Spaces:
Sleeping
Sleeping
Delete app.py
Browse files
app.py
DELETED
@@ -1,81 +0,0 @@
|
|
1 |
-
from gradio import Interface, File
|
2 |
-
import requests
|
3 |
-
import io
|
4 |
-
import base64
|
5 |
-
import json
|
6 |
-
|
7 |
-
# Replace with your Google AI Studio API key
|
8 |
-
API_KEY = "GOOGLE_API_KEY"
|
9 |
-
project_id= "genai-test-396922"
|
10 |
-
location= "us-central1"
|
11 |
-
model_id= "gemini-2.0-flash-exp"
|
12 |
-
# Endpoint for Gemini 2.0 Text Generation (adapt if needed for specific multimodal endpoints)
|
13 |
-
ENDPOINT = "https://aiplatform.googleapis.com/v1/projects/{project_id}/locations/{location}/models/{model_id}:textGenerate"
|
14 |
-
|
15 |
-
def encode_file(file_obj, file_type):
|
16 |
-
"""Encodes a file (image, video, or PDF) to base64."""
|
17 |
-
try:
|
18 |
-
if file_type == "pdf":
|
19 |
-
content = file_obj.read()
|
20 |
-
else:
|
21 |
-
content = file_obj.getvalue() # For images and videos
|
22 |
-
encoded_content = base64.b64encode(content).decode("utf-8")
|
23 |
-
return encoded_content
|
24 |
-
except Exception as e:
|
25 |
-
print(f"Error encoding file: {e}")
|
26 |
-
return None
|
27 |
-
|
28 |
-
def generate_response(text_prompt, files):
|
29 |
-
"""Handles multiple file inputs (images, videos, PDFs)."""
|
30 |
-
url = ENDPOINT.format(project_id=project_id, location=location, model_id=model_id)
|
31 |
-
headers = {"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"}
|
32 |
-
|
33 |
-
inputs = [{"text": text_prompt}]
|
34 |
-
for file_data in files:
|
35 |
-
if file_data is None:
|
36 |
-
continue
|
37 |
-
|
38 |
-
file_path = file_data.name
|
39 |
-
file_type = file_path.split(".")[-1].lower()
|
40 |
-
|
41 |
-
try:
|
42 |
-
with open(file_path, "rb") as file_obj:
|
43 |
-
encoded_content = encode_file(file_obj, file_type)
|
44 |
-
if encoded_content:
|
45 |
-
inputs.append({file_type: encoded_content}) # Use file type as key
|
46 |
-
except Exception as e:
|
47 |
-
return f"Error processing file {file_path}: {e}"
|
48 |
-
|
49 |
-
request_body = {"inputs": inputs}
|
50 |
-
|
51 |
-
try:
|
52 |
-
response = requests.post(url, headers=headers, json=request_body)
|
53 |
-
response.raise_for_status() # Raise an exception for bad status codes (4xx or 5xx)
|
54 |
-
return response.json().get("generated_text", "No generated text found in response.") # Handle missing key
|
55 |
-
|
56 |
-
except requests.exceptions.RequestException as e:
|
57 |
-
if response is not None:
|
58 |
-
try:
|
59 |
-
error_message = response.json().get("error", {}).get("message", "Unknown error")
|
60 |
-
except json.JSONDecodeError:
|
61 |
-
error_message = response.text
|
62 |
-
return f"API Error: {response.status_code} - {error_message}"
|
63 |
-
else:
|
64 |
-
return f"Request Exception: {e}"
|
65 |
-
except Exception as e:
|
66 |
-
return f"An unexpected error occurred: {e}"
|
67 |
-
|
68 |
-
|
69 |
-
# Define Gradio interface with multiple file input
|
70 |
-
interface = Interface(
|
71 |
-
fn=generate_response,
|
72 |
-
inputs=[
|
73 |
-
"text",
|
74 |
-
[File(label="Upload Files (Images, Videos, PDFs)", file_types=[".png", ".jpg", ".jpeg", ".mp4", ".mov", ".pdf"], file_count="multiple")], # Multiple files
|
75 |
-
],
|
76 |
-
outputs="text",
|
77 |
-
title="Multimodal Prompt with Gemini 2.0 (Experimental)",
|
78 |
-
description="Enter a text prompt and upload one or more files (images, videos, PDFs).",
|
79 |
-
)
|
80 |
-
|
81 |
-
interface.launch(share=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|