Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -29,7 +29,7 @@ from io import BytesIO
|
|
| 29 |
from pdf2image import convert_from_bytes, convert_from_path # pip install pdf2image
|
| 30 |
|
| 31 |
import requests
|
| 32 |
-
|
| 33 |
|
| 34 |
def setup_poppler_linux():
|
| 35 |
poppler_dir = "/tmp/poppler"
|
|
@@ -256,6 +256,8 @@ async def doc_parser(doc_path, prompt, model_id):
|
|
| 256 |
pages += chunk
|
| 257 |
yield extract_makrdown(pages), pages
|
| 258 |
all_pages.append(extract_makrdown(pages))
|
|
|
|
|
|
|
| 259 |
all_pages_raw.append(pages)
|
| 260 |
print(all_pages)
|
| 261 |
yield "\n---\n".join(all_pages), "\n\n".join(all_pages_raw)
|
|
|
|
| 29 |
from pdf2image import convert_from_bytes, convert_from_path # pip install pdf2image
|
| 30 |
|
| 31 |
import requests
|
| 32 |
+
from utils import convert_json_to_markdown
|
| 33 |
|
| 34 |
def setup_poppler_linux():
|
| 35 |
poppler_dir = "/tmp/poppler"
|
|
|
|
| 256 |
pages += chunk
|
| 257 |
yield extract_makrdown(pages), pages
|
| 258 |
all_pages.append(extract_makrdown(pages))
|
| 259 |
+
|
| 260 |
+
pages = convert_json_to_markdown(pages)
|
| 261 |
all_pages_raw.append(pages)
|
| 262 |
print(all_pages)
|
| 263 |
yield "\n---\n".join(all_pages), "\n\n".join(all_pages_raw)
|