Debugging

- Remove duplicate & unnecessary files

app/main.py

@@ -1,38 +0,0 @@
-# Path: app/main.py
-# 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
-
-load_dotenv()
-# Load HuggingFace API token from environment variable
-hf_token = os.getenv("HF_TOKEN")
-if not hf_token:
-    raise ValueError("HuggingFace API token not found!")
-
-def main():
-    """Main application entry point and interactive menu"""
-    while True:
-        print("\n" + "="*30)
-        print(" NL2SQL Application Main Menu")
-        print("\n" + "="*30)
-        print("1. Run Interactive Agent NL2SQL Mode (Ask a single question)")
-        print("2. Run Batch Evaluation of NL2SQL Agent (Evaluate on 15 test cases)")
-        print("3. Exit")
-        print("\n" + "="*30)
-        
-        choice = input("Select an option (1-3): ")
-
-        if choice == '1':
-            run_interactiveMode()
-        elif choice == '2':
-            run_evaluation()
-        elif choice == '3':
-            print("Exiting application. Goodbye!")
-            break
-        else:
-            print("Invalid choice. Please select a valid option (1, 2, or 3).")
-
-if __name__ == "__main__":
-    main()

backend/app/main.py

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

backend/hf_evaluation_results.json

@@ -2,108 +2,136 @@
     {
         "id": 1,
         "question": "List all the artists name in the database.",
-        "status": "PASS",
-        "generated_sql": "SELECT a.Name FROM Artist a;",
+        "taxonomy": "Selection",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT Name FROM Artist",
         "gold_sql": "SELECT Name FROM Artist;"
     },
     {
         "id": 2,
         "question": "How many genres are there?",
-        "status": "FAIL",
-        "generated_sql": "SELECT COUNT(DISTINCT g.GenreId) FROM Genre g;",
+        "taxonomy": "Aggregation",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT COUNT(*) FROM Genre",
         "gold_sql": "SELECT COUNT(*) FROM Genre;"
     },
     {
         "id": 3,
         "question": "List the names of the first 5 tracks.",
-        "status": "PASS",
-        "generated_sql": "SELECT t.Name FROM Track t ORDER BY t.TrackId LIMIT 5;",
+        "taxonomy": "Selection, Limit",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT Name FROM Track LIMIT 5",
         "gold_sql": "SELECT Name FROM Track LIMIT 5;"
     },
     {
         "id": 4,
         "question": "Count the number of customers located in the USA.",
-        "status": "FAIL",
-        "generated_sql": "SELECT COUNT(DISTINCT c.CustomerId) FROM Customer c WHERE c.Country = 'USA';",
+        "taxonomy": "Aggregation, Filtering",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT COUNT(*) FROM Customer WHERE Country = 'USA'",
         "gold_sql": "SELECT COUNT(*) FROM Customer WHERE Country = 'USA';"
     },
     {
         "id": 5,
         "question": "Find all invoices for the customer with ID 1.",
-        "status": "FAIL",
-        "generated_sql": "SELECT i.InvoiceId, i.InvoiceDate, i.Total FROM Invoice i JOIN Customer c ON i.CustomerId = c.CustomerId WHERE c.CustomerId = 1;",
+        "taxonomy": "Selection, Filtering",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT * FROM Invoice WHERE CustomerId = 1",
         "gold_sql": "SELECT * FROM Invoice WHERE CustomerId = 1;"
     },
     {
         "id": 6,
         "question": "List each album title along with the artist's name.",
-        "status": "PASS",
-        "generated_sql": "SELECT a.Title, ar.Name FROM Album a JOIN Artist ar ON a.ArtistId = ar.ArtistId;",
+        "taxonomy": "Simple Join",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT Album.Title, Artist.Name FROM Album JOIN Artist ON Album.ArtistId = Artist.ArtistId",
         "gold_sql": "SELECT Album.Title, Artist.Name FROM Album JOIN Artist ON Album.ArtistId = Artist.ArtistId;"
     },
     {
         "id": 7,
         "question": "How many tracks belong to the 'Rock' genre?",
-        "status": "FAIL",
-        "generated_sql": "SELECT COUNT(t.TrackId) FROM Track t JOIN Genre g ON t.GenreId = g.GenreId WHERE LOWER(g.Name) = 'rock';",
+        "taxonomy": "Simple Join, Filtering, Aggregation",
+        "ex_pass": true,
+        "esm_pass": false,
+        "generated_sql": "SELECT COUNT(*) FROM Track WHERE GenreId = (SELECT GenreId FROM Genre WHERE Name = 'Rock')",
         "gold_sql": "SELECT COUNT(*) FROM Track JOIN Genre ON Track.GenreId = Genre.GenreId WHERE Genre.Name = 'Rock';"
     },
     {
         "id": 8,
         "question": "Show the total revenue generated from each country.",
-        "status": "FAIL",
-        "generated_sql": "SELECT i.BillingCountry, SUM(i.Total) AS TotalRevenue FROM Invoice i GROUP BY i.BillingCountry ORDER BY TotalRevenue DESC NULLS LAST;",
+        "taxonomy": "Aggregation, Grouping",
+        "ex_pass": true,
+        "esm_pass": false,
+        "generated_sql": "SELECT BillingCountry, SUM(Total) AS TotalRevenue FROM Invoice GROUP BY BillingCountry",
         "gold_sql": "SELECT BillingCountry, SUM(Total) FROM Invoice GROUP BY BillingCountry;"
     },
     {
         "id": 9,
         "question": "Find the total number of items sold for each media type.",
-        "status": "FAIL",
-        "generated_sql": "SELECT mt.Name AS MediaType, SUM(il.Quantity) AS TotalQuantity FROM InvoiceLine il JOIN Track t ON il.TrackId = t.TrackId JOIN MediaType mt ON t.MediaTypeId = mt.MediaTypeId GROUP BY mt.Name ORDER BY TotalQuantity DESC NULLS LAST;",
-        "gold_sql": "SELECT MediaType.Name, COUNT(InvoiceLine.TrackId) FROM InvoiceLine JOIN Track ON InvoiceLine.TrackId = Track.TrackId JOIN MediaType ON Track.MediaTypeId = MediaType.MediaTypeId GROUP BY MediaType.Name;"
+        "taxonomy": "Multi-Join, Aggregation, Grouping",
+        "ex_pass": true,
+        "esm_pass": false,
+        "generated_sql": "SELECT M.Name, COUNT(T.TrackId) AS TotalItemsSold\nFROM MediaType M\nJOIN Track T ON M.MediaTypeId = T.MediaTypeId\nJOIN InvoiceLine IL ON T.TrackId = IL.TrackId\nGROUP BY M.Name",
+        "gold_sql": "SELECT MediaType.Name, SUM(InvoiceLine.Quantity) FROM InvoiceLine JOIN Track ON InvoiceLine.TrackId = Track.TrackId JOIN MediaType ON Track.MediaTypeId = MediaType.MediaTypeId GROUP BY MediaType.Name;"
     },
     {
         "id": 10,
         "question": "List the first and last names of all employees who are Sales Support Agents.",
-        "status": "PASS",
-        "generated_sql": "SELECT e.FirstName, e.LastName FROM Employee e WHERE e.Title = 'Sales Support Agent';",
+        "taxonomy": "Selection, Filtering",
+        "ex_pass": true,
+        "esm_pass": true,
+        "generated_sql": "SELECT FirstName, LastName FROM Employee WHERE Title = 'Sales Support Agent'",
         "gold_sql": "SELECT FirstName, LastName FROM Employee WHERE Title = 'Sales Support Agent';"
     },
     {
         "id": 11,
         "question": "List the top 5 customers who have spent the most money in total.",
-        "status": "FAIL",
-        "generated_sql": "SELECT c.CustomerId, SUM(i.Total) AS total_spent FROM Invoice i JOIN Customer c ON i.CustomerId = c.CustomerId GROUP BY c.CustomerId ORDER BY total_spent DESC LIMIT 5;",
+        "taxonomy": "Simple Join, Aggregation, Grouping, Ordering, Limit",
+        "ex_pass": false,
+        "esm_pass": false,
+        "generated_sql": "SELECT c.CustomerId, c.FirstName, c.LastName, SUM(i.Total) AS TotalSpent\nFROM Customer c\nJOIN Invoice i ON c.CustomerId = i.CustomerId\nGROUP BY c.CustomerId, c.FirstName, c.LastName\nORDER BY TotalSpent DESC\nLIMIT 5",
         "gold_sql": "SELECT c.FirstName, c.LastName, SUM(i.Total) as TotalSpent FROM Customer c JOIN Invoice i ON c.CustomerId = i.CustomerId GROUP BY c.CustomerId ORDER BY TotalSpent DESC LIMIT 5;"
     },
     {
         "id": 12,
         "question": "Which artist has the most tracks in the database? Give the name and count.",
-        "status": "ERROR",
-        "generated_sql": "SELECT a.Name, COUNT(t.TrackId) AS track_count FROM Track t JOIN Album a ON t.AlbumId = a.AlbumId GROUP BY a.Name ORDER BY track_count DESC LIMIT 1;",
-        "gold_sql": "SELECT ar.Name, COUNT(t.TrackId) as TrackCount FROM Artist ar JOIN Album al ON ar.ArtistId = al.ArtistId JOIN Track t ON al.AlbumId = t.AlbumId GROUP BY ar.ArtistId ORDER BY TrackCount DESC LIMIT 1;",
-        "error": "Execution failed on sql 'SELECT a.Name, COUNT(t.TrackId) AS track_count FROM Track t JOIN Album a ON t.AlbumId = a.AlbumId GROUP BY a.Name ORDER BY track_count DESC LIMIT 1;': no such column: a.Name"
+        "taxonomy": "Multi-Join, Aggregation, Grouping, Ordering, Limit",
+        "ex_pass": true,
+        "esm_pass": false,
+        "generated_sql": "SELECT T2.Name, COUNT(T3.TrackId) AS TrackCount FROM Album AS T1 JOIN Artist AS T2 ON T1.ArtistId = T2.ArtistId JOIN Track AS T3 ON T1.AlbumId = T3.AlbumId GROUP BY T2.Name ORDER BY TrackCount DESC LIMIT 1",
+        "gold_sql": "SELECT ar.Name, COUNT(t.TrackId) as TrackCount FROM Artist ar JOIN Album al ON ar.ArtistId = al.ArtistId JOIN Track t ON al.AlbumId = t.AlbumId GROUP BY ar.ArtistId ORDER BY TrackCount DESC LIMIT 1;"
     },
     {
         "id": 13,
         "question": "Which genres have more than 100 tracks? List the genre name and count.",
-        "status": "FAIL",
-        "generated_sql": "SELECT g.Name, COUNT(t.TrackId) AS track_count FROM Track t JOIN Genre g ON t.GenreId = g.GenreId GROUP BY g.Name HAVING COUNT(t.TrackId) > 100 ORDER BY track_count DESC NULLS LAST;",
+        "taxonomy": "Simple Join, Aggregation, Grouping, Having",
+        "ex_pass": true,
+        "esm_pass": false,
+        "generated_sql": "SELECT G.Name, COUNT(T.TrackId) AS TrackCount FROM Genre G JOIN Track T ON G.GenreId = T.GenreId GROUP BY G.Name HAVING COUNT(T.TrackId) > 100",
         "gold_sql": "SELECT g.Name, COUNT(t.TrackId) as TrackCount FROM Genre g JOIN Track t ON g.GenreId = t.GenreId GROUP BY g.GenreId HAVING TrackCount > 100;"
     },
     {
         "id": 14,
         "question": "Calculate the average track length in seconds for each genre.",
-        "status": "FAIL",
-        "generated_sql": "SELECT g.Name, AVG(t.Milliseconds) AS average_length FROM Track t JOIN Genre g ON t.GenreId = g.GenreId GROUP BY g.Name ORDER BY average_length NULLS LAST;",
+        "taxonomy": "Simple Join, Aggregation, Arithmetic, Grouping",
+        "ex_pass": false,
+        "esm_pass": false,
+        "generated_sql": "SELECT GenreId, AVG(Milliseconds / 1000.0) AS AverageLengthInSeconds FROM Track GROUP BY GenreId",
         "gold_sql": "SELECT g.Name, AVG(t.Milliseconds) / 1000.0 as AvgSeconds FROM Genre g JOIN Track t ON g.GenreId = t.GenreId GROUP BY g.GenreId;"
     },
     {
         "id": 15,
         "question": "Identify the artist who has earned the most revenue from customers in Canada.",
-        "status": "ERROR",
-        "generated_sql": "SELECT a.Name, SUM(i.Total) AS TotalRevenue FROM Invoice i JOIN Customer c ON i.CustomerId = c.CustomerId JOIN Album a ON c.SupportRepId = a.ArtistId WHERE c.Country = 'Canada' GROUP BY a.Name ORDER BY TotalRevenue DESC LIMIT 1;",
-        "gold_sql": "SELECT ar.Name, SUM(il.UnitPrice * il.Quantity) AS Revenue FROM Artist ar JOIN Album al ON ar.ArtistId = al.ArtistId JOIN Track t ON al.AlbumId = t.AlbumId JOIN InvoiceLine il ON t.TrackId = il.TrackId JOIN Invoice i ON il.InvoiceId = i.InvoiceId WHERE i.BillingCountry = 'Canada' GROUP BY ar.ArtistId ORDER BY Revenue DESC LIMIT 1;",
-        "error": "Execution failed on sql 'SELECT a.Name, SUM(i.Total) AS TotalRevenue FROM Invoice i JOIN Customer c ON i.CustomerId = c.CustomerId JOIN Album a ON c.SupportRepId = a.ArtistId WHERE c.Country = 'Canada' GROUP BY a.Name ORDER BY TotalRevenue DESC LIMIT 1;': no such column: a.Name"
+        "taxonomy": "Multi-Join, Aggregation, Grouping, Ordering, Limit",
+        "ex_pass": false,
+        "esm_pass": false,
+        "generated_sql": "SELECT T1.Name FROM Artist AS T1 JOIN Album AS T2 ON T1.ArtistId = T2.ArtistId JOIN Track AS T5 ON T2.AlbumId = T5.AlbumId JOIN InvoiceLine AS T3 ON T5.TrackId = T3.TrackId JOIN Invoice AS T4 ON T3.InvoiceId = T4.InvoiceId WHERE T4.BillingCountry = 'Canada' GROUP BY T1.ArtistId ORDER BY SUM(T3.UnitPrice * T3.Quantity) DESC LIMIT 1",
+        "gold_sql": "SELECT ar.Name, SUM(il.UnitPrice * il.Quantity) AS Revenue FROM Artist ar JOIN Album al ON ar.ArtistId = al.ArtistId JOIN Track t ON al.AlbumId = t.AlbumId JOIN InvoiceLine il ON t.TrackId = il.TrackId JOIN Invoice i ON il.InvoiceId = i.InvoiceId WHERE i.BillingCountry = 'Canada' GROUP BY ar.ArtistId ORDER BY Revenue DESC LIMIT 1;"
     }
 ]

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 backend.src.database.db_manager import get_db_connection, get_schema_context
+from src.database.db_manager import get_db_connection, get_schema_context
 from langchain_core.prompts import PromptTemplate
-from backend.src.nl2sql.hf_engine import get_llm
+from 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.