Upload 4 files

app.py

@@ -0,0 +1,71 @@
+import os, json, sqlite3, time, textwrap, requests
+import gradio as gr
+from dotenv import load_dotenv
+load_dotenv()
+
+HF_TOKEN   = os.getenv("HF_TOKEN")  # store in Space Secrets
+MODEL_ID   = "defog/sqlcoder-7b-2"
+API_URL    = f"https://api-inference.huggingface.co/models/{MODEL_ID}"
+HEADERS    = {"Authorization": f"Bearer {HF_TOKEN}"}
+DB_PATH    = "company.db"
+
+def nlp_to_sql(nl_query: str, schema_ddl: str) -> str:
+    prompt = textwrap.dedent(f"""
+    ### Task
+    Translate the following natural language question into ONE valid SQLite SQL query.
+
+    ### Schema
+    {schema_ddl}
+
+    ### Question
+    {nl_query}
+
+    ### SQL
+    """)
+    payload = {"inputs": prompt, "parameters": {"max_new_tokens": 256}}
+    response = requests.post(API_URL, headers=HEADERS, json=payload, timeout=45)
+    sql = response.json()[0]["generated_text"].split("### SQL")[-1].strip()
+    return sql
+
+def run_pipeline(nl_query):
+    trace = []
+    start = time.time()
+
+    # 1. load schema
+    with open("schema.sql") as f:
+        schema_ddl = f.read()
+    trace.append(("Load Schema", f"{len(schema_ddl.splitlines())} lines loaded"))
+
+    # 2. NL ➜ SQL
+    sql_query = nlp_to_sql(nl_query, schema_ddl)
+    trace.append(("LLM Output", sql_query))
+
+    # 3. Execute SQL
+    try:
+        with sqlite3.connect(DB_PATH) as conn:
+            cursor = conn.execute(sql_query)
+            rows = cursor.fetchall()
+            colnames = [d[0] for d in cursor.description] if cursor.description else []
+        result = {"columns": colnames, "rows": rows}
+        trace.append(("Execution", f"{len(rows)} rows returned"))
+    except Exception as e:
+        result = {"error": str(e)}
+        trace.append(("Execution Error", str(e)))
+
+    # 4. timing
+    trace.append(("Latency", f"{time.time() - start:0.2f}s"))
+
+    return sql_query, json.dumps(result, indent=2), "\n".join([f"{s}: {m}" for s, m in trace])
+
+with gr.Blocks(title="NLP ➜ SQL Demo") as demo:
+    gr.Markdown("### NLP to SQL Query • SQLite • Trace Visibility")
+    nl_input  = gr.Textbox(label="Natural-Language Question", placeholder="e.g. List employees in Engineering hired after 2021")
+    with gr.Row():
+        sql_out   = gr.Code(label="Generated SQL")
+        result_out= gr.Code(label="Query Result")
+    trace_out = gr.Textbox(label="Trace", lines=6)
+    submit = gr.Button("Run")
+    submit.click(fn=run_pipeline, inputs=nl_input, outputs=[sql_out, result_out, trace_out])
+
+if __name__ == "__main__":
+    demo.launch()

db_builder.py

@@ -0,0 +1,13 @@
+import sqlite3, pathlib
+
+DB_PATH = pathlib.Path("company.db")
+SCHEMA_FILE = pathlib.Path("schema.sql")
+
+def build_db():
+    DB_PATH.unlink(missing_ok=True)
+    with sqlite3.connect(DB_PATH) as conn, open(SCHEMA_FILE) as f:
+        conn.executescript(f.read())
+    print("Database created")
+
+if __name__ == "__main__":
+    build_db()

requirements.txt

@@ -0,0 +1,4 @@
+gradio==4.28.3
+requests
+sqlite-utils
+python-dotenv

schema.sql

@@ -0,0 +1,14 @@
+DROP TABLE IF EXISTS employees;
+CREATE TABLE employees (
+  emp_id     INTEGER PRIMARY KEY,
+  name       TEXT,
+  department TEXT,
+  hire_date  DATE,
+  salary     INTEGER
+);
+
+INSERT INTO employees (name, department, hire_date, salary) VALUES
+  ('Alice',  'Sales',      '2022-01-10', 95000),
+  ('Bob',    'Engineering','2023-03-14',115000),
+  ('Carlos', 'Finance',    '2021-07-22',100000),
+  ('Dana',   'Engineering','2020-11-05',125000);