Create utils.py
Browse files
utils.py
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from PyPDF2 import PdfReader
|
2 |
+
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
|
3 |
+
from gtts import gTTS
|
4 |
+
import os
|
5 |
+
|
6 |
+
# Download the summarization model and tokenizer
|
7 |
+
model_name = "ArtifactAI/led_large_16384_arxiv_summarization"
|
8 |
+
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
|
9 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
10 |
+
|
11 |
+
def summarize_and_speak_pdf_abstract(pdf_path):
|
12 |
+
"""
|
13 |
+
Reads a PDF file, extracts the abstract, summarizes it in one sentence, and generates an audio file of the summary.
|
14 |
+
|
15 |
+
Args:
|
16 |
+
pdf_path: Path to the PDF file.
|
17 |
+
"""
|
18 |
+
|
19 |
+
# Summarize the abstract
|
20 |
+
summary = summarize_pdf_abstract(pdf_path)
|
21 |
+
|
22 |
+
# Define language and audio format
|
23 |
+
language = "en" # Change this to your desired language
|
24 |
+
audio_format = "mp3"
|
25 |
+
|
26 |
+
# Create the text-to-speech object
|
27 |
+
tts = gTTS(text=summary, lang=language)
|
28 |
+
|
29 |
+
# Generate the audio file
|
30 |
+
audio_file_name = f"summary.{audio_format}"
|
31 |
+
tts.save(audio_file_name)
|
32 |
+
|
33 |
+
print(f"Audio file created: {audio_file_name}")
|
34 |
+
|
35 |
+
# Play the audio file (optional)
|
36 |
+
# os.system(f"play {audio_file_name}")
|
37 |
+
|
38 |
+
|
39 |
+
def summarize_pdf_abstract(pdf_path):
|
40 |
+
"""
|
41 |
+
Reads a PDF file, extracts the abstract, and summarizes it in one sentence.
|
42 |
+
|
43 |
+
Args:
|
44 |
+
pdf_path: Path to the PDF file.
|
45 |
+
|
46 |
+
Returns:
|
47 |
+
A string containing the one-sentence summary of the abstract.
|
48 |
+
"""
|
49 |
+
|
50 |
+
# Read the PDF file
|
51 |
+
reader = PdfReader(open(pdf_path, "rb"))
|
52 |
+
|
53 |
+
# Extract the abstract
|
54 |
+
abstract_text = ""
|
55 |
+
for page in reader.pages:
|
56 |
+
# Search for keywords like "Abstract" or "Introduction"
|
57 |
+
if (
|
58 |
+
"Abstract" in page.extract_text
|