app.py

# import streamlit as st
# from transformers import pipeline

# # Load the SQLCoder model
# sql_generator = pipeline('text-generation', model='defog/sqlcoder')

# st.title('SQL Table Extractor')

# # Text input for SQL query
# user_sql = st.text_input("Enter your SQL statement", "SELECT * FROM my_table WHERE condition;")

# # Button to parse SQL
# if st.button('Extract Tables'):
#     # Generate SQL or parse directly
#     results = sql_generator(user_sql)
#     # Assuming results contain SQL, extract table names (this part may require custom logic based on output)
#     tables = extract_tables_from_sql(results)
    
#     # Display extracted table names
#     st.write('Extracted Tables:', tables)

# def extract_tables_from_sql(sql):
#     # Dummy function: Implement logic to parse table names from SQL
#     return ["my_table"]  # Example output

# import streamlit as st
# from transformers import pipeline

# # Load the NER model
# ner = pipeline("ner", model="dbmdz/bert-large-cased-finetuned-conll03-english", grouped_entities=True)

# st.title('Hello World NER Parser')

# # User input for text
# user_input = st.text_area("Enter a sentence to parse for named entities:", "John Smith lives in San Francisco.")

# # Parse entities
# if st.button('Parse'):
#     entities = ner(user_input)
#     # Display extracted entities
#     for entity in entities:
#         st.write(f"Entity: {entity['word']}, Entity Type: {entity['entity_group']}")

import streamlit as st
from transformers import pipeline

# Load a smaller LLaMA model with permission to run custom code
text_generator = pipeline("text-generation", model="microsoft/Phi-3-mini-128k-instruct", trust_remote_code=True)

st.title('General Query Answerer')

# User input for a general question
user_query = st.text_area("Enter your question:", "Name all 50 US states.")

# Generate answer
if st.button('Answer Question'):
    answer = text_generator(user_query, max_length=150)[0]['generated_text']
    # Display the answer
    st.write('Answer:', answer)