Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -69,7 +69,7 @@ def pdf_pil(file_path,page_num):
|
|
| 69 |
pdf = pdfium.PdfDocument("data.pdf")
|
| 70 |
#n_pages = len(pdf)
|
| 71 |
#for page_number in range(n_pages):
|
| 72 |
-
page = pdf.get_page(int(page_num))
|
| 73 |
bitmap = page.render(
|
| 74 |
scale = 1, # 72dpi resolution
|
| 75 |
rotation = 0, # no additional rotation
|
|
@@ -88,7 +88,7 @@ def ocrpdf(file_path,pdf_lang,page_num):
|
|
| 88 |
|
| 89 |
for bound in bounds:
|
| 90 |
print(bound[1])
|
| 91 |
-
|
| 92 |
|
| 93 |
def scrape(instring):
|
| 94 |
html_src=(f'''
|
|
|
|
| 69 |
pdf = pdfium.PdfDocument("data.pdf")
|
| 70 |
#n_pages = len(pdf)
|
| 71 |
#for page_number in range(n_pages):
|
| 72 |
+
page = pdf.get_page(int(page_num)-1)
|
| 73 |
bitmap = page.render(
|
| 74 |
scale = 1, # 72dpi resolution
|
| 75 |
rotation = 0, # no additional rotation
|
|
|
|
| 88 |
|
| 89 |
for bound in bounds:
|
| 90 |
print(bound[1])
|
| 91 |
+
|
| 92 |
|
| 93 |
def scrape(instring):
|
| 94 |
html_src=(f'''
|