Folder and Files Structure Refactor

.env

@@ -1 +0,0 @@
-HF_TOKEN='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'

{app → backend/app}/main.py

@@ -2,8 +2,8 @@
 # Main entry point for the NL2SQL application
 import os
 from dotenv import load_dotenv
-from src.scripts.interactive_mode import run_interactiveMode
-from src.scripts.evaluation_mode import run_evaluation
+from backend.src.scripts.interactive_mode import run_interactiveMode
+from backend.src.scripts.evaluation_mode import run_evaluation
 
 load_dotenv()
 # Load HuggingFace API token from environment variable

hf_test_bench.py → backend/hf_test_bench.py

@@ -1,6 +1,6 @@
 # Test the Hugging Face inference
-from src.nl2sql.hf_engine import generate_sql
-from src.database.db_manager import get_db_connection, get_schema_context
+from backend.src.nl2sql.hf_engine import generate_sql
+from backend.src.database.db_manager import get_db_connection, get_schema_context
 import pandas as pd
 
 def test_single_query():

{src → backend/src}/nl2sql/sql_agent.py

@@ -1,8 +1,8 @@
 # Path: src/nl2sql/sql_agent.py
 # SQL Agent for handling NL2SQL conversion with Auto-Correct functionality
-from src.database.db_manager import get_db_connection, get_schema_context
+from backend.src.database.db_manager import get_db_connection, get_schema_context
 from langchain_core.prompts import PromptTemplate
-from src.nl2sql.hf_engine import get_llm
+from backend.src.nl2sql.hf_engine import get_llm
 
 # Craft the Prompt Template to instruct LLM on its persona
 SQL_PROMPT_TEMPLATE = """You are an expert SQLite developer.

{src → backend/src}/scripts/evaluate.py

@@ -2,8 +2,8 @@
 
 import json
 import pandas as pd
-from src.database.db_manager import get_db_connection, get_schema_context
-from src.nl2sql.qwen_engine import query_qwen, build_qwen_prompt
+from backend.src.database.db_manager import get_db_connection, get_schema_context
+from backend.src.nl2sql.qwen_engine import query_qwen, build_qwen_prompt
 
 def compare_results(df_generated_sql, df_gold_sql):
     """Compares the generated SQL query with the expected SQL query."""

{src → backend/src}/scripts/evaluation_mode.py

@@ -3,8 +3,8 @@
 import json
 from pathlib import Path
 import pandas as pd
-from src.database.db_manager import get_db_connection
-from src.nl2sql.sql_agent import nl2sql_agent
+from backend.src.database.db_manager import get_db_connection
+from backend.src.nl2sql.sql_agent import nl2sql_agent
 
 TEST_CASES_PATH = Path("src/scripts/test_cases.json")
 RESULTS_PATH = Path("hf_evaluation_results.json")

{src → backend/src}/scripts/interactive_mode.py

@@ -1,6 +1,6 @@
 # Path: src/scripts/interactive_mode.py
 # Interactive mode: Allows user to manually type questions and see the agent's response
-from src.nl2sql.sql_agent import nl2sql_agent
+from backend.src.nl2sql.sql_agent import nl2sql_agent
 
 def run_interactiveMode():
     """ Allows user to manually type questions and get agent's response."""

evaluation_results.json

@@ -1,72 +0,0 @@
-[
-    {
-        "id": 1,
-        "status": "ERROR: Execution failed on sql 'SELECT t.Genre, COUNT(t.TrackId) AS TrackCount FROM Track t GROUP BY t.Genre;': no such column: t.Genre"
-    },
-    {
-        "id": 2,
-        "status": "PASS",
-        "generated": "SELECT t1.Title, ar.Name FROM Album AS t1 INNER JOIN Artist AS ar ON t1.ArtistId = ar.ArtistId;"
-    },
-    {
-        "id": 3,
-        "status": "ERROR: Execution failed on sql 'SELECT T4.Country, SUM(T3.UnitPrice * T3.Quantity) as TotalRevenue\nFROM InvoiceLine AS T1\nJOIN Track AS T2 ON T1.TrackId = T2.TrackId\nJOIN Album AS T3 ON T2.AlbumId = T3.AlbumId\nJOIN Genre AS T4 ON T2.GenreId = T4.GenreId\nGROUP BY T4.Country;': no such column: T4.Country"
-    },
-    {
-        "id": 4,
-        "status": "ERROR: Execution failed on sql 'SELECT e.FirstName, e.LastName FROM Employee AS e JOIN Customer AS c ON e.SupportRepId = c.CustomerId WHERE c.Company LIKE 'Sales Support';': no such column: e.SupportRepId"
-    },
-    {
-        "id": 5,
-        "status": "FAIL",
-        "generated": "SELECT c.FirstName, c.LastName, SUM(il.UnitPrice * il.Quantity) as TotalSpent\nFROM Customer c\nJOIN Invoice i ON c.CustomerId = i.CustomerId\nJOIN InvoiceLine il ON i.InvoiceId = il.InvoiceId\nGROUP BY c.CustomerId\nORDER BY TotalSpent DESC\nLIMIT 5;"
-    },
-    {
-        "id": 6,
-        "status": "FAIL",
-        "generated": "SELECT T3.Name AS ArtistName, T1.Name AS SongName\nFROM Track AS T1\nJOIN Album AS T2 ON T1.AlbumId = T2.AlbumId\nJOIN Genre AS T3 ON T1.GenreId = T3.GenreId\nWHERE T3.Name = 'Rock'\nGROUP BY T1.TrackId;"
-    },
-    {
-        "id": 7,
-        "status": "ERROR: Execution failed on sql 'SELECT T3.Name AS MediaTypeName, COUNT(T1.TrackId) AS TotalTracksSold\nFROM Track AS T1\nJOIN MediaType AS T2 ON T1.MediaTypeId = T2.MediaTypeId\nGROUP BY T2.Name;': no such column: T3.Name"
-    },
-    {
-        "id": 8,
-        "status": "PASS",
-        "generated": "SELECT T3.Name FROM Playlist AS T1 INNER JOIN PlaylistTrack AS T2 ON T1.PlaylistId = T2.PlaylistId INNER JOIN Track AS T3 ON T2.TrackId = T3.TrackId WHERE T1.Name = 'TV Shows';"
-    },
-    {
-        "id": 9,
-        "status": "PASS",
-        "generated": "SELECT a.Name, COUNT(t.TrackId) AS TrackCount\nFROM Artist a\nJOIN Album ar ON a.ArtistId = ar.ArtistId\nJOIN Track t ON ar.AlbumId = t.AlbumId\nGROUP BY a.ArtistId\nORDER BY TrackCount DESC\nLIMIT 1;"
-    },
-    {
-        "id": 10,
-        "status": "FAIL",
-        "generated": "SELECT g.Name FROM Genre g JOIN Track t ON g.GenreId = t.GenreId GROUP BY g.Name HAVING COUNT(t.TrackId) > 100;"
-    },
-    {
-        "id": 11,
-        "status": "FAIL",
-        "generated": "SELECT a.Name, SUM(il.UnitPrice * il.Quantity) as TotalRevenue\nFROM Artist a\nJOIN Album al ON a.ArtistId = al.ArtistId\nJOIN Track tr ON al.AlbumId = tr.AlbumId\nJOIN InvoiceLine il ON tr.TrackId = il.TrackId\nGROUP BY a.ArtistId\nORDER BY TotalRevenue DESC\nLIMIT 1;"
-    },
-    {
-        "id": 12,
-        "status": "FAIL",
-        "generated": "SELECT a.Title, AVG(t.Milliseconds) / 1000 AS AverageLengthInSeconds\nFROM Album a\nJOIN Track t ON a.AlbumId = t.AlbumId\nGROUP BY a.AlbumId;"
-    },
-    {
-        "id": 13,
-        "status": "PASS",
-        "generated": "SELECT c.FirstName, c.LastName, e.FirstName, e.LastName\nFROM Customer c\nJOIN Employee e ON c.SupportRepId = e.EmployeeId\nWHERE e.LastName = 'Peacock';"
-    },
-    {
-        "id": 14,
-        "status": "PASS",
-        "generated": "SELECT t2.City FROM Invoice AS t1 INNER JOIN Customer AS t2 ON t1.CustomerId = t2.CustomerId WHERE strftime('%Y', t1.InvoiceDate) = '2013' GROUP BY t2.City ORDER BY COUNT(*) DESC LIMIT 1;"
-    },
-    {
-        "id": 15,
-        "status": "ERROR: Execution failed on sql 'SELECT T1.Title FROM Album AS T1 INNER JOIN InvoiceLine AS T2 ON T1.AlbumId = T2.AlbumId GROUP BY T1.AlbumId HAVING SUM(T2.UnitPrice * T2.Quantity) > 20;': no such column: T2.AlbumId"
-    }
-]

test_bench.py

@@ -1,36 +0,0 @@
-# Test the database connection and schema extraction
-
-import pandas as pd
-from src.database.db_manager import get_db_connection, get_schema_context
-from src.nl2sql.qwen_engine import query_qwen, build_qwen_prompt
-
-def run_test(question):
-    print(f"\n[USER QUESTION]: {question}")
-
-    # Step 1: Get the database schema context from db_manager
-    schema = get_schema_context()
-
-    # Step 2: Build the prompt and Generate SQL query
-    prompt = build_qwen_prompt(question, schema)
-    generated_sql = query_qwen(prompt)
-    print(f"\n[GENERATED SQL]:\n{generated_sql}")
-
-    # Step 3: Execute the generated SQL query against the database
-    try:
-        connection = get_db_connection()
-        # Use pandas to execute the SQL query and display results
-        result_df = pd.read_sql_query(generated_sql, connection)
-        connection.close()
-
-        print("[EXECUTION SUCCESS]:")
-        if result_df.empty:
-            print("Query returned no results.")
-        else:
-            print(result_df.head(5))  # Display the first 5 rows of the result
-    except Exception as e:
-        print(f"[EXECUTION ERROR]: {e}")
-
-if __name__ == "__main__":
-    # Example test question
-    test_question = "What is the total revenue generated from each country?"
-    run_test(test_question)