Upload sql_t5.py
Browse files
sql_t5.py
ADDED
|
@@ -0,0 +1,141 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
|
| 3 |
+
import torch
|
| 4 |
+
import os
|
| 5 |
+
|
| 6 |
+
def make_query(context, question):
|
| 7 |
+
result_query = f'''You are a SQL expert with extensive experience, you need to create a query to answer the question.
|
| 8 |
+
|
| 9 |
+
### Database schema (PostgreSQL):
|
| 10 |
+
{context}
|
| 11 |
+
|
| 12 |
+
### Question:
|
| 13 |
+
{question}
|
| 14 |
+
|
| 15 |
+
### SQL Query: '''
|
| 16 |
+
return result_query
|
| 17 |
+
|
| 18 |
+
st.set_page_config(page_title="SQL-to-Text with T5", page_icon="🤖")
|
| 19 |
+
st.title("SQL Query Generator with T5")
|
| 20 |
+
|
| 21 |
+
examples = [
|
| 22 |
+
{
|
| 23 |
+
"question": "Who was the music director in 1971 for the movie Kalyani?",
|
| 24 |
+
"description": "CREATE TABLE table_name_7 (Music VARCHAR, year VARCHAR, movie__in_kannada__ VARCHAR)"
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"question": "What's the highest with a capacity of greater than 4,000 and an average of 615?",
|
| 28 |
+
"description": "CREATE TABLE table_name_20 (highest INTEGER, average VARCHAR, capacity VARCHAR)"
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"question": "If the letters is φαν, what is the founding date?",
|
| 32 |
+
"description": "CREATE TABLE table_2538117_7 (founding_date VARCHAR, letters VARCHAR)"
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"question": "How many weeks had a game on November 26, 1978, and an attendance higher than 26,248?",
|
| 36 |
+
"description": "CREATE TABLE table_name_88 (week VARCHAR, date VARCHAR, attendance VARCHAR)"
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"question": "How many television service are in italian and n°is greater than 856.0?",
|
| 40 |
+
"description": "CREATE TABLE table_15887683_15 (television_service VARCHAR, language VARCHAR, n° VARCHAR)"
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"question": "What date was Bury the home team?",
|
| 44 |
+
"description": "CREATE TABLE table_name_67 (date VARCHAR, away_team VARCHAR)"
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"question": "What regular season result had an average attendance less than 942?",
|
| 48 |
+
"description": "CREATE TABLE table_name_16 (reg_season VARCHAR, avg_attendance INTEGER)"
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"question": "What is the value for 2011 when `a` is the value for 2009, and 4r is the value for 2013?",
|
| 52 |
+
"description": "CREATE TABLE table_name_89 (Id VARCHAR)"
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"question": "Who wrote episode with production code 1.01?",
|
| 56 |
+
"description": "CREATE TABLE table_28089666_1 (written_by VARCHAR, production_code VARCHAR)"
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"question": "find the names of museums which have more staff than the minimum staff number of all museums opened after 2010.",
|
| 60 |
+
"description": "CREATE TABLE museum (name VARCHAR, num_of_staff INTEGER, open_year INTEGER)"
|
| 61 |
+
}
|
| 62 |
+
]
|
| 63 |
+
|
| 64 |
+
@st.cache_resource
|
| 65 |
+
def load_model():
|
| 66 |
+
script_dir = os.path.dirname(os.path.abspath(__file__))
|
| 67 |
+
model_path = os.path.join(script_dir, "model")
|
| 68 |
+
|
| 69 |
+
if not os.path.exists(model_path):
|
| 70 |
+
raise FileNotFoundError(f"Model directory not found at {model_path}")
|
| 71 |
+
|
| 72 |
+
try:
|
| 73 |
+
tokenizer = AutoTokenizer.from_pretrained("google-t5/t5-small")
|
| 74 |
+
model = AutoModelForSeq2SeqLM.from_pretrained(model_path)
|
| 75 |
+
model.eval()
|
| 76 |
+
return model, tokenizer
|
| 77 |
+
except Exception as e:
|
| 78 |
+
raise RuntimeError(f"Error loading model: {str(e)}")
|
| 79 |
+
|
| 80 |
+
try:
|
| 81 |
+
model, tokenizer = load_model()
|
| 82 |
+
except Exception as e:
|
| 83 |
+
st.error(f"Failed to load model: {str(e)}")
|
| 84 |
+
st.stop()
|
| 85 |
+
|
| 86 |
+
if 'current_description' not in st.session_state:
|
| 87 |
+
st.session_state.current_description = """CREATE TABLE table_name_28 (played INTEGER, points VARCHAR, position VARCHAR)"""
|
| 88 |
+
if 'current_question' not in st.session_state:
|
| 89 |
+
st.session_state.current_question = "Which Played has a Points of 2, and a Position smaller than 8?"
|
| 90 |
+
|
| 91 |
+
def load_example(example):
|
| 92 |
+
st.session_state.current_description = example["description"]
|
| 93 |
+
st.session_state.current_question = example["question"]
|
| 94 |
+
|
| 95 |
+
st.subheader("Примеры:")
|
| 96 |
+
cols = st.columns(2)
|
| 97 |
+
for i, example in enumerate(examples):
|
| 98 |
+
col = cols[i % 2]
|
| 99 |
+
if col.button(
|
| 100 |
+
f"Пример {i+1}: {example['question'][:30]}...",
|
| 101 |
+
key=f"example_{i}",
|
| 102 |
+
):
|
| 103 |
+
load_example(example)
|
| 104 |
+
st.rerun()
|
| 105 |
+
|
| 106 |
+
with st.form("query_form"):
|
| 107 |
+
description = st.text_area(
|
| 108 |
+
"Описание таблицы (столбцы и их типы):",
|
| 109 |
+
st.session_state.current_description,
|
| 110 |
+
height=150,
|
| 111 |
+
key="desc_input"
|
| 112 |
+
)
|
| 113 |
+
|
| 114 |
+
question = st.text_input(
|
| 115 |
+
"Ваш вопрос:",
|
| 116 |
+
st.session_state.current_question,
|
| 117 |
+
key="question_input"
|
| 118 |
+
)
|
| 119 |
+
|
| 120 |
+
submitted = st.form_submit_button("Сгенерировать запрос")
|
| 121 |
+
|
| 122 |
+
if submitted:
|
| 123 |
+
if description and question:
|
| 124 |
+
input_text = make_query(description, question)
|
| 125 |
+
try:
|
| 126 |
+
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
| 127 |
+
with st.spinner("Генерация запроса..."):
|
| 128 |
+
outputs = model.generate(
|
| 129 |
+
input_ids,
|
| 130 |
+
max_length=200,
|
| 131 |
+
num_beams=5,
|
| 132 |
+
early_stopping=True,
|
| 133 |
+
pad_token_id=tokenizer.eos_token_id,
|
| 134 |
+
)
|
| 135 |
+
generated_sql = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
| 136 |
+
st.subheader("Результат:")
|
| 137 |
+
st.code(generated_sql, language="sql")
|
| 138 |
+
except Exception as e:
|
| 139 |
+
st.error(f"Ошибка при генерации: {str(e)}")
|
| 140 |
+
else:
|
| 141 |
+
st.warning("Пожалуйста, заполните описание таблицы и вопрос")
|