andreeabodea
commited on
Commit
•
388323e
1
Parent(s):
e60ebb7
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,173 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Hugging Face's logo
|
2 |
+
Hugging Face
|
3 |
+
Search models, datasets, users...
|
4 |
+
Models
|
5 |
+
Datasets
|
6 |
+
Spaces
|
7 |
+
Posts
|
8 |
+
Docs
|
9 |
+
Solutions
|
10 |
+
Pricing
|
11 |
+
|
12 |
+
|
13 |
+
|
14 |
+
Spaces:
|
15 |
+
|
16 |
+
andreeabodea
|
17 |
+
/
|
18 |
+
Extract_Project_Report_Section_1
|
19 |
+
|
20 |
+
|
21 |
+
like
|
22 |
+
0
|
23 |
+
|
24 |
+
Logs
|
25 |
+
App
|
26 |
+
Files
|
27 |
+
Community
|
28 |
+
Settings
|
29 |
+
Extract_Project_Report_Section_1
|
30 |
+
/
|
31 |
+
app.py
|
32 |
+
|
33 |
+
andreeabodea's picture
|
34 |
+
andreeabodea
|
35 |
+
Update app.py
|
36 |
+
536f374
|
37 |
+
VERIFIED
|
38 |
+
about 2 hours ago
|
39 |
+
raw
|
40 |
+
history
|
41 |
+
blame
|
42 |
+
edit
|
43 |
+
delete
|
44 |
+
No virus
|
45 |
+
5.51 kB
|
46 |
+
import os
|
47 |
+
import pdfplumber
|
48 |
+
import re
|
49 |
+
import gradio as gr
|
50 |
+
from transformers import pipeline, AutoModelForQuestionAnswering, AutoTokenizer
|
51 |
+
from io import BytesIO
|
52 |
+
import torch
|
53 |
+
|
54 |
+
"""
|
55 |
+
Extract the text from a section of a PDF file between 'wanted_section' and 'next_section'.
|
56 |
+
Parameters:
|
57 |
+
- path (str): The file path to the PDF file.
|
58 |
+
- wanted_section (str): The section to start extracting text from.
|
59 |
+
- next_section (str): The section to stop extracting text at.
|
60 |
+
Returns:
|
61 |
+
- text (str): The extracted text from the specified section range.
|
62 |
+
"""
|
63 |
+
|
64 |
+
|
65 |
+
def get_section(path, wanted_section, next_section):
|
66 |
+
print(wanted_section)
|
67 |
+
|
68 |
+
# Open the PDF file
|
69 |
+
doc = pdfplumber.open(BytesIO(path))
|
70 |
+
start_page = []
|
71 |
+
end_page = []
|
72 |
+
|
73 |
+
# Find the all the pages for the specified sections
|
74 |
+
for page in range(len(doc.pages)):
|
75 |
+
if len(doc.pages[page].search(wanted_section, return_chars=False, case=False)) > 0:
|
76 |
+
start_page.append(page)
|
77 |
+
if len(doc.pages[page].search(next_section, return_chars=False, case=False)) > 0:
|
78 |
+
end_page.append(page)
|
79 |
+
|
80 |
+
# Extract the text between the start and end page of the wanted section
|
81 |
+
text = []
|
82 |
+
for page_num in range(max(start_page), max(end_page)+1):
|
83 |
+
page = doc.pages[page_num]
|
84 |
+
text.append(page.extract_text())
|
85 |
+
text = " ".join(text)
|
86 |
+
final_text = text.replace("\n", " ")
|
87 |
+
return final_text
|
88 |
+
|
89 |
+
|
90 |
+
def extract_between(big_string, start_string, end_string):
|
91 |
+
# Use a non-greedy match for content between start_string and end_string
|
92 |
+
pattern = re.escape(start_string) + '(.*?)' + re.escape(end_string)
|
93 |
+
match = re.search(pattern, big_string, re.DOTALL)
|
94 |
+
|
95 |
+
if match:
|
96 |
+
# Return the content without the start and end strings
|
97 |
+
return match.group(1)
|
98 |
+
else:
|
99 |
+
# Return None if the pattern is not found
|
100 |
+
return None
|
101 |
+
|
102 |
+
def format_section1(section1_text):
|
103 |
+
result_section1_dict = {}
|
104 |
+
|
105 |
+
result_section1_dict['TOPIC'] = extract_between(section1_text, "Sektor", "EZ-Programm")
|
106 |
+
result_section1_dict['PROGRAM'] = extract_between(section1_text, "Sektor", "EZ-Programm")
|
107 |
+
result_section1_dict['PROJECT DESCRIPTION'] = extract_between(section1_text, "EZ-Programmziel", "Datum der letzten BE")
|
108 |
+
result_section1_dict['PROJECT NAME'] = extract_between(section1_text, "Modul", "Modulziel")
|
109 |
+
result_section1_dict['OBJECTIVE'] = extract_between(section1_text, "Modulziel", "Berichtszeitraum")
|
110 |
+
result_section1_dict['PROGRESS'] = extract_between(section1_text, "Zielerreichung des Moduls", "Massnahme im Zeitplan")
|
111 |
+
result_section1_dict['STATUS'] = extract_between(section1_text, "Massnahme im Zeitplan", "Risikoeinschätzung")
|
112 |
+
result_section1_dict['RECOMMENDATIONS'] = extract_between(section1_text, "Vorschläge zur Modulanpas-", "Voraussichtliche")
|
113 |
+
|
114 |
+
return result_section1_dict
|
115 |
+
|
116 |
+
def answer_questions(text,language="de"):
|
117 |
+
# Initialize the zero-shot classification pipeline
|
118 |
+
model_name = "deepset/gelectra-large-germanquad"
|
119 |
+
model = AutoModelForQuestionAnswering.from_pretrained(model_name)
|
120 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
121 |
+
|
122 |
+
# Initialize the QA pipeline
|
123 |
+
qa_pipeline = pipeline("question-answering", model=model, tokenizer=tokenizer)
|
124 |
+
questions = [
|
125 |
+
"Welches ist das Titel des Moduls?",
|
126 |
+
"Welches ist das Sektor oder das Kernthema?",
|
127 |
+
"Welches ist das Land?",
|
128 |
+
"Zu welchem Program oder EZ-Programm gehort das Projekt?"
|
129 |
+
#"Welche Durchführungsorganisation aus den 4 Varianten 'giz', 'kfw', 'ptb' und 'bgr' implementiert das Projekt?"
|
130 |
+
# "In dem Dokument was steht bei Sektor?",
|
131 |
+
# "In dem Dokument was steht von 'EZ-Programm' bis 'EZ-Programmziel'?",
|
132 |
+
# "In dem Dokument was steht bei EZ-Programmziel?",
|
133 |
+
# "In dem Dokument in dem Abschnitt '1. Kurzbeschreibung' was steht bei Modul?",
|
134 |
+
# "In dem Dokument was steht bei Zielerreichung des Moduls?",
|
135 |
+
# "In dem Dokument in dem Abschnitt '1. Kurzbeschreibung' was steht bei Maßnahme im Zeitplan?",
|
136 |
+
# "In dem Dokument was steht bei Vorschläge zur Modulanpassung?",
|
137 |
+
# "In dem Dokument in dem Abschnitt 'Anlage 1: Wirkungsmatrix des Moduls' was steht unter Laufzeit als erstes Datum?",
|
138 |
+
# "In dem Dokument in dem Abschnitt 'Anlage 1: Wirkungsmatrix des Moduls' was steht unter Laufzeit als zweites Datum?"
|
139 |
+
]
|
140 |
+
|
141 |
+
# Iterate over each question and get answers
|
142 |
+
answers_dict = {}
|
143 |
+
|
144 |
+
for question in questions:
|
145 |
+
result = qa_pipeline(question=question, context=text)
|
146 |
+
# print(f"Question: {question}")
|
147 |
+
# print(f"Answer: {result['answer']}\n")
|
148 |
+
answers_dict[question] = result['answer']
|
149 |
+
return answers_dict
|
150 |
+
|
151 |
+
|
152 |
+
def process_pdf(path):
|
153 |
+
results_dict = {}
|
154 |
+
results_dict["1. Kurzbeschreibung"] = \
|
155 |
+
get_section(path, "1. Kurzbeschreibung", "2. Einordnung des Moduls")
|
156 |
+
answers = answer_questions(results_dict["1. Kurzbeschreibung"])
|
157 |
+
return answers
|
158 |
+
|
159 |
+
def get_first_page_text(file_data):
|
160 |
+
doc = pdfplumber.open(BytesIO(file_data))
|
161 |
+
if len(doc.pages):
|
162 |
+
return doc.pages[0].extract_text()
|
163 |
+
|
164 |
+
if __name__ == "__main__":
|
165 |
+
|
166 |
+
# Define the Gradio interface
|
167 |
+
# iface = gr.Interface(fn=process_pdf,
|
168 |
+
demo = gr.Interface(fn=process_pdf,
|
169 |
+
inputs=gr.File(type="binary", label="Upload PDF"),
|
170 |
+
outputs=gr.Textbox(label="Extracted Text"),
|
171 |
+
title="PDF Text Extractor",
|
172 |
+
description="Upload a PDF file to extract.")
|
173 |
+
demo.launch()
|