Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -8,7 +8,6 @@ from huggingface_hub import login
|
|
| 8 |
# Space configuration
|
| 9 |
SPACE_DIR = os.environ.get("HF_HOME", os.getcwd())
|
| 10 |
|
| 11 |
-
|
| 12 |
def init_huggingface_auth():
|
| 13 |
"""Space-friendly authentication"""
|
| 14 |
token = os.getenv("HUGGINGFACE_TOKEN")
|
|
@@ -28,15 +27,12 @@ def init_huggingface_auth():
|
|
| 28 |
if not init_huggingface_auth():
|
| 29 |
print("Warning: Authentication failed")
|
| 30 |
|
| 31 |
-
|
| 32 |
# Load and preprocess the PDF content
|
| 33 |
pdf_path = os.path.join(SPACE_DIR, "LTDOCS.pdf")
|
| 34 |
with open(pdf_path, 'rb') as file:
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
|
| 41 |
# Initialize the pipeline
|
| 42 |
pipe = pipeline(
|
|
@@ -46,7 +42,7 @@ pipe = pipeline(
|
|
| 46 |
device="cpu", # replace with "mps" to run on a Mac device
|
| 47 |
)
|
| 48 |
|
| 49 |
-
# System prompt and welcome message
|
| 50 |
SYSTEM_PROMPT = f"""You Foton the chat bot assistant of the Company Lugha taussi, an AI language assistant specialized in African languages, with a focus on Swahili. Your primary tasks are:
|
| 51 |
1. Providing accurate translations between Swahili and other languages
|
| 52 |
2. Teaching Swahili vocabulary and grammar
|
|
@@ -59,10 +55,39 @@ Always maintain a friendly and patient demeanor, and provide cultural context wh
|
|
| 59 |
|
| 60 |
WELCOME_MESSAGE = "**Karibu Lugha Tausi!** Mimi ni Foton, msaidizi wako wa kibinafsi wa Kiswahili. Niko hapa kukusaidia kujifunza, kuelewa, na kuzungumza Kiswahili. **Ninaweza kukusaidiaje leo?** Hebu tuanze! 😊"
|
| 61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
def format_chat_message(messages, system_prompt=SYSTEM_PROMPT):
|
| 63 |
"""Format the chat messages with system prompt"""
|
| 64 |
formatted_prompt = f"{system_prompt}\n\n"
|
| 65 |
-
|
| 66 |
for message in messages:
|
| 67 |
if isinstance(message, tuple):
|
| 68 |
role, content = message
|
|
@@ -70,12 +95,11 @@ def format_chat_message(messages, system_prompt=SYSTEM_PROMPT):
|
|
| 70 |
formatted_prompt += f"User: {content}\nLugha Tausi: "
|
| 71 |
elif role == "assistant":
|
| 72 |
formatted_prompt += f"{content}\n"
|
| 73 |
-
|
| 74 |
return formatted_prompt
|
| 75 |
|
| 76 |
def chat_response(message, history):
|
| 77 |
"""Generate response for Gradio chat interface"""
|
| 78 |
-
# Convert history to the format expected by the model
|
| 79 |
messages = []
|
| 80 |
for user_msg, bot_msg in history:
|
| 81 |
messages.append(("user", user_msg))
|
|
@@ -93,18 +117,35 @@ def chat_response(message, history):
|
|
| 93 |
)
|
| 94 |
return outputs[0]["generated_text"].strip()
|
| 95 |
|
| 96 |
-
# Create Gradio interface
|
| 97 |
-
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
|
| 109 |
def user_input(message, history):
|
| 110 |
return "", history + [(message, None)]
|
|
@@ -143,5 +184,4 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
| 143 |
|
| 144 |
# Launch the interface
|
| 145 |
if __name__ == "__main__":
|
| 146 |
-
demo.launch(share=True,ssr_mode=False
|
| 147 |
-
|
|
|
|
| 8 |
# Space configuration
|
| 9 |
SPACE_DIR = os.environ.get("HF_HOME", os.getcwd())
|
| 10 |
|
|
|
|
| 11 |
def init_huggingface_auth():
|
| 12 |
"""Space-friendly authentication"""
|
| 13 |
token = os.getenv("HUGGINGFACE_TOKEN")
|
|
|
|
| 27 |
if not init_huggingface_auth():
|
| 28 |
print("Warning: Authentication failed")
|
| 29 |
|
|
|
|
| 30 |
# Load and preprocess the PDF content
|
| 31 |
pdf_path = os.path.join(SPACE_DIR, "LTDOCS.pdf")
|
| 32 |
with open(pdf_path, 'rb') as file:
|
| 33 |
+
pdf_reader = PyPDF2.PdfReader(file)
|
| 34 |
+
pdf_content = ' '.join([page.extract_text() for page in pdf_reader.pages])
|
| 35 |
+
pdf_content = pdf_content.lower().strip()
|
|
|
|
|
|
|
| 36 |
|
| 37 |
# Initialize the pipeline
|
| 38 |
pipe = pipeline(
|
|
|
|
| 42 |
device="cpu", # replace with "mps" to run on a Mac device
|
| 43 |
)
|
| 44 |
|
| 45 |
+
# System prompt and welcome message
|
| 46 |
SYSTEM_PROMPT = f"""You Foton the chat bot assistant of the Company Lugha taussi, an AI language assistant specialized in African languages, with a focus on Swahili. Your primary tasks are:
|
| 47 |
1. Providing accurate translations between Swahili and other languages
|
| 48 |
2. Teaching Swahili vocabulary and grammar
|
|
|
|
| 55 |
|
| 56 |
WELCOME_MESSAGE = "**Karibu Lugha Tausi!** Mimi ni Foton, msaidizi wako wa kibinafsi wa Kiswahili. Niko hapa kukusaidia kujifunza, kuelewa, na kuzungumza Kiswahili. **Ninaweza kukusaidiaje leo?** Hebu tuanze! 😊"
|
| 57 |
|
| 58 |
+
# CSS for custom styling
|
| 59 |
+
CUSTOM_CSS = """
|
| 60 |
+
.container {
|
| 61 |
+
max-width: 800px;
|
| 62 |
+
margin: auto;
|
| 63 |
+
padding: 20px;
|
| 64 |
+
}
|
| 65 |
+
.header {
|
| 66 |
+
text-align: center;
|
| 67 |
+
margin-bottom: 30px;
|
| 68 |
+
}
|
| 69 |
+
.icon {
|
| 70 |
+
width: 80px;
|
| 71 |
+
height: 80px;
|
| 72 |
+
margin: 0 auto 15px;
|
| 73 |
+
display: block;
|
| 74 |
+
}
|
| 75 |
+
.title {
|
| 76 |
+
font-size: 2.5em;
|
| 77 |
+
font-weight: bold;
|
| 78 |
+
margin-bottom: 10px;
|
| 79 |
+
}
|
| 80 |
+
.description {
|
| 81 |
+
font-size: 1.2em;
|
| 82 |
+
color: #666;
|
| 83 |
+
margin-bottom: 20px;
|
| 84 |
+
}
|
| 85 |
+
"""
|
| 86 |
+
|
| 87 |
def format_chat_message(messages, system_prompt=SYSTEM_PROMPT):
|
| 88 |
"""Format the chat messages with system prompt"""
|
| 89 |
formatted_prompt = f"{system_prompt}\n\n"
|
| 90 |
+
|
| 91 |
for message in messages:
|
| 92 |
if isinstance(message, tuple):
|
| 93 |
role, content = message
|
|
|
|
| 95 |
formatted_prompt += f"User: {content}\nLugha Tausi: "
|
| 96 |
elif role == "assistant":
|
| 97 |
formatted_prompt += f"{content}\n"
|
| 98 |
+
|
| 99 |
return formatted_prompt
|
| 100 |
|
| 101 |
def chat_response(message, history):
|
| 102 |
"""Generate response for Gradio chat interface"""
|
|
|
|
| 103 |
messages = []
|
| 104 |
for user_msg, bot_msg in history:
|
| 105 |
messages.append(("user", user_msg))
|
|
|
|
| 117 |
)
|
| 118 |
return outputs[0]["generated_text"].strip()
|
| 119 |
|
| 120 |
+
# Create Gradio interface with custom theme and styling
|
| 121 |
+
with gr.Blocks(theme=gr.themes.Soft(), css=CUSTOM_CSS) as demo:
|
| 122 |
+
with gr.Column(elem_classes="container"):
|
| 123 |
+
# Header section with icon, title and description
|
| 124 |
+
with gr.Column(elem_classes="header"):
|
| 125 |
+
gr.Image("/api/placeholder/80/80", elem_classes="icon", label="Foton Icon")
|
| 126 |
+
gr.Markdown("# Foton - Lugha Tausi Assistant", elem_classes="title")
|
| 127 |
+
gr.Markdown(
|
| 128 |
+
"""Your personal Swahili language assistant, powered by AI.
|
| 129 |
+
Specialized in translations, teaching, and cultural context.
|
| 130 |
+
Let's explore the beauty of Swahili together! 🌍✨""",
|
| 131 |
+
elem_classes="description"
|
| 132 |
+
)
|
| 133 |
+
|
| 134 |
+
# Chat interface
|
| 135 |
+
chatbot = gr.Chatbot(
|
| 136 |
+
value=[(None, WELCOME_MESSAGE)],
|
| 137 |
+
height=500,
|
| 138 |
+
show_label=False,
|
| 139 |
+
elem_classes="chatbox"
|
| 140 |
+
)
|
| 141 |
+
|
| 142 |
+
with gr.Row():
|
| 143 |
+
msg = gr.Textbox(
|
| 144 |
+
placeholder="Type your message here...",
|
| 145 |
+
show_label=False,
|
| 146 |
+
scale=9
|
| 147 |
+
)
|
| 148 |
+
clear = gr.Button("Clear Chat", scale=1)
|
| 149 |
|
| 150 |
def user_input(message, history):
|
| 151 |
return "", history + [(message, None)]
|
|
|
|
| 184 |
|
| 185 |
# Launch the interface
|
| 186 |
if __name__ == "__main__":
|
| 187 |
+
demo.launch(share=True, ssr_mode=False)
|
|
|