Spaces:
Sleeping
Sleeping
Neurolingua
commited on
Commit
•
29191e5
1
Parent(s):
93f2b25
Update other_function.py
Browse files- other_function.py +15 -0
other_function.py
CHANGED
@@ -7,12 +7,27 @@ from io import BytesIO
|
|
7 |
import pandas as pd
|
8 |
from urllib.parse import urlparse
|
9 |
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
from pypdf import PdfReader
|
11 |
from ai71 import AI71
|
12 |
import os
|
13 |
import PyPDF2
|
14 |
import pandas as pd
|
|
|
|
|
|
|
|
|
15 |
|
|
|
|
|
|
|
|
|
16 |
from inference_sdk import InferenceHTTPClient
|
17 |
import base64
|
18 |
UPLOAD_FOLDER = '/code/uploads'
|
|
|
7 |
import pandas as pd
|
8 |
from urllib.parse import urlparse
|
9 |
import os
|
10 |
+
import py
|
11 |
+
from pdf2image import convert_from_path
|
12 |
+
import cv2
|
13 |
+
import numpy as np
|
14 |
+
import pytesseract
|
15 |
+
import subprocess
|
16 |
+
from PIL import Image
|
17 |
from pypdf import PdfReader
|
18 |
from ai71 import AI71
|
19 |
import os
|
20 |
import PyPDF2
|
21 |
import pandas as pd
|
22 |
+
def extract_text_from_image(image_path):
|
23 |
+
img = cv2.imread(image_path)
|
24 |
+
if img is None:
|
25 |
+
raise ValueError("Image not found or unable to load")
|
26 |
|
27 |
+
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
|
28 |
+
text = pytesseract.image_to_string(img_rgb)
|
29 |
+
|
30 |
+
return text
|
31 |
from inference_sdk import InferenceHTTPClient
|
32 |
import base64
|
33 |
UPLOAD_FOLDER = '/code/uploads'
|