Jeremy Live
commited on
Commit
·
107a3f2
1
Parent(s):
d435e17
Revert "API solved v2"
Browse filesThis reverts commit 56fb484b2a4d4edf150659764c8a3c1f6022bf8e.
app.py
CHANGED
@@ -14,81 +14,6 @@ import plotly.graph_objects as go
|
|
14 |
from plotly.subplots import make_subplots
|
15 |
from shared import initialize_llm, setup_database_connection, create_agent
|
16 |
|
17 |
-
try:
|
18 |
-
from langchain_core.messages import HumanMessage, AIMessage
|
19 |
-
LANGCHAIN_AVAILABLE = True
|
20 |
-
except ImportError:
|
21 |
-
# Fallback if langchain not available
|
22 |
-
class HumanMessage:
|
23 |
-
def __init__(self, content):
|
24 |
-
self.content = content
|
25 |
-
|
26 |
-
class AIMessage:
|
27 |
-
def __init__(self, content):
|
28 |
-
self.content = content
|
29 |
-
LANGCHAIN_AVAILABLE = False
|
30 |
-
|
31 |
-
# Configure logging
|
32 |
-
logging.basicConfig(level=logging.INFO)
|
33 |
-
logger = logging.getLogger(__name__)
|
34 |
-
|
35 |
-
def create_ui():
|
36 |
-
"""Create the Gradio UI components."""
|
37 |
-
# Custom CSS for styling
|
38 |
-
custom_css = """
|
39 |
-
.gradio-container {
|
40 |
-
max-width: 1200px !important;
|
41 |
-
}
|
42 |
-
.chat-container {
|
43 |
-
height: 600px;
|
44 |
-
overflow-y: auto;
|
45 |
-
}
|
46 |
-
.chart-container {
|
47 |
-
height: 600px;
|
48 |
-
overflow-y: auto;
|
49 |
-
}
|
50 |
-
"""
|
51 |
-
|
52 |
-
with gr.Blocks(css=custom_css, title="🤖 SQL Database Assistant") as demo:
|
53 |
-
gr.Markdown("# 🤖 SQL Database Assistant")
|
54 |
-
gr.Markdown("Ask questions about your database in natural language!")
|
55 |
-
|
56 |
-
with gr.Row():
|
57 |
-
with gr.Column(scale=2):
|
58 |
-
chatbot = gr.Chatbot(
|
59 |
-
label="Chat",
|
60 |
-
elem_classes="chat-container",
|
61 |
-
type="messages",
|
62 |
-
height=500
|
63 |
-
)
|
64 |
-
|
65 |
-
with gr.Row():
|
66 |
-
question_input = gr.Textbox(
|
67 |
-
label="Ask your question",
|
68 |
-
placeholder="Type your question here...",
|
69 |
-
lines=2,
|
70 |
-
scale=4
|
71 |
-
)
|
72 |
-
submit_button = gr.Button("Send", variant="primary", scale=1)
|
73 |
-
|
74 |
-
streaming_output_display = gr.Markdown(visible=False)
|
75 |
-
|
76 |
-
with gr.Column(scale=1):
|
77 |
-
chart_display = gr.Plot(
|
78 |
-
label="Charts",
|
79 |
-
elem_classes="chart-container",
|
80 |
-
height=500
|
81 |
-
)
|
82 |
-
|
83 |
-
# Status indicators
|
84 |
-
with gr.Row():
|
85 |
-
status_indicator = gr.Markdown(
|
86 |
-
"### ✅ System Status\n- **Database**: Ready\n- **AI Model**: Ready\n- **API**: Available",
|
87 |
-
elem_id="status"
|
88 |
-
)
|
89 |
-
|
90 |
-
return demo, chatbot, chart_display, question_input, submit_button, streaming_output_display
|
91 |
-
|
92 |
# ... (resto del código existente sin cambios) ...
|
93 |
|
94 |
def create_application():
|
@@ -196,112 +121,6 @@ def create_application():
|
|
196 |
|
197 |
return demo
|
198 |
|
199 |
-
async def stream_agent_response(question: str, chat_history: List[List[str]]) -> Tuple[str, Optional[go.Figure]]:
|
200 |
-
"""Process a question through the SQL agent and return response with optional chart."""
|
201 |
-
|
202 |
-
# Initialize components
|
203 |
-
llm, llm_error = initialize_llm()
|
204 |
-
if llm_error:
|
205 |
-
return f"**LLM Error:** {llm_error}", None
|
206 |
-
|
207 |
-
db_connection, db_error = setup_database_connection()
|
208 |
-
if db_error:
|
209 |
-
return f"**Database Error:** {db_error}", None
|
210 |
-
|
211 |
-
agent, agent_error = create_agent(llm, db_connection)
|
212 |
-
if agent_error:
|
213 |
-
return f"**Agent Error:** {agent_error}", None
|
214 |
-
|
215 |
-
try:
|
216 |
-
logger.info(f"Processing question: {question}")
|
217 |
-
|
218 |
-
# Prepare the input with chat history
|
219 |
-
input_data = {"input": question}
|
220 |
-
if chat_history:
|
221 |
-
# Format chat history for the agent
|
222 |
-
formatted_history = []
|
223 |
-
for human, ai in chat_history:
|
224 |
-
formatted_history.extend([
|
225 |
-
HumanMessage(content=human),
|
226 |
-
AIMessage(content=ai)
|
227 |
-
])
|
228 |
-
input_data["chat_history"] = formatted_history
|
229 |
-
|
230 |
-
# Execute the agent
|
231 |
-
response = agent.invoke(input_data)
|
232 |
-
|
233 |
-
# Extract the response text
|
234 |
-
if hasattr(response, 'output') and response.output:
|
235 |
-
response_text = response.output
|
236 |
-
elif isinstance(response, dict) and 'output' in response:
|
237 |
-
response_text = response['output']
|
238 |
-
elif isinstance(response, str):
|
239 |
-
response_text = response
|
240 |
-
else:
|
241 |
-
response_text = str(response)
|
242 |
-
|
243 |
-
# Check for SQL queries in the response
|
244 |
-
sql_pattern = r'```sql\s*(.*?)\s*```'
|
245 |
-
sql_matches = re.findall(sql_pattern, response_text, re.DOTALL)
|
246 |
-
|
247 |
-
chart_fig = None
|
248 |
-
if sql_matches:
|
249 |
-
# Try to execute the SQL and create a chart
|
250 |
-
try:
|
251 |
-
sql_query = sql_matches[-1].strip()
|
252 |
-
logger.info(f"Executing SQL query: {sql_query}")
|
253 |
-
|
254 |
-
# Execute the query
|
255 |
-
result = db_connection.run(sql_query)
|
256 |
-
|
257 |
-
if result:
|
258 |
-
# Convert result to DataFrame
|
259 |
-
import pandas as pd
|
260 |
-
if isinstance(result, list) and result:
|
261 |
-
df = pd.DataFrame(result)
|
262 |
-
|
263 |
-
# Determine chart type based on data
|
264 |
-
if len(df.columns) >= 2:
|
265 |
-
# Simple bar chart for categorical data
|
266 |
-
fig = go.Figure()
|
267 |
-
|
268 |
-
if len(df) <= 20: # Bar chart for smaller datasets
|
269 |
-
fig.add_trace(go.Bar(
|
270 |
-
x=df.iloc[:, 0],
|
271 |
-
y=df.iloc[:, 1],
|
272 |
-
name=str(df.columns[1])
|
273 |
-
))
|
274 |
-
fig.update_layout(
|
275 |
-
title=f"{df.columns[0]} vs {df.columns[1]}",
|
276 |
-
xaxis_title=str(df.columns[0]),
|
277 |
-
yaxis_title=str(df.columns[1])
|
278 |
-
)
|
279 |
-
else: # Line chart for larger datasets
|
280 |
-
fig.add_trace(go.Scatter(
|
281 |
-
x=df.iloc[:, 0],
|
282 |
-
y=df.iloc[:, 1],
|
283 |
-
mode='lines+markers',
|
284 |
-
name=str(df.columns[1])
|
285 |
-
))
|
286 |
-
fig.update_layout(
|
287 |
-
title=f"{df.columns[0]} vs {df.columns[1]}",
|
288 |
-
xaxis_title=str(df.columns[0]),
|
289 |
-
yaxis_title=str(df.columns[1])
|
290 |
-
)
|
291 |
-
|
292 |
-
chart_fig = fig
|
293 |
-
|
294 |
-
except Exception as e:
|
295 |
-
logger.warning(f"Could not create chart: {e}")
|
296 |
-
# Continue without chart
|
297 |
-
|
298 |
-
return response_text, chart_fig
|
299 |
-
|
300 |
-
except Exception as e:
|
301 |
-
error_msg = f"**Error processing question:** {str(e)}"
|
302 |
-
logger.error(error_msg, exc_info=True)
|
303 |
-
return error_msg, None
|
304 |
-
|
305 |
# Create the application
|
306 |
demo = create_application()
|
307 |
|
|
|
14 |
from plotly.subplots import make_subplots
|
15 |
from shared import initialize_llm, setup_database_connection, create_agent
|
16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
# ... (resto del código existente sin cambios) ...
|
18 |
|
19 |
def create_application():
|
|
|
121 |
|
122 |
return demo
|
123 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
124 |
# Create the application
|
125 |
demo = create_application()
|
126 |
|