Spaces:
Running
Running
Oh Gyuhyeok
commited on
Commit
โข
59e8fda
1
Parent(s):
360c8b1
Implement PDF Upload (#2)
Browse files* Implement PDF Upload
1. Remove line breaks
2. Translate pdfs
- easyword_translator/pdf.py +35 -0
- easyword_translator/run.py +109 -2
easyword_translator/pdf.py
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import pymupdf # PyMuPDF
|
2 |
+
|
3 |
+
|
4 |
+
def read_pdf(file_path):
|
5 |
+
# Open the PDF file
|
6 |
+
document = pymupdf.open(file_path)
|
7 |
+
text = ""
|
8 |
+
|
9 |
+
# Iterate through the pages
|
10 |
+
for page_num in range(len(document)):
|
11 |
+
# Extract text from each page
|
12 |
+
page = document.load_page(page_num)
|
13 |
+
text += page.get_text()
|
14 |
+
|
15 |
+
# Close the PDF document
|
16 |
+
document.close()
|
17 |
+
|
18 |
+
return text
|
19 |
+
|
20 |
+
|
21 |
+
def remove_line_breaks(text):
|
22 |
+
# remove only single line breaks, not paragraphs
|
23 |
+
# find line breaks and it is not followed by a period
|
24 |
+
for i in range(len(text)):
|
25 |
+
if i == 0 or i == len(text) - 1:
|
26 |
+
continue
|
27 |
+
if text[i] == "\n" and text[i - 1] != "." and text[i + 1] != "\n":
|
28 |
+
text = text[:i] + " " + text[i + 1 :]
|
29 |
+
return text
|
30 |
+
|
31 |
+
|
32 |
+
if __name__ == "__main__":
|
33 |
+
file_path = "example3.pdf"
|
34 |
+
pdf_text = read_pdf(file_path)
|
35 |
+
print(remove_line_breaks(pdf_text))
|
easyword_translator/run.py
CHANGED
@@ -9,6 +9,8 @@ from langchain_core.prompts import ChatPromptTemplate
|
|
9 |
from langchain_upstage import ChatUpstage
|
10 |
from rapidfuzz import process
|
11 |
|
|
|
|
|
12 |
warnings.filterwarnings("ignore")
|
13 |
|
14 |
|
@@ -125,15 +127,102 @@ def translate(sentence: str) -> str:
|
|
125 |
f"์ ๋ฌธ์ฉ์ด๋ฅผ ๋ฒ์ญํ์ผ๋ฉด ๋ฐ๋์ ์์ด๋ฅผ ๊ดํธ[]์ ๋ฃ์ด์ ๋ฐ๋ผ ๋ถ์ฌ์ผ ํด. '์คํํ๋ฆ[control]'์ฒ๋ผ. ๋ฐฉ๊ธ ๋ฒ์ญํ '{refined_translation}'์์, ์๋ ๋ฌธ์ฅ '{sentence}'์ ์ฌ์ฉ๋ ์์ด๋ฅผ ์ฉ์ด ๋ฐ๋ก ๋ค์ ๊ดํธ []์ ๋ฃ์ด์ ๋ฐ๋ผ ๋ถ์ฌ์ค.",
|
126 |
),
|
127 |
]
|
128 |
-
|
|
|
|
|
|
|
|
|
|
|
129 |
logger.info(refined_translation)
|
130 |
|
131 |
refined_translation = refined_translation.replace("[", "(").replace("]", ")")
|
132 |
return refined_translation
|
133 |
|
134 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
with gr.Blocks() as demo:
|
136 |
-
with gr.Tab("
|
137 |
chatbot = gr.Interface(
|
138 |
fn=translate,
|
139 |
inputs=gr.Textbox(label="Enter your text"),
|
@@ -150,6 +239,24 @@ with gr.Blocks() as demo:
|
|
150 |
title=TITLE,
|
151 |
description=DESCRIPTION,
|
152 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
153 |
|
154 |
|
155 |
def main():
|
|
|
9 |
from langchain_upstage import ChatUpstage
|
10 |
from rapidfuzz import process
|
11 |
|
12 |
+
import pymupdf
|
13 |
+
|
14 |
warnings.filterwarnings("ignore")
|
15 |
|
16 |
|
|
|
127 |
f"์ ๋ฌธ์ฉ์ด๋ฅผ ๋ฒ์ญํ์ผ๋ฉด ๋ฐ๋์ ์์ด๋ฅผ ๊ดํธ[]์ ๋ฃ์ด์ ๋ฐ๋ผ ๋ถ์ฌ์ผ ํด. '์คํํ๋ฆ[control]'์ฒ๋ผ. ๋ฐฉ๊ธ ๋ฒ์ญํ '{refined_translation}'์์, ์๋ ๋ฌธ์ฅ '{sentence}'์ ์ฌ์ฉ๋ ์์ด๋ฅผ ์ฉ์ด ๋ฐ๋ก ๋ค์ ๊ดํธ []์ ๋ฃ์ด์ ๋ฐ๋ผ ๋ถ์ฌ์ค.",
|
128 |
),
|
129 |
]
|
130 |
+
try:
|
131 |
+
refined_translation = chainer(messages).invoke({})
|
132 |
+
except Exception as e:
|
133 |
+
logger.error(e)
|
134 |
+
break
|
135 |
+
|
136 |
logger.info(refined_translation)
|
137 |
|
138 |
refined_translation = refined_translation.replace("[", "(").replace("]", ")")
|
139 |
return refined_translation
|
140 |
|
141 |
|
142 |
+
class PDFFile:
|
143 |
+
def __init__(self):
|
144 |
+
self.file_list = []
|
145 |
+
|
146 |
+
def read_pdf(self, file_path: str) -> str:
|
147 |
+
# Open the PDF file
|
148 |
+
document = pymupdf.open(file_path)
|
149 |
+
text = ""
|
150 |
+
|
151 |
+
# Iterate through the pages
|
152 |
+
for page_num in range(len(document)):
|
153 |
+
# Extract text from each page
|
154 |
+
page = document.load_page(page_num)
|
155 |
+
text += page.get_text()
|
156 |
+
|
157 |
+
# Close the PDF document
|
158 |
+
document.close()
|
159 |
+
|
160 |
+
return text
|
161 |
+
|
162 |
+
def remove_line_breaks(self, text: str) -> list[str]:
|
163 |
+
# remove only single line breaks, not paragraphs
|
164 |
+
# find line breaks and it is not followed by a period
|
165 |
+
for i in range(len(text)):
|
166 |
+
if i == 0 or i == len(text) - 1:
|
167 |
+
continue
|
168 |
+
if text[i] == "\n" and text[i - 1] != "." and text[i + 1] != "\n":
|
169 |
+
text = text[:i] + " " + text[i + 1 :]
|
170 |
+
return text
|
171 |
+
|
172 |
+
def upload_file(self, file_path: str) -> list[str]:
|
173 |
+
self.file_list.append(file_path)
|
174 |
+
return self.file_list
|
175 |
+
|
176 |
+
def transalte_pdf(
|
177 |
+
self,
|
178 |
+
remove_line_breaks: bool,
|
179 |
+
save_before_translation: bool,
|
180 |
+
) -> str:
|
181 |
+
if not self.file_list:
|
182 |
+
return "No file uploaded yet."
|
183 |
+
|
184 |
+
file_out_list = []
|
185 |
+
for file in self.file_list:
|
186 |
+
directory = os.path.dirname(file)
|
187 |
+
filename = os.path.basename(file)
|
188 |
+
|
189 |
+
# remove extension
|
190 |
+
filename = ".".join(filename.split(".")[:-1])
|
191 |
+
pdf_text = self.read_pdf(file)
|
192 |
+
if remove_line_breaks:
|
193 |
+
pdf_text = self.remove_line_breaks(pdf_text)
|
194 |
+
|
195 |
+
if save_before_translation:
|
196 |
+
with open(f"{directory}/{filename}_pre.txt", "w") as f:
|
197 |
+
f.write(pdf_text)
|
198 |
+
file_out_list.append(f"{directory}/{filename}_pre.txt")
|
199 |
+
|
200 |
+
# translation = translate(pdf_text)
|
201 |
+
# Translation with divide and conquer with 50 sentences
|
202 |
+
translation = ""
|
203 |
+
# seperate the text into sentences
|
204 |
+
sentences = pdf_text.split(".")
|
205 |
+
for i in range(0, len(sentences), 50):
|
206 |
+
translation += translate(".".join(sentences[i : i + 50])) + ". "
|
207 |
+
|
208 |
+
with open(f"{directory}/{filename}_translated.txt", "w") as f:
|
209 |
+
f.write(translation)
|
210 |
+
file_out_list.append(f"{directory}/{filename}_translated.txt")
|
211 |
+
|
212 |
+
self.file_list = []
|
213 |
+
# Zip the files
|
214 |
+
# import zipfile
|
215 |
+
|
216 |
+
# with zipfile.ZipFile(f"{directory}/translated_files.zip", "w") as z:
|
217 |
+
# for file in file_out_list:
|
218 |
+
# z.write(file)
|
219 |
+
|
220 |
+
# return f"{directory}/translated_files.zip"
|
221 |
+
return file_out_list
|
222 |
+
|
223 |
+
|
224 |
with gr.Blocks() as demo:
|
225 |
+
with gr.Tab("TEXT"):
|
226 |
chatbot = gr.Interface(
|
227 |
fn=translate,
|
228 |
inputs=gr.Textbox(label="Enter your text"),
|
|
|
239 |
title=TITLE,
|
240 |
description=DESCRIPTION,
|
241 |
)
|
242 |
+
with gr.Tab("PDF"):
|
243 |
+
pdf_file = PDFFile()
|
244 |
+
upload_button = gr.UploadButton(
|
245 |
+
label="Upload PDF",
|
246 |
+
file_types=[".pdf"],
|
247 |
+
)
|
248 |
+
upload_file_list_box = gr.File(label="Uploaded Files")
|
249 |
+
|
250 |
+
upload_button.upload(pdf_file.upload_file, upload_button, upload_file_list_box)
|
251 |
+
|
252 |
+
run_translator = gr.Interface(
|
253 |
+
fn=pdf_file.transalte_pdf,
|
254 |
+
inputs=[
|
255 |
+
gr.Checkbox(label="Remove line breaks"),
|
256 |
+
gr.Checkbox(label="Save before translation"),
|
257 |
+
],
|
258 |
+
outputs=[gr.File(label="Download Translated Files")],
|
259 |
+
)
|
260 |
|
261 |
|
262 |
def main():
|