signalpilot-ai-internal 0.3.1__py3-none-any.whl → 0.3.3__py3-none-any.whl

signalpilot_ai_internal/snowflake_schema_service.py

@@ -0,0 +1,639 @@
+"""
+Snowflake schema service handlers for SignalPilot AI.
+Provides REST API handlers for Snowflake database schema retrieval and query execution.
+Supports multiple databases within a single Snowflake connection.
+
+Behavior:
+- Picks the smallest RUNNING warehouse.
+- If none running, resumes the smallest SUSPENDED warehouse.
+- If none exist, attempts to CREATE a tiny warehouse (requires privilege).
+- Builds a catalog with parallel schema processing for performance.
+- For each table, includes detailed column information: name, type, ordinal position,
+  nullable, description, default value, and type-specific attributes.
+"""
+
+import json
+import os
+import re
+import subprocess
+import sys
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from typing import Any, Dict, Optional, List
+
+from jupyter_server.base.handlers import APIHandler
+import tornado
+
+SIZE_ORDER = ["XSMALL", "SMALL", "MEDIUM", "LARGE", "XLARGE", "XXLARGE", "XXXLARGE", "X4LARGE", "X5LARGE", "X6LARGE"]
+
+
+class SnowflakeSchemaHandler(APIHandler):
+    """Handler for Snowflake schema operations"""
+    
+    def _setup_snowflake_environment(self):
+        """Install required Snowflake packages if not available"""
+        def install_package(package_name):
+            try:
+                subprocess.check_call([sys.executable, "-m", "pip", "install", package_name])
+                return True
+            except subprocess.CalledProcessError:
+                return False
+
+        missing_packages = []
+        
+        try:
+            import snowflake.connector
+        except ImportError:
+            if install_package("snowflake-connector-python"):
+                try:
+                    import snowflake.connector
+                except ImportError as e:
+                    missing_packages.append(f"snowflake-connector-python: {str(e)}")
+            else:
+                missing_packages.append("snowflake-connector-python: installation failed")
+
+        if missing_packages:
+            raise ImportError("Required modules could not be installed: " + ", ".join(missing_packages))
+        
+        import snowflake.connector
+        return snowflake.connector
+
+    def _get_snowflake_config(self, provided_config: Optional[Dict] = None) -> Optional[Dict]:
+        """Get Snowflake configuration from request or environment variables"""
+        if provided_config:
+            return provided_config
+        
+        # Look for Snowflake database configuration in the environment
+        for key, value in os.environ.items():
+            if key.endswith('_CONNECTION_JSON'):
+                try:
+                    config = json.loads(value)
+                    if config.get('type') == 'snowflake':
+                        return config
+                except Exception as e:
+                    print(f"[SnowflakeSchemaHandler] Error parsing database config {key}: {e}")
+                    continue
+        
+        return None
+    
+    def _get_connection_params(self, config: Dict) -> Dict[str, Any]:
+        """Build Snowflake connection parameters from configuration"""
+        # Extract account from connectionUrl
+        connection_url = config.get('connectionUrl', '')
+        if not connection_url:
+            raise ValueError("connectionUrl is required for Snowflake")
+        
+        # Extract the account identifier from the connectionUrl
+        url_match = re.match(r'https?://([^/]+)', connection_url)
+        if not url_match:
+            raise ValueError(f"Invalid Snowflake connectionUrl format: {connection_url}")
+        
+        account = url_match.group(1)
+        # Strip .snowflakecomputing.com if present
+        account_identifier = account.replace('.snowflakecomputing.com', '')
+        
+        conn_params = {
+            'account': account_identifier,
+            'user': config['username'],
+            'password': config['password'],
+        }
+        
+        warehouse = config.get('warehouse')
+        database = config.get('database')
+        role = config.get('role')
+        
+        if warehouse:
+            conn_params['warehouse'] = warehouse
+        if database:
+            conn_params['database'] = database
+        if role:
+            conn_params['role'] = role
+        
+        return conn_params
+    
+    def _fetch_result_scan(self, cur, sql: str, name_col: str = "name") -> List[str]:
+        """Execute SQL and fetch results using RESULT_SCAN"""
+        cur.execute(sql)
+        cur.execute(f'SELECT "{name_col}" FROM TABLE(RESULT_SCAN(LAST_QUERY_ID()))')
+        rows = cur.fetchall()
+        # Handle both DictCursor (returns dicts) and regular cursor (returns tuples)
+        if rows and isinstance(rows[0], dict):
+            return [r[name_col] for r in rows]
+        return [r[0] for r in rows]
+    
+    def _get_warehouses(self, cur) -> List[Dict]:
+        """Get all warehouses with their state and size"""
+        cur.execute("SHOW WAREHOUSES")
+        cur.execute("""
+            SELECT "name","state","size","auto_suspend","auto_resume"
+            FROM TABLE(RESULT_SCAN(LAST_QUERY_ID()))
+        """)
+        rows = cur.fetchall()
+        # Handle DictCursor (returns dicts) and regular cursor (returns tuples)
+        if rows and isinstance(rows[0], dict):
+            # DictCursor: normalize keys to lowercase
+            return [{k.lower(): v for k, v in row.items()} for row in rows]
+        else:
+            # Regular cursor: manually create dicts
+            cols = [d[0].lower() for d in cur.description]
+            return [dict(zip(cols, row)) for row in rows]
+    
+    def _size_rank(self, sz: str) -> int:
+        """Get the numeric rank of a warehouse size"""
+        s = (sz or "").upper()
+        return SIZE_ORDER.index(s) if s in SIZE_ORDER else len(SIZE_ORDER) + 1
+    
+    def _choose_smallest_running(self, warehouses: List[Dict]) -> Optional[str]:
+        """Choose the smallest running warehouse"""
+        running = [w for w in warehouses if (w.get("state") or "").upper() == "STARTED"]
+        if not running:
+            return None
+        running.sort(key=lambda w: self._size_rank(w.get("size")))
+        return running[0]["name"]
+    
+    def _choose_smallest_suspended(self, warehouses: List[Dict]) -> Optional[str]:
+        """Choose the smallest suspended warehouse"""
+        suspended = [w for w in warehouses if (w.get("state") or "").upper() in ("SUSPENDED", "RESIZING")]
+        if not suspended:
+            return None
+        suspended.sort(key=lambda w: self._size_rank(w.get("size")))
+        return suspended[0]["name"]
+    
+    def _resume_warehouse(self, cur, name: str) -> None:
+        """Resume a suspended warehouse"""
+        cur.execute(f'ALTER WAREHOUSE "{name}" RESUME')
+    
+    def _create_tiny_warehouse(self, cur, name: str = "SPAI_TINY_WH") -> str:
+        """Create a tiny warehouse (requires proper privilege)"""
+        cur.execute(f'''
+            CREATE WAREHOUSE IF NOT EXISTS "{name}"
+            WITH WAREHOUSE_SIZE = XSMALL
+                 AUTO_SUSPEND = 60
+                 AUTO_RESUME = TRUE
+                 INITIALLY_SUSPENDED = TRUE
+        ''')
+        # Start it
+        cur.execute(f'ALTER WAREHOUSE "{name}" RESUME')
+        return name
+    
+    def _ensure_warehouse(self, cur, preferred: Optional[str]) -> str:
+        """Ensure a warehouse is available and running"""
+        # Respect an explicitly provided warehouse first, if any
+        if preferred:
+            try:
+                cur.execute(f'SHOW WAREHOUSES LIKE \'{preferred}\'')
+                cur.execute('SELECT "name","state","size" FROM TABLE(RESULT_SCAN(LAST_QUERY_ID()))')
+                row = cur.fetchone()
+                if row:
+                    # Handle DictCursor vs regular cursor
+                    state = (row["state"] if isinstance(row, dict) else row[1] or "").upper()
+                    if state != "STARTED":
+                        self._resume_warehouse(cur, preferred)
+                    return preferred
+            except Exception as e:
+                # Fall back to discovery below
+                print(f"Note: preferred warehouse '{preferred}' not available or cannot be resumed ({e}). Falling back.")
+
+        warehouses = self._get_warehouses(cur)
+        name = self._choose_smallest_running(warehouses)
+        if name:
+            return name
+
+        name = self._choose_smallest_suspended(warehouses)
+        if name:
+            self._resume_warehouse(cur, name)
+            return name
+
+        # None exist → create tiny one
+        return self._create_tiny_warehouse(cur)
+    
+    def _list_databases(self, cur) -> List[str]:
+        """List all databases"""
+        return self._fetch_result_scan(cur, "SHOW DATABASES", "name")
+    
+    def _list_schemas_for_db(self, cur, db: str) -> List[str]:
+        """List all schemas for a database (excluding INFORMATION_SCHEMA)"""
+        cur.execute(f'USE DATABASE "{db}"')
+        schemas = self._fetch_result_scan(cur, "SHOW SCHEMAS", "name")
+        return [s for s in schemas if s.upper() != "INFORMATION_SCHEMA"]
+
+    def _list_tables_with_columns_for_schema(self, connector, conn, db: str, schema: str, limit: int = 5000) -> List[Dict]:
+        """Get tables and their columns for a schema using optimized bulk query."""
+        cur = conn.cursor(connector.DictCursor)
+        try:
+            cur.execute(f'USE DATABASE "{db}"')
+            cur.execute(f'USE SCHEMA "{schema}"')
+            
+            # Get all tables and columns in one query for better performance
+            cur.execute("""
+                SELECT 
+                    t.TABLE_SCHEMA,
+                    t.TABLE_NAME,
+                    t.TABLE_TYPE,
+                    c.COLUMN_NAME,
+                    c.DATA_TYPE,
+                    c.ORDINAL_POSITION,
+                    c.IS_NULLABLE,
+                    c.COLUMN_DEFAULT,
+                    c.CHARACTER_MAXIMUM_LENGTH,
+                    c.NUMERIC_PRECISION,
+                    c.NUMERIC_SCALE,
+                    c.COMMENT
+                FROM INFORMATION_SCHEMA.TABLES t
+                LEFT JOIN INFORMATION_SCHEMA.COLUMNS c
+                    ON t.TABLE_SCHEMA = c.TABLE_SCHEMA 
+                    AND t.TABLE_NAME = c.TABLE_NAME
+                WHERE t.TABLE_SCHEMA = %s
+                ORDER BY t.TABLE_NAME, c.ORDINAL_POSITION
+                LIMIT 50000
+            """, (schema,))
+            rows = cur.fetchall()
+            
+            # Group by table
+            tables_dict = {}
+            for r in rows:
+                if isinstance(r, dict):
+                    table_key = r["TABLE_NAME"]
+                    if table_key not in tables_dict:
+                        tables_dict[table_key] = {
+                            "schema": r["TABLE_SCHEMA"],
+                            "table": r["TABLE_NAME"],
+                            "type": r["TABLE_TYPE"],
+                            "columns": []
+                        }
+                    
+                    if r.get("COLUMN_NAME"):
+                        col = {
+                            "name": r["COLUMN_NAME"],
+                            "type": r["DATA_TYPE"],
+                            "ordinal": r["ORDINAL_POSITION"],
+                            "nullable": r["IS_NULLABLE"] == "YES",
+                        }
+                        if r.get("COMMENT"):
+                            col["description"] = r["COMMENT"]
+                        if r.get("COLUMN_DEFAULT"):
+                            col["default"] = r["COLUMN_DEFAULT"]
+                        if r.get("CHARACTER_MAXIMUM_LENGTH"):
+                            col["max_length"] = r["CHARACTER_MAXIMUM_LENGTH"]
+                        if r.get("NUMERIC_PRECISION"):
+                            col["precision"] = r["NUMERIC_PRECISION"]
+                        if r.get("NUMERIC_SCALE") is not None:
+                            col["scale"] = r["NUMERIC_SCALE"]
+                        tables_dict[table_key]["columns"].append(col)
+                else:
+                    table_key = r[1]
+                    if table_key not in tables_dict:
+                        tables_dict[table_key] = {
+                            "schema": r[0],
+                            "table": r[1],
+                            "type": r[2],
+                            "columns": []
+                        }
+                    
+                    if r[3]:  # COLUMN_NAME
+                        col = {
+                            "name": r[3],
+                            "type": r[4],
+                            "ordinal": r[5],
+                            "nullable": r[6] == "YES",
+                        }
+                        if r[11]:  # COMMENT
+                            col["description"] = r[11]
+                        if r[7]:  # COLUMN_DEFAULT
+                            col["default"] = r[7]
+                        if r[8]:  # CHARACTER_MAXIMUM_LENGTH
+                            col["max_length"] = r[8]
+                        if r[9]:  # NUMERIC_PRECISION
+                            col["precision"] = r[9]
+                        if r[10] is not None:  # NUMERIC_SCALE
+                            col["scale"] = r[10]
+                        tables_dict[table_key]["columns"].append(col)
+            
+            return list(tables_dict.values())[:limit]
+        finally:
+            cur.close()
+    
+    def _process_schema(self, connector, conn, db: str, schema: str) -> Dict:
+        """Process a single schema with its tables and columns."""
+        try:
+            tables = self._list_tables_with_columns_for_schema(connector, conn, db, schema)
+            return {"schema": schema, "tables": tables, "error": None}
+        except Exception as e:
+            print(f"Warning: Error processing schema {db}.{schema}: {e}", file=sys.stderr)
+            return {"schema": schema, "tables": [], "error": str(e)}
+    
+    def _build_catalog(self, connector, conn, max_workers: int = 5) -> Dict:
+        """Build complete catalog with parallel schema processing"""
+        cur = conn.cursor(connector.DictCursor)
+        try:
+            # 1) Ensure a small warehouse is available
+            preferred_wh = None
+            # Extract warehouse from conn if available
+            try:
+                cur.execute("SELECT CURRENT_WAREHOUSE()")
+                row = cur.fetchone()
+                if row:
+                    preferred_wh = row[0] if isinstance(row, tuple) else row.get("CURRENT_WAREHOUSE()")
+            except:
+                pass
+            
+            wh = self._ensure_warehouse(cur, preferred_wh)
+            cur.execute(f'USE WAREHOUSE "{wh}"')
+
+            # 2) Databases (no compute requirement, but we already have compute)
+            dbs = self._list_databases(cur)
+            cur.close()
+
+            catalog = []
+            print(f"Processing {len(dbs)} databases...", file=sys.stderr)
+            
+            for db in dbs:
+                print(f"  Processing database: {db}", file=sys.stderr)
+                cur = conn.cursor(connector.DictCursor)
+                try:
+                    schemas = self._list_schemas_for_db(cur, db)
+                    print(f"    Found {len(schemas)} schemas", file=sys.stderr)
+                finally:
+                    cur.close()
+                
+                # Process schemas in parallel for this database
+                schema_objs = []
+                with ThreadPoolExecutor(max_workers=max_workers) as executor:
+                    future_to_schema = {
+                        executor.submit(self._process_schema, connector, conn, db, s): s 
+                        for s in schemas
+                    }
+                    
+                    for future in as_completed(future_to_schema):
+                        schema_name = future_to_schema[future]
+                        try:
+                            result = future.result()
+                            schema_objs.append(result)
+                            print(f"      Completed schema: {schema_name} ({len(result['tables'])} tables)", file=sys.stderr)
+                        except Exception as e:
+                            print(f"      Error with schema {schema_name}: {e}", file=sys.stderr)
+                            schema_objs.append({"schema": schema_name, "tables": [], "error": str(e)})
+                
+                catalog.append({"database": db, "schemas": schema_objs})
+
+            return {"warehouse": wh, "databases": catalog}
+        finally:
+            if not cur.is_closed():
+                cur.close()
+    
+    def _format_catalog_as_json(self, catalog: Dict) -> Dict:
+        """Format the catalog for JSON response"""
+        return catalog
+
+    @tornado.web.authenticated
+    def post(self):
+        """Get Snowflake database schema information"""
+        try:
+            # Parse request body
+            try:
+                body = json.loads(self.request.body.decode('utf-8'))
+            except json.JSONDecodeError:
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "Invalid JSON in request body"
+                }))
+                return
+            
+            # Get Snowflake configuration from request or environment
+            config = self._get_snowflake_config(body.get('config'))
+            
+            if not config:
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "No Snowflake configuration provided and no Snowflake configurations found in environment"
+                }))
+                return
+            
+            # Setup Snowflake environment
+            try:
+                connector = self._setup_snowflake_environment()
+            except ImportError as e:
+                self.set_status(500)
+                self.finish(json.dumps({
+                    "error": str(e)
+                }))
+                return
+            
+            # Get database schema using optimized catalog building
+            try:
+                conn_params = self._get_connection_params(config)
+                max_workers = int(body.get('max_workers', 5))
+                
+                print(f"[SnowflakeSchemaHandler] Connecting with account={conn_params['account']}, user={conn_params['user']}, warehouse={conn_params.get('warehouse')}, role={conn_params.get('role')}")
+                
+                connection = connector.connect(**conn_params, client_session_keep_alive=False)
+                
+                try:
+                    catalog = self._build_catalog(connector, connection, max_workers=max_workers)
+                    result = self._format_catalog_as_json(catalog)
+                    self.finish(json.dumps(result, indent=2))
+                finally:
+                    connection.close()
+                
+            except Exception as e:
+                self.set_status(500)
+                self.finish(json.dumps({
+                    "error": f"Error connecting to Snowflake: {str(e)}"
+                }))
+                
+        except Exception as e:
+            self.set_status(500)
+            self.finish(json.dumps({
+                "error": "Internal server error",
+                "message": str(e)
+            }))
+
+
+class SnowflakeQueryHandler(APIHandler):
+    """Handler for Snowflake query execution"""
+    
+    def _setup_snowflake_environment(self):
+        """Install required Snowflake packages if not available"""
+        def install_package(package_name):
+            try:
+                subprocess.check_call([sys.executable, "-m", "pip", "install", package_name])
+                return True
+            except subprocess.CalledProcessError:
+                return False
+
+        missing_packages = []
+        
+        try:
+            import snowflake.connector
+        except ImportError:
+            if install_package("snowflake-connector-python"):
+                try:
+                    import snowflake.connector
+                except ImportError as e:
+                    missing_packages.append(f"snowflake-connector-python: {str(e)}")
+            else:
+                missing_packages.append("snowflake-connector-python: installation failed")
+
+        if missing_packages:
+            raise ImportError("Required modules could not be installed: " + ", ".join(missing_packages))
+        
+        import snowflake.connector
+        return snowflake.connector
+    
+    def _get_snowflake_config(self, provided_config: Optional[Dict] = None) -> Optional[Dict]:
+        """Get Snowflake configuration from request or environment variables"""
+        if provided_config:
+            return provided_config
+        
+        # Look for Snowflake database configuration in the environment
+        for key, value in os.environ.items():
+            if key.endswith('_CONNECTION_JSON'):
+                try:
+                    config = json.loads(value)
+                    if config.get('type') == 'snowflake':
+                        return config
+                except Exception as e:
+                    print(f"[SnowflakeQueryHandler] Error parsing database config {key}: {e}")
+                    continue
+        
+        return None
+    
+    def _get_connection_params(self, config: Dict) -> Dict[str, Any]:
+        """Build Snowflake connection parameters from configuration"""
+        # Extract account from connectionUrl
+        connection_url = config.get('connectionUrl', '')
+        if not connection_url:
+            raise ValueError("connectionUrl is required for Snowflake")
+        
+        # Extract the account identifier from the connectionUrl
+        # Expected format: https://account.snowflakecomputing.com or https://account-region.snowflakecomputing.com
+        import re
+        url_match = re.match(r'https?://([^/]+)', connection_url)
+        if not url_match:
+            raise ValueError(f"Invalid Snowflake connectionUrl format: {connection_url}")
+        
+        account = url_match.group(1)
+        # Strip .snowflakecomputing.com if present
+        account_identifier = account.replace('.snowflakecomputing.com', '')
+        
+        conn_params = {
+            'account': account_identifier,
+            'user': config['username'],
+            'password': config['password'],
+        }
+        
+        warehouse = config.get('warehouse')
+        database = config.get('database')
+        role = config.get('role')
+        
+        if warehouse:
+            conn_params['warehouse'] = warehouse
+        if database:
+            conn_params['database'] = database
+        if role:
+            conn_params['role'] = role
+        
+        return conn_params
+    
+    @tornado.web.authenticated
+    def post(self):
+        """Execute a read-only SQL query on Snowflake"""
+        try:
+            # Parse request body
+            try:
+                body = json.loads(self.request.body.decode('utf-8'))
+            except json.JSONDecodeError:
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "Invalid JSON in request body"
+                }))
+                return
+            
+            # Get query from request
+            query = body.get('query')
+            if not query:
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "Missing 'query' field in request body"
+                }))
+                return
+            
+            # Basic validation for read-only queries
+            normalized_query = query.strip().upper()
+            if not normalized_query.startswith('SELECT') and not normalized_query.startswith('WITH') and not normalized_query.startswith('SHOW') and not normalized_query.startswith('DESCRIBE'):
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "Only SELECT, WITH, SHOW, or DESCRIBE statements are allowed for read queries."
+                }))
+                return
+            
+            # Get Snowflake configuration from request or environment
+            config = self._get_snowflake_config(body.get('config'))
+            
+            if not config:
+                self.set_status(400)
+                self.finish(json.dumps({
+                    "error": "No Snowflake configuration provided and no Snowflake configurations found in environment"
+                }))
+                return
+            
+            # Setup Snowflake environment
+            try:
+                connector = self._setup_snowflake_environment()
+            except ImportError as e:
+                self.set_status(500)
+                self.finish(json.dumps({
+                    "error": str(e)
+                }))
+                return
+            
+            # Execute query
+            try:
+                conn_params = self._get_connection_params(config)
+                
+                # Allow specifying a specific database for the query
+                database = body.get('database')
+                if database:
+                    conn_params['database'] = database
+                
+                # Ensure we have a warehouse for querying
+                if not conn_params.get('warehouse'):
+                    raise ValueError("A warehouse is required to execute queries.")
+                
+                connection = connector.connect(**conn_params)
+                cursor = connection.cursor()
+                
+                try:
+                    cursor.execute(query)
+                    
+                    # Get column names from cursor description
+                    columns = [desc[0] for desc in cursor.description] if cursor.description else []
+                    
+                    # Fetch all results
+                    rows = cursor.fetchall()
+                    
+                    # Convert result to list of dictionaries
+                    result_rows = [
+                        {columns[i]: row[i] for i in range(len(columns))}
+                        for row in rows
+                    ]
+                    
+                    self.finish(json.dumps({
+                        "result": result_rows
+                    }))
+                    
+                finally:
+                    cursor.close()
+                    connection.close()
+                    
+            except Exception as e:
+                self.set_status(500)
+                self.finish(json.dumps({
+                    "error": f"Snowflake query failed: {str(e)}"
+                }))
+                
+        except Exception as e:
+            self.set_status(500)
+            self.finish(json.dumps({
+                "error": "Internal server error",
+                "message": str(e)
+            }))