Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -2,6 +2,7 @@ from groq import Groq
|
|
| 2 |
from pydantic import BaseModel
|
| 3 |
import json
|
| 4 |
import gradio as gr
|
|
|
|
| 5 |
|
| 6 |
class ValidationStatus(BaseModel):
|
| 7 |
is_valid: bool
|
|
@@ -19,14 +20,41 @@ class SQLQueryGeneration(BaseModel):
|
|
| 19 |
execution_results: str
|
| 20 |
optimization_notes: list[str]
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
def generate_sql_query(api_key, user_query):
|
| 23 |
"""Generate SQL query from natural language using GROQ API"""
|
| 24 |
try:
|
| 25 |
if not api_key:
|
| 26 |
-
return "Error: Please enter your GROQ API key", "", "", "",
|
| 27 |
|
| 28 |
if not user_query:
|
| 29 |
-
return "Error: Please enter a query description", "", "", "",
|
| 30 |
|
| 31 |
client = Groq(api_key=api_key)
|
| 32 |
|
|
@@ -41,12 +69,24 @@ After generating the SQL query, you must:
|
|
| 41 |
2. Populate the table with realistic sample data that demonstrates the query's functionality
|
| 42 |
3. Execute the generated SQL query against the sample table
|
| 43 |
4. Display the SQL table structure and data clearly
|
| 44 |
-
5. Show the query execution results
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
Always present your response in this order:
|
| 46 |
- Generated SQL query with syntax explanation
|
| 47 |
- Table schema (CREATE TABLE statement)
|
| 48 |
- Sample data (INSERT statements or table visualization)
|
| 49 |
-
- Query execution results
|
| 50 |
- Any relevant notes about assumptions made or query optimization suggestions""",
|
| 51 |
},
|
| 52 |
{
|
|
@@ -87,18 +127,21 @@ Execution Notes:
|
|
| 87 |
Optimization Notes:
|
| 88 |
{chr(10).join(f"- {note}" for note in sql_query_generation.optimization_notes)}"""
|
| 89 |
|
|
|
|
|
|
|
|
|
|
| 90 |
return (
|
| 91 |
sql_query_generation.query,
|
| 92 |
metadata,
|
| 93 |
sql_query_generation.table_schema,
|
| 94 |
sql_query_generation.sample_data,
|
| 95 |
-
|
| 96 |
validation_text
|
| 97 |
)
|
| 98 |
|
| 99 |
except Exception as e:
|
| 100 |
error_msg = f"Error: {str(e)}"
|
| 101 |
-
return error_msg, "", "", "",
|
| 102 |
|
| 103 |
# Create Gradio interface
|
| 104 |
with gr.Blocks(title="SQL Query Generator", theme=gr.themes.Soft()) as demo:
|
|
@@ -173,9 +216,14 @@ with gr.Blocks(title="SQL Query Generator", theme=gr.themes.Soft()) as demo:
|
|
| 173 |
)
|
| 174 |
|
| 175 |
with gr.Row():
|
| 176 |
-
execution_output = gr.
|
| 177 |
-
label="Execution Results",
|
| 178 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 179 |
)
|
| 180 |
|
| 181 |
generate_btn.click(
|
|
@@ -202,7 +250,7 @@ with gr.Blocks(title="SQL Query Generator", theme=gr.themes.Soft()) as demo:
|
|
| 202 |
The app will provide:
|
| 203 |
- A validated SQL query
|
| 204 |
- Table schema and sample data
|
| 205 |
-
- Execution results
|
| 206 |
- Optimization suggestions
|
| 207 |
"""
|
| 208 |
)
|
|
|
|
| 2 |
from pydantic import BaseModel
|
| 3 |
import json
|
| 4 |
import gradio as gr
|
| 5 |
+
import pandas as pd
|
| 6 |
|
| 7 |
class ValidationStatus(BaseModel):
|
| 8 |
is_valid: bool
|
|
|
|
| 20 |
execution_results: str
|
| 21 |
optimization_notes: list[str]
|
| 22 |
|
| 23 |
+
def parse_execution_results_to_dataframe(execution_results):
|
| 24 |
+
"""Convert text-based table results to pandas DataFrame"""
|
| 25 |
+
try:
|
| 26 |
+
lines = execution_results.strip().split('\n')
|
| 27 |
+
if len(lines) < 3:
|
| 28 |
+
return None
|
| 29 |
+
|
| 30 |
+
# Extract header
|
| 31 |
+
header_line = lines[0]
|
| 32 |
+
headers = [col.strip() for col in header_line.split('|')]
|
| 33 |
+
|
| 34 |
+
# Extract data rows (skip separator line)
|
| 35 |
+
data_rows = []
|
| 36 |
+
for line in lines[2:]:
|
| 37 |
+
if line.strip() and not line.strip().startswith('-'):
|
| 38 |
+
row = [cell.strip() for cell in line.split('|')]
|
| 39 |
+
if len(row) == len(headers):
|
| 40 |
+
data_rows.append(row)
|
| 41 |
+
|
| 42 |
+
if data_rows:
|
| 43 |
+
df = pd.DataFrame(data_rows, columns=headers)
|
| 44 |
+
return df
|
| 45 |
+
return None
|
| 46 |
+
except Exception as e:
|
| 47 |
+
print(f"Error parsing results: {e}")
|
| 48 |
+
return None
|
| 49 |
+
|
| 50 |
def generate_sql_query(api_key, user_query):
|
| 51 |
"""Generate SQL query from natural language using GROQ API"""
|
| 52 |
try:
|
| 53 |
if not api_key:
|
| 54 |
+
return "Error: Please enter your GROQ API key", "", "", "", None, ""
|
| 55 |
|
| 56 |
if not user_query:
|
| 57 |
+
return "Error: Please enter a query description", "", "", "", None, ""
|
| 58 |
|
| 59 |
client = Groq(api_key=api_key)
|
| 60 |
|
|
|
|
| 69 |
2. Populate the table with realistic sample data that demonstrates the query's functionality
|
| 70 |
3. Execute the generated SQL query against the sample table
|
| 71 |
4. Display the SQL table structure and data clearly
|
| 72 |
+
5. Show the query execution results in a pipe-delimited table format
|
| 73 |
+
|
| 74 |
+
IMPORTANT: The execution_results field must contain a properly formatted table with:
|
| 75 |
+
- Header row with column names separated by pipes (|)
|
| 76 |
+
- A separator row with dashes
|
| 77 |
+
- Data rows with values separated by pipes (|)
|
| 78 |
+
|
| 79 |
+
Example format:
|
| 80 |
+
column1 | column2 | column3
|
| 81 |
+
--------|---------|--------
|
| 82 |
+
value1 | value2 | value3
|
| 83 |
+
value4 | value5 | value6
|
| 84 |
+
|
| 85 |
Always present your response in this order:
|
| 86 |
- Generated SQL query with syntax explanation
|
| 87 |
- Table schema (CREATE TABLE statement)
|
| 88 |
- Sample data (INSERT statements or table visualization)
|
| 89 |
+
- Query execution results (in pipe-delimited table format)
|
| 90 |
- Any relevant notes about assumptions made or query optimization suggestions""",
|
| 91 |
},
|
| 92 |
{
|
|
|
|
| 127 |
Optimization Notes:
|
| 128 |
{chr(10).join(f"- {note}" for note in sql_query_generation.optimization_notes)}"""
|
| 129 |
|
| 130 |
+
# Convert execution results to DataFrame
|
| 131 |
+
results_df = parse_execution_results_to_dataframe(sql_query_generation.execution_results)
|
| 132 |
+
|
| 133 |
return (
|
| 134 |
sql_query_generation.query,
|
| 135 |
metadata,
|
| 136 |
sql_query_generation.table_schema,
|
| 137 |
sql_query_generation.sample_data,
|
| 138 |
+
results_df,
|
| 139 |
validation_text
|
| 140 |
)
|
| 141 |
|
| 142 |
except Exception as e:
|
| 143 |
error_msg = f"Error: {str(e)}"
|
| 144 |
+
return error_msg, "", "", "", None, ""
|
| 145 |
|
| 146 |
# Create Gradio interface
|
| 147 |
with gr.Blocks(title="SQL Query Generator", theme=gr.themes.Soft()) as demo:
|
|
|
|
| 216 |
)
|
| 217 |
|
| 218 |
with gr.Row():
|
| 219 |
+
execution_output = gr.Dataframe(
|
| 220 |
+
label="📊 Execution Results",
|
| 221 |
+
headers=None,
|
| 222 |
+
datatype="str",
|
| 223 |
+
row_count=10,
|
| 224 |
+
col_count=None,
|
| 225 |
+
wrap=True,
|
| 226 |
+
interactive=False
|
| 227 |
)
|
| 228 |
|
| 229 |
generate_btn.click(
|
|
|
|
| 250 |
The app will provide:
|
| 251 |
- A validated SQL query
|
| 252 |
- Table schema and sample data
|
| 253 |
+
- Execution results in Excel-style table format
|
| 254 |
- Optimization suggestions
|
| 255 |
"""
|
| 256 |
)
|