Spaces:
Sleeping
Sleeping
arithescientist
commited on
Commit
·
31e79df
1
Parent(s):
a04de99
Update app.py
Browse files
app.py
CHANGED
@@ -27,12 +27,23 @@ bert_legal_model = Summarizer(custom_model=custom_model, custom_tokenizer=custom
|
|
27 |
|
28 |
|
29 |
def pdf(file):
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
|
35 |
-
return
|
36 |
#pageObject.extractText()
|
37 |
|
38 |
|
@@ -41,7 +52,7 @@ def pdf(file):
|
|
41 |
|
42 |
iface = gr.Interface(
|
43 |
pdf,
|
44 |
-
"file", "
|
45 |
)
|
46 |
|
47 |
if __name__ == "__main__":
|
|
|
27 |
|
28 |
|
29 |
def pdf(file):
|
30 |
+
|
31 |
+
path = os.getcwd()
|
32 |
+
folder_name = './'
|
33 |
+
path = os.path.join(path, folder_name)
|
34 |
+
|
35 |
+
list_of_files = []
|
36 |
+
for root, dirs, files in os.walk(path):
|
37 |
+
for file in files:
|
38 |
+
if(file.endswith(".pdf")):
|
39 |
+
# print(os.path.join(root,file))
|
40 |
+
list_of_files.append(os.path.join(root,file))
|
41 |
+
|
42 |
+
total_pages = print("\nProcessing {} files...\n".format(len(list_of_files)))
|
43 |
+
|
44 |
+
|
45 |
|
46 |
+
return total_pages
|
47 |
#pageObject.extractText()
|
48 |
|
49 |
|
|
|
52 |
|
53 |
iface = gr.Interface(
|
54 |
pdf,
|
55 |
+
"file", "text"
|
56 |
)
|
57 |
|
58 |
if __name__ == "__main__":
|