|
227 |
+
word_freq = Counter()
|
228 |
+
|
229 |
+
for token in doc:
|
230 |
+
if (token.text.lower() not in stopwords and # No es stopword
|
231 |
+
token.is_alpha and # Es alfabético
|
232 |
+
len(token.text) >= min_length and # Longitud mínima
|
233 |
+
not token.is_punct and # No es puntuación
|
234 |
+
not token.like_num): # No es número
|
235 |
+
|
236 |
+
# Usar el lema en lugar del token para unificar variantes
|
237 |
+
word_freq[token.lemma_] += 1
|
238 |
+
|
239 |
+
# Filtrar por frecuencia mínima y ordenar por frecuencia
|
240 |
+
key_concepts = [(word, freq) for word, freq in word_freq.items()
|
241 |
+
if freq >= min_freq]
|
242 |
+
key_concepts.sort(key=lambda x: x[1], reverse=True)
|
243 |
+
|
244 |
+
return key_concepts[:10] # Retornar los 10 conceptos más frecuentes
|
245 |
|
246 |
|
247 |
def create_concept_graph(doc, key_concepts):
|
248 |
+
"""
|
249 |
+
Crea un grafo de relaciones entre conceptos.
|
250 |
+
"""
|
251 |
G = nx.Graph()
|
252 |
+
concept_words = {concept[0] for concept in key_concepts}
|
253 |
+
|
254 |
for sent in doc.sents:
|
255 |
+
sentence_concepts = []
|
256 |
+
for token in sent:
|
257 |
+
if token.lemma_ in concept_words:
|
258 |
+
sentence_concepts.append(token.lemma_)
|
259 |
+
|
260 |
+
# Crear conexiones entre conceptos en la misma oración
|
261 |
+
for i, concept1 in enumerate(sentence_concepts):
|
262 |
+
for concept2 in sentence_concepts[i+1:]:
|
263 |
if G.has_edge(concept1, concept2):
|
264 |
G[concept1][concept2]['weight'] += 1
|
265 |
else:
|
266 |
G.add_edge(concept1, concept2, weight=1)
|
267 |
+
|
268 |
return G
|
269 |
|
270 |
def visualize_concept_graph(G, lang_code):
|