Upload folder using huggingface_hub

README.md

@@ -58,12 +58,16 @@ Open the printed URL (default **http://127.0.0.1:7861** if the port is free). Us
 Push the contents of this folder as the Space root, not the whole training repo:
 
 ```bash
-cd sql_compare_ui_qwen
-uv run hf repo create Tuana/qwen-text2sql-demo --type space --sdk gradio
-uv run hf upload Tuana/qwen-text2sql-demo . . \
+uv run hf repo create Tuana/qwen-text2sql-demo \
+  --type space \
+  --space-sdk gradio \
+  --exist-ok
+
+uv run hf upload Tuana/qwen-text2sql-demo sql_compare_ui_qwen . \
   --repo-type space \
   --exclude ".env" \
-  --exclude "__pycache__/*" \
+  --exclude "data/spider_eval_synthetic/synthetic.db" \
+  --exclude "**/__pycache__/**" \
   --exclude "local_inference.py" \
   --exclude "inference_device.py"
 ```

app.py

@@ -390,8 +390,69 @@ def _compare_sqlite_db_path() -> Path:
     return Path(raw).expanduser().resolve()
 
 
-def _database_preview_rows(limit: int = 5) -> list[dict[str, str]]:
+def _load_csv_rows(data_dir: Path) -> tuple[list[dict[str, str]], list[dict[str, str]], list[dict[str, str]]]:
+    with (data_dir / "department.csv").open(newline="", encoding="utf-8") as f:
+        departments = list(csv.DictReader(f))
+    with (data_dir / "head.csv").open(newline="", encoding="utf-8") as f:
+        heads = list(csv.DictReader(f))
+    with (data_dir / "management.csv").open(newline="", encoding="utf-8") as f:
+        management = list(csv.DictReader(f))
+    return departments, heads, management
+
+
+def _ensure_compare_sqlite_db() -> Path:
     db = _compare_sqlite_db_path()
+    if db.is_file():
+        return db
+
+    data_dir = _demo_data_dir()
+    departments, heads, management = _load_csv_rows(data_dir)
+    db.parent.mkdir(parents=True, exist_ok=True)
+    conn = sqlite3.connect(db)
+    try:
+        conn.executescript(
+            """
+            DROP TABLE IF EXISTS department;
+            DROP TABLE IF EXISTS management;
+            DROP TABLE IF EXISTS head;
+
+            CREATE TABLE department (
+              department_id VARCHAR,
+              name VARCHAR,
+              creation VARCHAR
+            );
+            CREATE TABLE management (
+              department_id VARCHAR,
+              head_id VARCHAR,
+              temporary_acting VARCHAR
+            );
+            CREATE TABLE head (
+              head_id VARCHAR,
+              name VARCHAR,
+              born_state VARCHAR
+            );
+            """
+        )
+        conn.executemany(
+            "INSERT INTO department (department_id, name, creation) VALUES (?, ?, ?)",
+            [(r["department_id"], r["name"], r["creation"]) for r in departments],
+        )
+        conn.executemany(
+            "INSERT INTO head (head_id, name, born_state) VALUES (?, ?, ?)",
+            [(r["head_id"], r["name"], r["born_state"]) for r in heads],
+        )
+        conn.executemany(
+            "INSERT INTO management (department_id, head_id, temporary_acting) VALUES (?, ?, ?)",
+            [(r["department_id"], r["head_id"], r["temporary_acting"]) for r in management],
+        )
+        conn.commit()
+    finally:
+        conn.close()
+    return db
+
+
+def _database_preview_rows(limit: int = 5) -> list[dict[str, str]]:
+    db = _ensure_compare_sqlite_db()
     if db.is_file():
         conn = sqlite3.connect(f"file:{db}?mode=ro", uri=True)
         conn.row_factory = sqlite3.Row
@@ -418,12 +479,9 @@ def _database_preview_rows(limit: int = 5) -> list[dict[str, str]]:
             conn.close()
 
     data_dir = _demo_data_dir()
-    with (data_dir / "department.csv").open(newline="", encoding="utf-8") as f:
-        departments = {r["department_id"]: r for r in csv.DictReader(f)}
-    with (data_dir / "head.csv").open(newline="", encoding="utf-8") as f:
-        heads = {r["head_id"]: r for r in csv.DictReader(f)}
-    with (data_dir / "management.csv").open(newline="", encoding="utf-8") as f:
-        management = list(csv.DictReader(f))
+    dept_rows, head_rows, management = _load_csv_rows(data_dir)
+    departments = {r["department_id"]: r for r in dept_rows}
+    heads = {r["head_id"]: r for r in head_rows}
 
     preview: list[dict[str, str]] = []
     for rel in management:
@@ -563,9 +621,7 @@ def _execute_compare_sql(sql: str, *, row_limit: int = 150) -> str:
     ok, stmt = _compare_validate_select(sql)
     if not ok:
         return f"Error: {stmt}"
-    db = _compare_sqlite_db_path()
-    if not db.is_file():
-        return f"Error: database file not found: {db}"
+    db = _ensure_compare_sqlite_db()
     try:
         conn = sqlite3.connect(f"file:{db}?mode=ro", uri=True)
         conn.row_factory = sqlite3.Row