Spaces:
Paused
Paused
Commit
·
cc60679
1
Parent(s):
0159ca5
file loading testing
Browse files- helper_functions.py +24 -11
helper_functions.py
CHANGED
|
@@ -9,18 +9,31 @@ import os
|
|
| 9 |
import functools
|
| 10 |
import requests
|
| 11 |
|
| 12 |
-
def process_file(
|
| 13 |
-
#
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
else:
|
| 22 |
-
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
def load_documents_from_url(url):
|
| 26 |
try:
|
|
|
|
| 9 |
import functools
|
| 10 |
import requests
|
| 11 |
|
| 12 |
+
def process_file(uploaded_file):
|
| 13 |
+
# Save the file temporarily to process it
|
| 14 |
+
temp_file_path = f"/tmp/{uploaded_file.name}"
|
| 15 |
+
|
| 16 |
+
with open(temp_file_path, "wb") as temp_file:
|
| 17 |
+
# Write the uploaded file content to the temporary file
|
| 18 |
+
temp_file.write(uploaded_file.read())
|
| 19 |
+
|
| 20 |
+
# Determine the file type and load it accordingly
|
| 21 |
+
if uploaded_file.name.endswith(".pdf"):
|
| 22 |
+
# Load PDF with PyMuPDFLoader
|
| 23 |
+
loader = PyMuPDFLoader(temp_file_path)
|
| 24 |
+
elif uploaded_file.name.endswith(".txt"):
|
| 25 |
+
# Load text file with TextLoader
|
| 26 |
+
loader = TextLoader(temp_file_path)
|
| 27 |
else:
|
| 28 |
+
raise ValueError("Unsupported file format. Only PDF and TXT are supported.")
|
| 29 |
+
|
| 30 |
+
# Load documents from the file
|
| 31 |
+
documents = loader.load()
|
| 32 |
+
|
| 33 |
+
# Clean up the temporary file
|
| 34 |
+
os.remove(temp_file_path)
|
| 35 |
+
|
| 36 |
+
return documents
|
| 37 |
|
| 38 |
def load_documents_from_url(url):
|
| 39 |
try:
|