Ali Abdullah
commited on
Update app.py
Browse files
app.py
CHANGED
|
@@ -2,20 +2,27 @@ import gradio as gr
|
|
| 2 |
import requests
|
| 3 |
import mimetypes
|
| 4 |
import os
|
|
|
|
|
|
|
| 5 |
import re
|
| 6 |
|
| 7 |
-
# === Fix surrogate character issue ===
|
| 8 |
def sanitize(text):
|
| 9 |
if isinstance(text, str):
|
| 10 |
text = re.sub(r'[\uD800-\uDFFF]', '', text)
|
| 11 |
return text.encode("utf-8", "ignore").decode("utf-8")
|
| 12 |
return text
|
| 13 |
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
chat_sessions = {"file": [], "url": []}
|
| 21 |
|
|
@@ -24,137 +31,86 @@ def format_chat(history):
|
|
| 24 |
|
| 25 |
def ask_from_file(file_obj, question, session_id):
|
| 26 |
if not file_obj or not question.strip():
|
| 27 |
-
return "⚠️
|
| 28 |
if os.path.getsize(file_obj.name) > 10 * 1024 * 1024:
|
| 29 |
-
return "❌ File too large.
|
| 30 |
-
|
| 31 |
try:
|
| 32 |
-
|
| 33 |
with open(file_obj.name, "rb") as f:
|
| 34 |
-
files = {"file": (file_obj.name, f,
|
| 35 |
history = chat_sessions["file"]
|
| 36 |
context = "\n".join([f"Q: {q}\nA: {a}" for q, a in history]) + f"\nQ: {question}"
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
answer = sanitize(response.json().get("answer", "⚠️ No answer returned."))
|
| 40 |
chat_sessions["file"].append((question, answer))
|
| 41 |
return format_chat(chat_sessions["file"]), chat_sessions["file"]
|
| 42 |
except Exception as e:
|
| 43 |
-
return f"❌ Error: {
|
| 44 |
|
| 45 |
def ask_from_url(url, question, session_id):
|
| 46 |
if not url.strip() or not question.strip():
|
| 47 |
-
return "⚠️
|
| 48 |
try:
|
| 49 |
history = chat_sessions["url"]
|
| 50 |
context = "\n".join([f"Q: {q}\nA: {a}" for q, a in history]) + f"\nQ: {question}"
|
| 51 |
-
|
| 52 |
-
answer = sanitize(
|
| 53 |
chat_sessions["url"].append((question, answer))
|
| 54 |
return format_chat(chat_sessions["url"]), chat_sessions["url"]
|
| 55 |
except Exception as e:
|
| 56 |
-
return f"❌ Error: {
|
| 57 |
|
| 58 |
def extract_text_from_image(image_path):
|
| 59 |
try:
|
| 60 |
with open(image_path, "rb") as f:
|
| 61 |
files = {"file": ("image.png", f, "image/png")}
|
| 62 |
-
|
| 63 |
-
return sanitize(
|
| 64 |
except Exception as e:
|
| 65 |
-
return f"❌ Error: {
|
| 66 |
|
| 67 |
def transcribe_audio(audio_path):
|
| 68 |
try:
|
| 69 |
with open(audio_path, "rb") as f:
|
| 70 |
files = {"file": ("audio.wav", f, "audio/wav")}
|
| 71 |
-
|
| 72 |
-
return sanitize(
|
| 73 |
except Exception as e:
|
| 74 |
-
return f"❌ Error: {
|
| 75 |
-
|
| 76 |
-
def clear_file_chat():
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
.
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
color: white !important;
|
| 111 |
-
border: 1px solid #00FF88 !important;
|
| 112 |
-
border-radius: 6px !important;
|
| 113 |
-
}
|
| 114 |
-
.gr-tabitem[data-selected="true"] > button {
|
| 115 |
-
background-color: #00FF88 !important;
|
| 116 |
-
color: black !important;
|
| 117 |
-
}
|
| 118 |
-
.gr-tabitem > button {
|
| 119 |
-
background-color: transparent !important;
|
| 120 |
-
color: white !important;
|
| 121 |
-
}
|
| 122 |
-
"""
|
| 123 |
-
|
| 124 |
-
# UI
|
| 125 |
-
with gr.Blocks(css=custom_css) as demo:
|
| 126 |
-
gr.Markdown("# 🤖 AI Chatbot with File, Web, Image, Audio & Chat History")
|
| 127 |
-
|
| 128 |
-
with gr.Tab("📂 Chat with File"):
|
| 129 |
-
file_input = gr.File(label="Upload File (.txt, .csv, .docx, .pdf)", file_types=[".txt", ".csv", ".docx", ".pdf"])
|
| 130 |
-
file_question = gr.Textbox(label="Your Question", placeholder="Ask something based on file")
|
| 131 |
-
file_button = gr.Button("Ask from File")
|
| 132 |
-
file_output = gr.Textbox(label="Chat History", lines=10)
|
| 133 |
-
clear_file = gr.Button("New File Chat")
|
| 134 |
-
|
| 135 |
-
file_button.click(fn=ask_from_file, inputs=[file_input, file_question, gr.State("file")], outputs=[file_output, gr.State("file")])
|
| 136 |
-
clear_file.click(fn=clear_file_chat, inputs=[], outputs=[file_output, gr.State("file")])
|
| 137 |
-
|
| 138 |
-
with gr.Tab("🌐 Chat with Website"):
|
| 139 |
-
url_input = gr.Textbox(label="Website URL", placeholder="https://example.com")
|
| 140 |
-
url_question = gr.Textbox(label="Your Question", placeholder="Ask something based on webpage")
|
| 141 |
-
url_button = gr.Button("Ask from URL")
|
| 142 |
-
url_output = gr.Textbox(label="Chat History", lines=10)
|
| 143 |
-
clear_url = gr.Button("New URL Chat")
|
| 144 |
-
|
| 145 |
-
url_button.click(fn=ask_from_url, inputs=[url_input, url_question, gr.State("url")], outputs=[url_output, gr.State("url")])
|
| 146 |
-
clear_url.click(fn=clear_url_chat, inputs=[], outputs=[url_output, gr.State("url")])
|
| 147 |
-
|
| 148 |
-
with gr.Tab("🖼️ Extract Text from Image"):
|
| 149 |
-
image_input = gr.Image(label="Upload Image", type="filepath")
|
| 150 |
-
image_button = gr.Button("Extract Text")
|
| 151 |
-
image_output = gr.Textbox(label="Extracted Text", lines=8)
|
| 152 |
-
image_button.click(fn=extract_text_from_image, inputs=image_input, outputs=image_output)
|
| 153 |
-
|
| 154 |
-
with gr.Tab("🎤 Transcribe Audio"):
|
| 155 |
-
audio_input = gr.Audio(label="Upload Audio", type="filepath")
|
| 156 |
-
audio_button = gr.Button("Transcribe Audio")
|
| 157 |
-
audio_output = gr.Textbox(label="Transcript", lines=8)
|
| 158 |
-
audio_button.click(fn=transcribe_audio, inputs=audio_input, outputs=audio_output)
|
| 159 |
|
| 160 |
demo.launch(share=False)
|
|
|
|
| 2 |
import requests
|
| 3 |
import mimetypes
|
| 4 |
import os
|
| 5 |
+
import threading
|
| 6 |
+
import uvicorn
|
| 7 |
import re
|
| 8 |
|
|
|
|
| 9 |
def sanitize(text):
|
| 10 |
if isinstance(text, str):
|
| 11 |
text = re.sub(r'[\uD800-\uDFFF]', '', text)
|
| 12 |
return text.encode("utf-8", "ignore").decode("utf-8")
|
| 13 |
return text
|
| 14 |
|
| 15 |
+
def run_fastapi():
|
| 16 |
+
uvicorn.run("main:app", host="0.0.0.0", port=8000)
|
| 17 |
+
|
| 18 |
+
threading.Thread(target=run_fastapi, daemon=True).start()
|
| 19 |
+
|
| 20 |
+
API_URLS = {
|
| 21 |
+
"file": "http://127.0.0.1:8000/chat-with-file",
|
| 22 |
+
"url": "http://127.0.0.1:8000/chat-with-url",
|
| 23 |
+
"image": "http://127.0.0.1:8000/extract-text-from-image",
|
| 24 |
+
"audio": "http://127.0.0.1:8000/transcribe-audio"
|
| 25 |
+
}
|
| 26 |
|
| 27 |
chat_sessions = {"file": [], "url": []}
|
| 28 |
|
|
|
|
| 31 |
|
| 32 |
def ask_from_file(file_obj, question, session_id):
|
| 33 |
if not file_obj or not question.strip():
|
| 34 |
+
return "⚠️ Upload a file and enter a question.", chat_sessions["file"]
|
| 35 |
if os.path.getsize(file_obj.name) > 10 * 1024 * 1024:
|
| 36 |
+
return "❌ File too large. Max 10MB.", chat_sessions["file"]
|
|
|
|
| 37 |
try:
|
| 38 |
+
mime, _ = mimetypes.guess_type(file_obj.name)
|
| 39 |
with open(file_obj.name, "rb") as f:
|
| 40 |
+
files = {"file": (file_obj.name, f, mime or "application/octet-stream")}
|
| 41 |
history = chat_sessions["file"]
|
| 42 |
context = "\n".join([f"Q: {q}\nA: {a}" for q, a in history]) + f"\nQ: {question}"
|
| 43 |
+
res = requests.post(API_URLS["file"], files=files, data={"question": context})
|
| 44 |
+
answer = sanitize(res.json().get("answer", "⚠️ No answer."))
|
|
|
|
| 45 |
chat_sessions["file"].append((question, answer))
|
| 46 |
return format_chat(chat_sessions["file"]), chat_sessions["file"]
|
| 47 |
except Exception as e:
|
| 48 |
+
return f"❌ Error: {e}", chat_sessions["file"]
|
| 49 |
|
| 50 |
def ask_from_url(url, question, session_id):
|
| 51 |
if not url.strip() or not question.strip():
|
| 52 |
+
return "⚠️ URL and question required.", chat_sessions["url"]
|
| 53 |
try:
|
| 54 |
history = chat_sessions["url"]
|
| 55 |
context = "\n".join([f"Q: {q}\nA: {a}" for q, a in history]) + f"\nQ: {question}"
|
| 56 |
+
res = requests.post(API_URLS["url"], json={"url": url, "question": context})
|
| 57 |
+
answer = sanitize(res.json().get("answer", "⚠️ No answer."))
|
| 58 |
chat_sessions["url"].append((question, answer))
|
| 59 |
return format_chat(chat_sessions["url"]), chat_sessions["url"]
|
| 60 |
except Exception as e:
|
| 61 |
+
return f"❌ Error: {e}", chat_sessions["url"]
|
| 62 |
|
| 63 |
def extract_text_from_image(image_path):
|
| 64 |
try:
|
| 65 |
with open(image_path, "rb") as f:
|
| 66 |
files = {"file": ("image.png", f, "image/png")}
|
| 67 |
+
res = requests.post(API_URLS["image"], files=files)
|
| 68 |
+
return sanitize(res.json().get("answer", "⚠️ No text extracted."))
|
| 69 |
except Exception as e:
|
| 70 |
+
return f"❌ Error: {e}"
|
| 71 |
|
| 72 |
def transcribe_audio(audio_path):
|
| 73 |
try:
|
| 74 |
with open(audio_path, "rb") as f:
|
| 75 |
files = {"file": ("audio.wav", f, "audio/wav")}
|
| 76 |
+
res = requests.post(API_URLS["audio"], files=files)
|
| 77 |
+
return sanitize(res.json().get("answer", "⚠️ No transcript returned."))
|
| 78 |
except Exception as e:
|
| 79 |
+
return f"❌ Error: {e}"
|
| 80 |
+
|
| 81 |
+
def clear_file_chat(): chat_sessions["file"] = []; return "", chat_sessions["file"]
|
| 82 |
+
def clear_url_chat(): chat_sessions["url"] = []; return "", chat_sessions["url"]
|
| 83 |
+
|
| 84 |
+
with gr.Blocks(css="") as demo:
|
| 85 |
+
gr.Markdown("# 🤖 AI Chatbot with File, Web, Image, Audio")
|
| 86 |
+
with gr.Tab("📂 File"):
|
| 87 |
+
file = gr.File(file_types=[".txt", ".csv", ".docx", ".pdf"])
|
| 88 |
+
q1 = gr.Textbox(label="Question")
|
| 89 |
+
b1 = gr.Button("Ask from File")
|
| 90 |
+
out1 = gr.Textbox(label="Chat", lines=10)
|
| 91 |
+
clr1 = gr.Button("Clear")
|
| 92 |
+
b1.click(fn=ask_from_file, inputs=[file, q1, gr.State("file")], outputs=[out1, gr.State("file")])
|
| 93 |
+
clr1.click(fn=clear_file_chat, outputs=[out1, gr.State("file")])
|
| 94 |
+
|
| 95 |
+
with gr.Tab("🌐 URL"):
|
| 96 |
+
url = gr.Textbox(label="Website URL")
|
| 97 |
+
q2 = gr.Textbox(label="Question")
|
| 98 |
+
b2 = gr.Button("Ask from URL")
|
| 99 |
+
out2 = gr.Textbox(label="Chat", lines=10)
|
| 100 |
+
clr2 = gr.Button("Clear")
|
| 101 |
+
b2.click(fn=ask_from_url, inputs=[url, q2, gr.State("url")], outputs=[out2, gr.State("url")])
|
| 102 |
+
clr2.click(fn=clear_url_chat, outputs=[out2, gr.State("url")])
|
| 103 |
+
|
| 104 |
+
with gr.Tab("🖼️ Image OCR"):
|
| 105 |
+
img = gr.Image(type="filepath")
|
| 106 |
+
b3 = gr.Button("Extract Text")
|
| 107 |
+
out3 = gr.Textbox(label="Text", lines=8)
|
| 108 |
+
b3.click(fn=extract_text_from_image, inputs=img, outputs=out3)
|
| 109 |
+
|
| 110 |
+
with gr.Tab("🎤 Audio"):
|
| 111 |
+
aud = gr.Audio(type="filepath")
|
| 112 |
+
b4 = gr.Button("Transcribe")
|
| 113 |
+
out4 = gr.Textbox(label="Transcript", lines=8)
|
| 114 |
+
b4.click(fn=transcribe_audio, inputs=aud, outputs=out4)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 115 |
|
| 116 |
demo.launch(share=False)
|