Spaces:
Sleeping
Sleeping
drmurataltun
commited on
Commit
•
cccd854
1
Parent(s):
cc4a955
Upload 2 files
Browse files- app.py +153 -0
- requirements.txt +6 -0
app.py
ADDED
@@ -0,0 +1,153 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
import streamlit as st
|
3 |
+
import google.generativeai as genai
|
4 |
+
import markdown
|
5 |
+
from docx import Document
|
6 |
+
from bs4 import BeautifulSoup
|
7 |
+
from PyPDF2 import PdfFileReader
|
8 |
+
import tempfile
|
9 |
+
import os
|
10 |
+
|
11 |
+
# Configure the API key
|
12 |
+
genai.configure(api_key=os.getenv('gemini_api'))
|
13 |
+
|
14 |
+
# Function to convert PDF to text
|
15 |
+
def pdf_to_text(file):
|
16 |
+
with open(file, 'rb') as f:
|
17 |
+
pdf = PdfFileReader(f)
|
18 |
+
text = ""
|
19 |
+
for page_num in range(pdf.numPages):
|
20 |
+
page = pdf.getPage(page_num)
|
21 |
+
text += page.extract_text()
|
22 |
+
return text
|
23 |
+
|
24 |
+
# Function to upload file to the Generative AI API
|
25 |
+
def upload_file(file_path):
|
26 |
+
st.write("Uploading file...")
|
27 |
+
text_file = genai.upload_file(path=file_path)
|
28 |
+
st.write(f"Completed upload: {text_file.uri}")
|
29 |
+
return text_file
|
30 |
+
|
31 |
+
# Function to convert text to Markdown
|
32 |
+
def to_markdown(text):
|
33 |
+
text = text.replace('•', ' *')
|
34 |
+
return textwrap.indent(text, '> ', predicate=lambda _: True)
|
35 |
+
|
36 |
+
chat_session = None
|
37 |
+
|
38 |
+
# Function to build the model
|
39 |
+
def build_model(text_file):
|
40 |
+
global chat_session
|
41 |
+
generation_config = {
|
42 |
+
"temperature": 0.2,
|
43 |
+
"top_p": 0.95,
|
44 |
+
"top_k": 64,
|
45 |
+
"max_output_tokens": 8192,
|
46 |
+
"response_mime_type": "text/plain",
|
47 |
+
}
|
48 |
+
|
49 |
+
model = genai.GenerativeModel(
|
50 |
+
model_name="gemini-1.5-flash",
|
51 |
+
generation_config=generation_config,
|
52 |
+
system_instruction="""Yüklenen belgedeki bilgilere göre Türkçe cevap ver.
|
53 |
+
Eğer sorunun cevabı belgede bulunmuyorsa 'Belgede Cevap Bulunmuyor' yaz.
|
54 |
+
""",
|
55 |
+
)
|
56 |
+
|
57 |
+
chat_session = model.start_chat(history=[])
|
58 |
+
response = chat_session.send_message(["Yüklenen belgeyi bir cümle ile özetle", text_file])
|
59 |
+
|
60 |
+
st.markdown(to_markdown(response.text))
|
61 |
+
|
62 |
+
# Function to interact with the chat model
|
63 |
+
def chat(prompt):
|
64 |
+
try:
|
65 |
+
response = chat_session.send_message(prompt)
|
66 |
+
markdown_text = to_markdown(response.text)
|
67 |
+
st.markdown(markdown_text)
|
68 |
+
return response.text
|
69 |
+
except ValueError:
|
70 |
+
st.write(response.prompt_feedback)
|
71 |
+
st.write(response.candidates[0].finish_reason)
|
72 |
+
st.write(response.candidates[0].safety_ratings)
|
73 |
+
except Exception as e:
|
74 |
+
st.write("An unexpected error occurred:", e)
|
75 |
+
|
76 |
+
# Function to generate a report based on questions
|
77 |
+
def generate_report(questions):
|
78 |
+
report_text = "\n## SORULARINIZ VE CEVAPLARI\n"
|
79 |
+
for question in questions:
|
80 |
+
report_text += f"\n## {question}\n"
|
81 |
+
answer = chat(question)
|
82 |
+
report_text += f"\n{answer}\n"
|
83 |
+
return report_text
|
84 |
+
|
85 |
+
# Function to convert Markdown to HTML
|
86 |
+
def convert_Markdown_to_HTML(report_text):
|
87 |
+
html_text = markdown.markdown(report_text)
|
88 |
+
return html_text
|
89 |
+
|
90 |
+
# Function to add HTML to a Word document
|
91 |
+
def add_html_to_word(html_text, doc):
|
92 |
+
soup = BeautifulSoup(html_text, 'html.parser')
|
93 |
+
for element in soup:
|
94 |
+
if element.name == 'h1':
|
95 |
+
doc.add_heading(element.get_text(), level=1)
|
96 |
+
elif element.name == 'h2':
|
97 |
+
doc.add_heading(element.get_text(), level=2)
|
98 |
+
elif element.name == 'h3':
|
99 |
+
doc.add_heading(element.get_text(), level=3)
|
100 |
+
elif element.name == 'h4':
|
101 |
+
doc.add_heading(element.get_text(), level=4)
|
102 |
+
elif element.name == 'h5':
|
103 |
+
doc.add_heading(element.get_text(), level=5)
|
104 |
+
elif element.name == 'h6':
|
105 |
+
doc.add_heading(element.get_text(), level=6)
|
106 |
+
elif element.name == 'p':
|
107 |
+
doc.add_paragraph(element.get_text())
|
108 |
+
elif element.name == 'ul':
|
109 |
+
for li in element.find_all('li'):
|
110 |
+
doc.add_paragraph(li.get_text(), style='List Bullet')
|
111 |
+
elif element.name == 'ol':
|
112 |
+
for li in element.find_all('li'):
|
113 |
+
doc.add_paragraph(li.get_text(), style='List Number')
|
114 |
+
elif element.name:
|
115 |
+
doc.add_paragraph(element.get_text())
|
116 |
+
|
117 |
+
# Streamlit interface
|
118 |
+
st.title("REPORT GENERATOR: ASK YOUR QUESTIONS TO A PDF FILE BY MURAT KARAKAYA AKADEMI")
|
119 |
+
st.write("Upload a PDF to ask questions and get the answers.")
|
120 |
+
|
121 |
+
uploaded_file = st.file_uploader("Upload PDF", type="pdf")
|
122 |
+
questions_input = st.text_area("Enter Questions", placeholder="Type your questions here, one per line.", height=150)
|
123 |
+
|
124 |
+
if uploaded_file and questions_input:
|
125 |
+
with tempfile.NamedTemporaryFile(delete=False) as temp_file:
|
126 |
+
temp_file.write(uploaded_file.read())
|
127 |
+
temp_file_path = temp_file.name
|
128 |
+
|
129 |
+
text_content = pdf_to_text(temp_file_path)
|
130 |
+
text_file = upload_file(temp_file_path)
|
131 |
+
build_model(text_file)
|
132 |
+
|
133 |
+
questions = questions_input.split("\n")
|
134 |
+
report_text = generate_report(questions)
|
135 |
+
|
136 |
+
html_text = convert_Markdown_to_HTML(report_text)
|
137 |
+
doc = Document()
|
138 |
+
add_html_to_word(html_text, doc)
|
139 |
+
|
140 |
+
doc_name = os.path.basename(temp_file_path).replace(".pdf", ".docx")
|
141 |
+
doc_name = "Rapor " + doc_name
|
142 |
+
doc.save(doc_name)
|
143 |
+
|
144 |
+
st.markdown(report_text)
|
145 |
+
st.write("Document generated successfully!")
|
146 |
+
|
147 |
+
with open(doc_name, "rb") as file:
|
148 |
+
st.download_button(label="Download Report", data=file, file_name=doc_name)
|
149 |
+
|
150 |
+
os.remove(temp_file_path)
|
151 |
+
os.remove(doc_name)
|
152 |
+
|
153 |
+
genai.delete_file(text_file.name)
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit
|
2 |
+
google-generativeai
|
3 |
+
markdown
|
4 |
+
python-docx
|
5 |
+
beautifulsoup4
|
6 |
+
PyPDF2
|