Create semantic_analysis.py
Browse files
modules/text_analysis/semantic_analysis.py
ADDED
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import spacy
|
3 |
+
import networkx as nx
|
4 |
+
import matplotlib.pyplot as plt
|
5 |
+
from collections import Counter
|
6 |
+
from sklearn.feature_extraction.text import TfidfVectorizer
|
7 |
+
from sklearn.metrics.pairwise import cosine_similarity
|
8 |
+
|
9 |
+
# ... (mantén las definiciones de POS_COLORS, POS_TRANSLATIONS, y ENTITY_LABELS como están)
|
10 |
+
|
11 |
+
def identify_key_concepts(doc):
|
12 |
+
word_freq = Counter([token.lemma_.lower() for token in doc if token.pos_ in ['NOUN', 'VERB'] and not token.is_stop])
|
13 |
+
return word_freq.most_common(10) # Top 10 conceptos clave
|
14 |
+
|
15 |
+
def create_concept_graph(text, concepts):
|
16 |
+
vectorizer = TfidfVectorizer()
|
17 |
+
tfidf_matrix = vectorizer.fit_transform([text])
|
18 |
+
concept_vectors = vectorizer.transform([c[0] for c in concepts])
|
19 |
+
similarity_matrix = cosine_similarity(concept_vectors, concept_vectors)
|
20 |
+
|
21 |
+
G = nx.Graph()
|
22 |
+
for i, (concept, weight) in enumerate(concepts):
|
23 |
+
G.add_node(concept, weight=weight)
|
24 |
+
for j in range(i+1, len(concepts)):
|
25 |
+
if similarity_matrix[i][j] > 0.1:
|
26 |
+
G.add_edge(concept, concepts[j][0], weight=similarity_matrix[i][j])
|
27 |
+
|
28 |
+
return G
|
29 |
+
|
30 |
+
def visualize_concept_graph(G, lang):
|
31 |
+
fig, ax = plt.subplots(figsize=(15, 10))
|
32 |
+
pos = nx.spring_layout(G, k=0.5, iterations=50)
|
33 |
+
|
34 |
+
node_sizes = [G.nodes[node]['weight'] * 100 for node in G.nodes()]
|
35 |
+
nx.draw_networkx_nodes(G, pos, node_size=node_sizes, node_color='lightblue', alpha=0.8, ax=ax)
|
36 |
+
nx.draw_networkx_labels(G, pos, font_size=10, font_weight="bold", ax=ax)
|
37 |
+
nx.draw_networkx_edges(G, pos, width=1, alpha=0.5, ax=ax)
|
38 |
+
|
39 |
+
edge_labels = nx.get_edge_attributes(G, 'weight')
|
40 |
+
nx.draw_networkx_edge_labels(G, pos, edge_labels=edge_labels, font_size=8, ax=ax)
|
41 |
+
|
42 |
+
title = {
|
43 |
+
'es': "Relaciones entre Conceptos Clave",
|
44 |
+
'en': "Key Concept Relations",
|
45 |
+
'fr': "Relations entre Concepts Clés"
|
46 |
+
}
|
47 |
+
ax.set_title(title[lang], fontsize=16)
|
48 |
+
ax.axis('off')
|
49 |
+
|
50 |
+
plt.tight_layout()
|
51 |
+
return fig
|
52 |
+
|
53 |
+
def perform_semantic_analysis(text, nlp, lang):
|
54 |
+
doc = nlp(text)
|
55 |
+
|
56 |
+
# Identificar conceptos clave
|
57 |
+
key_concepts = identify_key_concepts(doc)
|
58 |
+
|
59 |
+
# Crear y visualizar grafo de conceptos
|
60 |
+
concept_graph = create_concept_graph(text, key_concepts)
|
61 |
+
relations_graph = visualize_concept_graph(concept_graph, lang)
|
62 |
+
|
63 |
+
return {
|
64 |
+
'key_concepts': key_concepts,
|
65 |
+
'relations_graph': relations_graph
|
66 |
+
}
|
67 |
+
|
68 |
+
__all__ = ['perform_semantic_analysis', 'ENTITY_LABELS', 'POS_TRANSLATIONS']
|