Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -31,6 +31,15 @@ from pdf2image import convert_from_bytes, convert_from_path # pip install pd
|
|
| 31 |
import requests
|
| 32 |
from utils import convert_json_to_markdown, extract_json_content
|
| 33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
def setup_poppler_linux():
|
| 35 |
poppler_dir = "/tmp/poppler"
|
| 36 |
if not os.path.exists(poppler_dir):
|
|
@@ -258,7 +267,7 @@ async def doc_parser(doc_path, prompt, model_id):
|
|
| 258 |
|
| 259 |
try:
|
| 260 |
json_pages = extract_json_content(pages)
|
| 261 |
-
json_pages =
|
| 262 |
except Exception as e:
|
| 263 |
json_pages = pages
|
| 264 |
all_pages.append(extract_makrdown(json_pages))
|
|
|
|
| 31 |
import requests
|
| 32 |
from utils import convert_json_to_markdown, extract_json_content
|
| 33 |
|
| 34 |
+
def json_inline_list(data):
|
| 35 |
+
if isinstance(data, dict):
|
| 36 |
+
for k, v in data.items():
|
| 37 |
+
if isinstance(v, list):
|
| 38 |
+
data[k] = "[" + ", ".join(str(x) for x in v) + "]"
|
| 39 |
+
pretty = json.dumps(data, indent=2, ensure_ascii=False)
|
| 40 |
+
return f"```json\n{pretty}\n```"
|
| 41 |
+
|
| 42 |
+
|
| 43 |
def setup_poppler_linux():
|
| 44 |
poppler_dir = "/tmp/poppler"
|
| 45 |
if not os.path.exists(poppler_dir):
|
|
|
|
| 267 |
|
| 268 |
try:
|
| 269 |
json_pages = extract_json_content(pages)
|
| 270 |
+
json_pages = json_inline_list(json.loads(json_pages))
|
| 271 |
except Exception as e:
|
| 272 |
json_pages = pages
|
| 273 |
all_pages.append(extract_makrdown(json_pages))
|