Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
import os
|
2 |
import io
|
3 |
-
from flask import Flask, request, jsonify
|
4 |
from werkzeug.utils import secure_filename
|
5 |
from PyPDF2 import PdfReader
|
6 |
from docx import Document
|
@@ -14,36 +14,43 @@ ALLOWED_EXTENSIONS = {"pdf", "docx", "pptx", "txt"}
|
|
14 |
def allowed_file(filename):
|
15 |
return "." in filename and filename.rsplit(".", 1)[1].lower() in ALLOWED_EXTENSIONS
|
16 |
|
|
|
|
|
|
|
|
|
17 |
@app.route("/summarize", methods=["POST"])
|
18 |
def summarize():
|
19 |
if "file" not in request.files:
|
20 |
return jsonify({"error": "No file uploaded"}), 400
|
21 |
-
|
22 |
file = request.files["file"]
|
23 |
|
24 |
if file.filename == "":
|
25 |
return jsonify({"error": "No selected file"}), 400
|
26 |
-
|
27 |
if not allowed_file(file.filename):
|
28 |
return jsonify({"error": "Unsupported file format"}), 400
|
29 |
-
|
30 |
filename = secure_filename(file.filename)
|
31 |
file_content = file.read()
|
32 |
-
|
33 |
# Process file based on type
|
34 |
summary = None
|
35 |
file_ext = filename.rsplit(".", 1)[1].lower()
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
|
|
|
|
|
|
47 |
|
48 |
# Summarization functions
|
49 |
def summarize_pdf(file_content):
|
@@ -58,12 +65,16 @@ def summarize_docx(file_content):
|
|
58 |
|
59 |
def summarize_pptx(file_content):
|
60 |
ppt = Presentation(io.BytesIO(file_content))
|
61 |
-
text =
|
62 |
-
|
|
|
|
|
|
|
|
|
63 |
|
64 |
def summarize_txt(file_content):
|
65 |
text = file_content.decode("utf-8")
|
66 |
return text[:500]
|
67 |
|
68 |
if __name__ == "__main__":
|
69 |
-
app.run(host="0.0.0.0", port=7860)
|
|
|
1 |
import os
|
2 |
import io
|
3 |
+
from flask import Flask, request, jsonify, render_template
|
4 |
from werkzeug.utils import secure_filename
|
5 |
from PyPDF2 import PdfReader
|
6 |
from docx import Document
|
|
|
14 |
def allowed_file(filename):
|
15 |
return "." in filename and filename.rsplit(".", 1)[1].lower() in ALLOWED_EXTENSIONS
|
16 |
|
17 |
+
@app.route("/", methods=["GET"])
|
18 |
+
def index():
|
19 |
+
return "Document Summarizer API is running! Use /summarize endpoint for POST requests."
|
20 |
+
|
21 |
@app.route("/summarize", methods=["POST"])
|
22 |
def summarize():
|
23 |
if "file" not in request.files:
|
24 |
return jsonify({"error": "No file uploaded"}), 400
|
25 |
+
|
26 |
file = request.files["file"]
|
27 |
|
28 |
if file.filename == "":
|
29 |
return jsonify({"error": "No selected file"}), 400
|
30 |
+
|
31 |
if not allowed_file(file.filename):
|
32 |
return jsonify({"error": "Unsupported file format"}), 400
|
33 |
+
|
34 |
filename = secure_filename(file.filename)
|
35 |
file_content = file.read()
|
36 |
+
|
37 |
# Process file based on type
|
38 |
summary = None
|
39 |
file_ext = filename.rsplit(".", 1)[1].lower()
|
40 |
+
|
41 |
+
try:
|
42 |
+
if file_ext == "pdf":
|
43 |
+
summary = summarize_pdf(file_content)
|
44 |
+
elif file_ext == "docx":
|
45 |
+
summary = summarize_docx(file_content)
|
46 |
+
elif file_ext == "pptx":
|
47 |
+
summary = summarize_pptx(file_content)
|
48 |
+
elif file_ext == "txt":
|
49 |
+
summary = summarize_txt(file_content)
|
50 |
+
|
51 |
+
return jsonify({"filename": filename, "summary": summary})
|
52 |
+
except Exception as e:
|
53 |
+
return jsonify({"error": f"Error processing file: {str(e)}"}), 500
|
54 |
|
55 |
# Summarization functions
|
56 |
def summarize_pdf(file_content):
|
|
|
65 |
|
66 |
def summarize_pptx(file_content):
|
67 |
ppt = Presentation(io.BytesIO(file_content))
|
68 |
+
text = []
|
69 |
+
for slide in ppt.slides:
|
70 |
+
for shape in slide.shapes:
|
71 |
+
if hasattr(shape, "text"):
|
72 |
+
text.append(shape.text)
|
73 |
+
return "\n".join(text)[:500]
|
74 |
|
75 |
def summarize_txt(file_content):
|
76 |
text = file_content.decode("utf-8")
|
77 |
return text[:500]
|
78 |
|
79 |
if __name__ == "__main__":
|
80 |
+
app.run(host="0.0.0.0", port=7860, debug=True)
|