File size: 4,988 Bytes
4e66fbc
 
 
 
 
 
 
 
7f0f733
 
 
 
4e66fbc
 
 
 
 
7f0f733
8a35b07
7f0f733
 
 
4e66fbc
3f1d535
 
 
 
 
 
 
 
 
 
 
 
 
4e66fbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3f1d535
4e66fbc
64ecac0
e858d98
 
 
 
 
3f1d535
 
 
 
 
 
 
 
 
 
e858d98
 
3f1d535
4e66fbc
 
 
921f45c
4e66fbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# AUTOGENERATED! DO NOT EDIT! File to edit: ../drive/MyDrive/Codici/Python/Apps/Gradio_App/Langchain_apps/langchain_summarization_app.ipynb.

# %% auto 0
__all__ = ['hub_llm', 'title', 'description', 'combine_prompt_template', 'pdf_example_1', 'pdf_example_2', 'prompt_example_1',
           'prompt_example_2', 'upload_file_input', 'custom_prompt_input', 'custom_chunk_input', 'chunk_size_input',
           'chunk_overlap_input', 'examples', 'outputs', 'iface', 'summarize']

# %% ../drive/MyDrive/Codici/Python/Apps/Gradio_App/Langchain_apps/langchain_summarization_app.ipynb 3
from langchain_community.llms import HuggingFaceHub
from langchain_community.document_loaders import PyPDFLoader
from langchain_core.prompts import PromptTemplate
from langchain.chains import LLMChain
from langchain.text_splitter import CharacterTextSplitter
from langchain.chains.mapreduce import MapReduceChain
from langchain.prompts import PromptTemplate
from langchain.docstore.document import Document
from langchain.chains.summarize import load_summarize_chain

import os
import dotenv
from dotenv import load_dotenv
load_dotenv()

# %% ../drive/MyDrive/Codici/Python/Apps/Gradio_App/Langchain_apps/langchain_summarization_app.ipynb 5
hub_llm = HuggingFaceHub(
    repo_id="facebook/bart-large-cnn", # facebook/bart-large-cnn or "google/flan-t5-base" or "google/pegasus-xsum"
    model_kwargs={
        "temperature": 0.01, # Controls randomness (0.0: deterministic, 1.0: very random)
        "max_new_tokens": 256*2,  # Maximum number of tokens to generate in the summary
        "min_length": 30,  # Minimum length of the generated summary
        "repetition_penalty": 1.2,  # Penalizes repeated tokens (higher value = less repetition)
        "top_k": 50,  # Consider only the top k most likely tokens when generating
        "top_p": 0.95,  # Consider tokens with cumulative probability up to top_p
        "early_stopping": True, # Stops generation when a certain condition is met (e.g., end-of-sequence token)
    }
)

# %% ../drive/MyDrive/Codici/Python/Apps/Gradio_App/Langchain_apps/langchain_summarization_app.ipynb 15
from langchain.text_splitter import RecursiveCharacterTextSplitter
import gradio as gr
import time

title="PDF Summarizer"
description="Summarize your PDF using a custom combine prompt."

# Default combine_prompt
combine_prompt_template = """Write a comprehensive summary of this academic article.

Divide the summary in:
1. Main Objective of the paper
2. Results

{text}

SUMMARY:"""

# Example PDF files and prompts
pdf_example_1 = './ZeroShotDataAug.pdf'
pdf_example_2 = './bert.pdf'
prompt_example_1 = """Write a comprehensive summary of this academic article.

Divide the summary in:
1. Main Objective of the paper
2. Results

{text}

SUMMARY:"""
prompt_example_2 = """Summarize the following document focusing on the key findings and methodology.

{text}

Summary:"""

# Implementation
def summarize(pdf_file, custom_prompt, custom_chunk, chunk_size, chunk_overlap):
    try:
        # Get the uploaded file path
        file_path = pdf_file.name

        # Load and process the PDF
        loader = PyPDFLoader(file_path)
        if custom_chunk:
            text_splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap)
            docs = loader.load_and_split(text_splitter=text_splitter)
        else:
            docs = loader.load_and_split()

        PROMPT = PromptTemplate(template=custom_prompt, input_variables=['text'])
        chain = load_summarize_chain(hub_llm, chain_type='map_reduce', combine_prompt=PROMPT)

        # Introduce a delay before calling the API
        time.sleep(2)
        summary = chain.invoke(docs)['output_text']
        return summary
    except Exception as e:
        return f"An error occurred: {e}"

upload_file_input = gr.UploadButton(label="Upload PDF", file_types=[".pdf"], file_count="single")
custom_prompt_input = gr.Textbox(label="Custom Prompt",
                                 lines=10,
                                 value=combine_prompt_template,
                                 info="Define your own prompt or leave empty for default.")
custom_chunk_input = gr.Checkbox(label="Custom Chunk", value=False, info="Recommended to be left unchecked")
chunk_size_input = gr.Number(label="Chunk Size", value=700,minimum=500,maximum=1000,step=100)
chunk_overlap_input = gr.Number(label="Chunk Overlap", value=50,minimum=10,maximum=100,step=100)

examples=[
    [pdf_example_1, prompt_example_1, False, 700, 50],
    # [pdf_example_2, prompt_example_2, False, 700, 50]
    ]

outputs = gr.Textbox(label="Summary")

iface = gr.Interface(
    title=title,
    description=description,
    fn=summarize,
    inputs=[upload_file_input,
            custom_prompt_input,
            custom_chunk_input,
            chunk_size_input,
            chunk_overlap_input
            ],
    outputs=outputs,
    examples=examples,
)

iface.launch(
    debug=False,
    # share=False,
    )