File size: 5,391 Bytes
6d5272f
56da2e5
 
 
 
6d5272f
56da2e5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6d5272f
 
56da2e5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
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
152
153
154
155
156
157
158
import gradio as gr
from modules.extractive import TFIDFSummarizer, TextRankSummarizer, CombinedSummarizer, BERTSummarizer
from modules.abstractive import load_summarizers, abstractive_summary
from modules.preprocessing import Preprocessor, PDFProcessor
from modules.utils import handle_long_text

# Cargar modelos abstractivos finetuneados
summarizers = load_summarizers()


# Función principal para generar resúmenes
def summarize(input_text, file, summary_type, method, num_sentences, model_name, max_length, num_beams):
    preprocessor = Preprocessor()

    if file is not None:
        pdf_processor = PDFProcessor()
        input_text = pdf_processor.pdf_to_text(file.name)

    if not input_text:
        return "Por favor, ingrese texto o cargue un archivo válido."

    cleaned_text = preprocessor.clean_text(input_text)

    if summary_type == "Extractivo":
        if method == "TF-IDF":
            summarizer = TFIDFSummarizer()
        elif method == "TextRank":
            summarizer = TextRankSummarizer()
        elif method == "BERT":
            summarizer = BERTSummarizer()
        elif method == "TF-IDF + TextRank":
            summarizer = CombinedSummarizer()
        else:
            return "Método no válido para resumen extractivo."

        return summarizer.summarize(
            preprocessor.split_into_sentences(cleaned_text),
            preprocessor.clean_sentences(preprocessor.split_into_sentences(cleaned_text)),
            num_sentences,
        )

    elif summary_type == "Abstractivo":
        if model_name not in summarizers:
            return "Modelo no disponible para resumen abstractivo."
        return handle_long_text(
            cleaned_text,
            summarizers[model_name][0],
            summarizers[model_name][1],
            max_length=max_length,
            stride=128,
        )

    elif summary_type == "Combinado":
        if model_name not in summarizers:
            return "Modelo no disponible para resumen abstractivo."
        extractive_summary = TFIDFSummarizer().summarize(
            preprocessor.split_into_sentences(cleaned_text),
            preprocessor.clean_sentences(preprocessor.split_into_sentences(cleaned_text)),
            num_sentences,
        )
        return handle_long_text(
            extractive_summary,
            summarizers[model_name][0],
            summarizers[model_name][1],
            max_length=max_length,
            stride=128,
        )

    return "Seleccione un tipo de resumen válido."


# Interfaz dinámica
with gr.Blocks() as interface:
    gr.Markdown("# Demo: Generador de Resúmenes Inteligente")

    # Entrada de texto o archivo
    with gr.Row():
        input_text = gr.Textbox(lines=9, label="Ingrese texto")
        file = gr.File(label="Subir archivo (PDF, TXT)")

    # Selección de tipo de resumen
    summary_type = gr.Radio(
        ["Extractivo", "Abstractivo", "Combinado"],
        label="Tipo de resumen",
        value="Extractivo",
    )

    # Opciones dinámicas
    method = gr.Radio(
        ["TF-IDF", "TextRank", "BERT", "TF-IDF + TextRank"],
        label="Método Extractivo",
        visible=True,
    )
    num_sentences = gr.Slider(
        1, 10, value=3, step=1, label="Número de oraciones (Extractivo)", visible=True
    )
    model_name = gr.Radio(
        ["Pegasus", "T5", "BART"],
        label="Modelo Abstractivo",
        visible=False,
    )
    max_length = gr.Slider(
        50, 300, value=128, step=10, label="Longitud máxima (Abstractivo)", visible=False
    )
    num_beams = gr.Slider(
        1, 10, value=4, step=1, label="Número de haces (Abstractivo)", visible=False
    )


    def update_options(summary_type):
        if summary_type == "Extractivo":
            return (
                gr.update(visible=True), gr.update(visible=True), gr.update(visible=False), gr.update(visible=False),
                gr.update(visible=False))
        elif summary_type == "Abstractivo":
            return (
                gr.update(visible=False), gr.update(visible=False), gr.update(visible=True), gr.update(visible=True),
                gr.update(visible=True))
        elif summary_type == "Combinado":
            return (gr.update(visible=True), gr.update(visible=True), gr.update(visible=True), gr.update(visible=True),
                    gr.update(visible=True))
        else:
            return (
                gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False),
                gr.update(visible=False))


    summary_type.change(
        update_options,
        inputs=[summary_type],
        outputs=[method, num_sentences, model_name, max_length, num_beams],
    )

    summarize_button = gr.Button("Generar Resumen")
    output = gr.Textbox(lines=10, label="Resumen generado", interactive=True)
    copy_button = gr.Button("Copiar Resumen")

    summarize_button.click(
        summarize,
        inputs=[input_text, file, summary_type, method, num_sentences, model_name, max_length, num_beams],
        outputs=output,
    )


    def copy_summary(summary):
        return summary


    copy_button.click(
        fn=copy_summary,
        inputs=[output],
        outputs=[output],
        js="""function(summary) { navigator.clipboard.writeText(summary); return summary; }""",
    )

if __name__ == "__main__":
    interface.launch()