Spaces:
Sleeping
Sleeping
Update backend/main.py
Browse files- backend/main.py +188 -59
backend/main.py
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
from fastapi import FastAPI, UploadFile, File, Form
|
2 |
from fastapi.staticfiles import StaticFiles
|
3 |
from fastapi.responses import FileResponse, JSONResponse
|
4 |
import cloudinary
|
@@ -6,12 +6,18 @@ import cloudinary.uploader
|
|
6 |
import requests
|
7 |
import io
|
8 |
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
import pandas as pd
|
10 |
from PyPDF2 import PdfReader
|
11 |
-
|
12 |
import textract
|
13 |
-
from google import genai
|
14 |
-
from google.genai import types
|
15 |
|
16 |
# Cloudinary Config
|
17 |
cloudinary.config(
|
@@ -23,79 +29,202 @@ cloudinary.config(
|
|
23 |
# Google Gemini Client
|
24 |
genai_client = genai.Client(api_key=os.getenv("GOOGLE_API_KEY"))
|
25 |
|
26 |
-
def
|
27 |
try:
|
28 |
-
result = cloudinary.uploader.upload(
|
|
|
|
|
|
|
29 |
return result.get("secure_url")
|
30 |
except Exception as e:
|
31 |
print(f"Cloudinary upload error: {e}")
|
32 |
return None
|
33 |
|
34 |
-
def
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
text += page.extract_text()
|
41 |
-
return text
|
42 |
-
|
43 |
-
# Convert Excel to text
|
44 |
-
elif file.content_type == "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" or file.content_type == "application/vnd.ms-excel":
|
45 |
-
df = pd.read_excel(io.BytesIO(file.file.read()))
|
46 |
-
return df.to_string()
|
47 |
-
|
48 |
-
# Convert PowerPoint to text
|
49 |
-
elif file.content_type == "application/vnd.openxmlformats-officedocument.presentationml.presentation":
|
50 |
-
prs = Presentation(io.BytesIO(file.file.read()))
|
51 |
-
text = ""
|
52 |
-
for slide in prs.slides:
|
53 |
-
for shape in slide.shapes:
|
54 |
-
if hasattr(shape, "text"):
|
55 |
-
text += shape.text
|
56 |
-
return text
|
57 |
-
|
58 |
-
# Handle TXT files
|
59 |
-
elif file.content_type == "text/plain":
|
60 |
-
return file.file.read().decode("utf-8")
|
61 |
-
|
62 |
-
# Unsupported file type
|
63 |
-
else:
|
64 |
return None
|
65 |
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
content={"error": "Unsupported file type or conversion failed."}
|
75 |
-
)
|
76 |
|
77 |
-
|
78 |
-
|
79 |
-
|
|
|
80 |
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
content={"error": "File upload failed."}
|
85 |
)
|
86 |
|
87 |
-
|
88 |
-
|
89 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
90 |
|
91 |
-
def
|
92 |
try:
|
93 |
-
#
|
|
|
|
|
|
|
94 |
response = genai_client.models.generate_content(
|
95 |
-
model="gemini-2.0-flash
|
96 |
-
contents=[
|
|
|
|
|
|
|
|
|
|
|
|
|
97 |
)
|
98 |
|
99 |
return response.text
|
100 |
except Exception as e:
|
101 |
return f"Google Gemini API error: {e}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from fastapi import FastAPI, UploadFile, File, Form, HTTPException
|
2 |
from fastapi.staticfiles import StaticFiles
|
3 |
from fastapi.responses import FileResponse, JSONResponse
|
4 |
import cloudinary
|
|
|
6 |
import requests
|
7 |
import io
|
8 |
import os
|
9 |
+
import mimetypes
|
10 |
+
import tempfile
|
11 |
+
from pathlib import Path
|
12 |
+
|
13 |
+
from google import genai
|
14 |
+
from google.genai import types
|
15 |
+
|
16 |
+
# Import required libraries for file conversion
|
17 |
import pandas as pd
|
18 |
from PyPDF2 import PdfReader
|
19 |
+
import pptx
|
20 |
import textract
|
|
|
|
|
21 |
|
22 |
# Cloudinary Config
|
23 |
cloudinary.config(
|
|
|
29 |
# Google Gemini Client
|
30 |
genai_client = genai.Client(api_key=os.getenv("GOOGLE_API_KEY"))
|
31 |
|
32 |
+
def upload_file_to_cloudinary(file_bytes: bytes, file_type: str) -> str:
|
33 |
try:
|
34 |
+
result = cloudinary.uploader.upload(
|
35 |
+
io.BytesIO(file_bytes),
|
36 |
+
resource_type="raw" if file_type == "text" else "auto"
|
37 |
+
)
|
38 |
return result.get("secure_url")
|
39 |
except Exception as e:
|
40 |
print(f"Cloudinary upload error: {e}")
|
41 |
return None
|
42 |
|
43 |
+
def upload_image(file_bytes: bytes) -> str:
|
44 |
+
try:
|
45 |
+
result = cloudinary.uploader.upload(io.BytesIO(file_bytes))
|
46 |
+
return result.get("secure_url")
|
47 |
+
except Exception as e:
|
48 |
+
print(f"Cloudinary upload error: {e}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
return None
|
50 |
|
51 |
+
def convert_to_text(file_bytes: bytes, file_type: str, filename: str) -> str:
|
52 |
+
"""Convert various file types to plain text"""
|
53 |
+
try:
|
54 |
+
# Create a temporary file to process
|
55 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=Path(filename).suffix) as temp_file:
|
56 |
+
temp_file.write(file_bytes)
|
57 |
+
temp_path = temp_file.name
|
58 |
+
|
59 |
+
text_content = ""
|
60 |
+
|
61 |
+
if file_type == "application/pdf":
|
62 |
+
# Convert PDF to text
|
63 |
+
pdf = PdfReader(temp_path)
|
64 |
+
for page in pdf.pages:
|
65 |
+
text_content += page.extract_text() + "\n"
|
66 |
+
|
67 |
+
elif file_type in ["application/vnd.ms-excel", "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"]:
|
68 |
+
# Convert Excel to text
|
69 |
+
df = pd.read_excel(temp_path)
|
70 |
+
text_content = df.to_string()
|
71 |
+
|
72 |
+
elif file_type in ["application/vnd.ms-powerpoint", "application/vnd.openxmlformats-officedocument.presentationml.presentation"]:
|
73 |
+
# Convert PowerPoint to text
|
74 |
+
ppt = pptx.Presentation(temp_path)
|
75 |
+
for slide in ppt.slides:
|
76 |
+
for shape in slide.shapes:
|
77 |
+
if hasattr(shape, "text"):
|
78 |
+
text_content += shape.text + "\n"
|
79 |
+
text_content += "\n---\n"
|
80 |
+
|
81 |
+
elif file_type == "text/plain":
|
82 |
+
# Already text, just read it
|
83 |
+
with open(temp_path, 'r', encoding='utf-8', errors='ignore') as f:
|
84 |
+
text_content = f.read()
|
85 |
+
|
86 |
+
else:
|
87 |
+
# Try using textract for other types
|
88 |
+
text_content = textract.process(temp_path).decode('utf-8', errors='ignore')
|
89 |
+
|
90 |
+
# Clean up temporary file
|
91 |
+
os.unlink(temp_path)
|
92 |
+
|
93 |
+
return text_content
|
94 |
|
95 |
+
except Exception as e:
|
96 |
+
print(f"Conversion error: {e}")
|
97 |
+
raise HTTPException(status_code=500, detail=f"File conversion failed: {str(e)}")
|
|
|
|
|
98 |
|
99 |
+
def generate_response(user_prompt: str, image_url: str) -> str:
|
100 |
+
try:
|
101 |
+
# Download image from URL
|
102 |
+
image_bytes = requests.get(image_url).content
|
103 |
|
104 |
+
# Prepare image part
|
105 |
+
image_part = types.Part.from_bytes(
|
106 |
+
data=image_bytes, mime_type="image/jpeg"
|
|
|
107 |
)
|
108 |
|
109 |
+
# Send to Gemini
|
110 |
+
response = genai_client.models.generate_content(
|
111 |
+
model="gemini-2.0-flash-exp",
|
112 |
+
contents=[user_prompt, image_part],
|
113 |
+
)
|
114 |
+
|
115 |
+
return response.text
|
116 |
+
except Exception as e:
|
117 |
+
return f"Google Gemini API error: {e}"
|
118 |
|
119 |
+
def generate_response_for_document(user_prompt: str, doc_url: str, mime_type: str) -> str:
|
120 |
try:
|
121 |
+
# Download document from URL
|
122 |
+
doc_data = requests.get(doc_url).content
|
123 |
+
|
124 |
+
# Send to Gemini
|
125 |
response = genai_client.models.generate_content(
|
126 |
+
model="gemini-2.0-flash",
|
127 |
+
contents=[
|
128 |
+
types.Part.from_bytes(
|
129 |
+
data=doc_data,
|
130 |
+
mime_type=mime_type,
|
131 |
+
),
|
132 |
+
user_prompt
|
133 |
+
]
|
134 |
)
|
135 |
|
136 |
return response.text
|
137 |
except Exception as e:
|
138 |
return f"Google Gemini API error: {e}"
|
139 |
+
|
140 |
+
app = FastAPI()
|
141 |
+
|
142 |
+
# Serve static files
|
143 |
+
app.mount("/static", StaticFiles(directory="frontend"), name="static")
|
144 |
+
|
145 |
+
@app.get("/")
|
146 |
+
async def home():
|
147 |
+
return FileResponse("frontend/index.html")
|
148 |
+
|
149 |
+
@app.get("/text-generator")
|
150 |
+
async def text_generator():
|
151 |
+
return FileResponse("frontend/text-generator.html")
|
152 |
+
|
153 |
+
@app.get("/about")
|
154 |
+
async def about():
|
155 |
+
return FileResponse("frontend/about.html")
|
156 |
+
|
157 |
+
@app.get("/features")
|
158 |
+
async def features():
|
159 |
+
return FileResponse("frontend/features.html")
|
160 |
+
|
161 |
+
@app.post("/process/")
|
162 |
+
async def process_file(file: UploadFile = File(...), user_prompt: str = Form(...)):
|
163 |
+
file_bytes = await file.read()
|
164 |
+
content_type = file.content_type or mimetypes.guess_type(file.filename)[0] or "application/octet-stream"
|
165 |
+
|
166 |
+
# Handle image files
|
167 |
+
if content_type.startswith("image/"):
|
168 |
+
image_url = upload_image(file_bytes)
|
169 |
+
|
170 |
+
if not image_url:
|
171 |
+
return JSONResponse(
|
172 |
+
status_code=500,
|
173 |
+
content={"error": "Image upload failed."}
|
174 |
+
)
|
175 |
+
|
176 |
+
result = generate_response(user_prompt, image_url)
|
177 |
+
return {"response": result}
|
178 |
+
|
179 |
+
# Handle document files (Excel, PDF, TXT, PowerPoint)
|
180 |
+
elif content_type in [
|
181 |
+
"application/pdf",
|
182 |
+
"application/vnd.ms-excel",
|
183 |
+
"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
|
184 |
+
"application/vnd.ms-powerpoint",
|
185 |
+
"application/vnd.openxmlformats-officedocument.presentationml.presentation",
|
186 |
+
"text/plain"
|
187 |
+
]:
|
188 |
+
try:
|
189 |
+
# Convert the file to text
|
190 |
+
text_content = convert_to_text(file_bytes, content_type, file.filename)
|
191 |
+
|
192 |
+
# Save text content to a file
|
193 |
+
text_filename = f"{Path(file.filename).stem}.txt"
|
194 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".txt", mode='w', encoding='utf-8') as text_file:
|
195 |
+
text_file.write(text_content)
|
196 |
+
text_path = text_file.name
|
197 |
+
|
198 |
+
# Upload text file to Cloudinary
|
199 |
+
with open(text_path, 'rb') as f:
|
200 |
+
text_url = upload_file_to_cloudinary(f.read(), "text")
|
201 |
+
|
202 |
+
# Clean up temporary text file
|
203 |
+
os.unlink(text_path)
|
204 |
+
|
205 |
+
if not text_url:
|
206 |
+
return JSONResponse(
|
207 |
+
status_code=500,
|
208 |
+
content={"error": "Text file upload failed."}
|
209 |
+
)
|
210 |
+
|
211 |
+
# Process with Gemini
|
212 |
+
result = generate_response_for_document(user_prompt, text_url, "text/plain")
|
213 |
+
|
214 |
+
return {
|
215 |
+
"response": result,
|
216 |
+
"text_content": text_content[:500] + "..." if len(text_content) > 500 else text_content,
|
217 |
+
"text_url": text_url
|
218 |
+
}
|
219 |
+
|
220 |
+
except Exception as e:
|
221 |
+
return JSONResponse(
|
222 |
+
status_code=500,
|
223 |
+
content={"error": f"Processing failed: {str(e)}"}
|
224 |
+
)
|
225 |
+
|
226 |
+
else:
|
227 |
+
return JSONResponse(
|
228 |
+
status_code=400,
|
229 |
+
content={"error": "Unsupported file type. Please upload an image, PDF, Excel, PowerPoint, or text file."}
|
230 |
+
)
|