feat: make Backend._from_url() consistently use kwargs as overrides

Author: NickCrews

ibis/backends/bigquery/__init__.py

@@ -328,12 +328,14 @@ def read_json(
         )
         return self._read_file(path, table_name=table_name, job_config=job_config)
 
-    def _from_url(self, url: ParseResult, **kwargs):
-        return self.connect(
-            project_id=url.netloc or kwargs.get("project_id", [""])[0],
-            dataset_id=url.path[1:] or kwargs.get("dataset_id", [""])[0],
-            **kwargs,
-        )
+    def _from_url(self, url: ParseResult, **kwarg_overrides):
+        kwargs = {}
+        if url.netloc:
+            kwargs["project_id"] = url.netloc
+        if url.path:
+            kwargs["dataset_id"] = url.path[1:]
+        kwargs.update(kwarg_overrides)
+        return self.connect(**kwargs)
 
     def do_connect(
         self,

ibis/backends/clickhouse/__init__.py

@@ -69,36 +69,20 @@ def _register_in_memory_table(self, op: ops.InMemoryTable) -> None:
     def _finalize_memtable(self, name: str) -> None:
         """No-op."""
 
-    def _from_url(self, url: ParseResult, **kwargs) -> BaseBackend:
-        """Connect to a backend using a URL `url`.
-
-        Parameters
-        ----------
-        url
-            URL with which to connect to a backend.
-        kwargs
-            Additional keyword arguments
-
-        Returns
-        -------
-        BaseBackend
-            A backend instance
-
-        """
-        database = url.path[1:]
-
-        connect_args = {
-            "user": url.username,
-            "password": unquote_plus(url.password or ""),
-            "host": url.hostname,
-            "database": database or "",
-            "port": url.port,
-            **kwargs,
-        }
-
-        kwargs.update(connect_args)
+    def _from_url(self, url: ParseResult, **kwarg_overrides) -> BaseBackend:
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if url.port:
+            kwargs["port"] = url.port
+        if database := url.path[1:]:
+            kwargs["database"] = database
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
-
         return self.connect(**kwargs)
 
     def _convert_kwargs(self, kwargs):

ibis/backends/druid/__init__.py

@@ -44,35 +44,20 @@ def version(self) -> str:
             [(version,)] = result.fetchall()
         return version
 
-    def _from_url(self, url: ParseResult, **kwargs):
-        """Connect to a backend using a URL `url`.
-
-        Parameters
-        ----------
-        url
-            URL with which to connect to a backend.
-        kwargs
-            Additional keyword arguments
-
-        Returns
-        -------
-        BaseBackend
-            A backend instance
-
-        """
-        kwargs = {
-            "user": url.username,
-            "password": unquote_plus(url.password)
-            if url.password is not None
-            else None,
-            "host": url.hostname,
-            "path": url.path,
-            "port": url.port,
-            **kwargs,
-        }
-
+    def _from_url(self, url: ParseResult, **kwarg_overrides):
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if url.path:
+            kwargs["path"] = url.path
+        if url.port:
+            kwargs["port"] = url.port
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
-
         return self.connect(**kwargs)
 
     @property

ibis/backends/exasol/__init__.py

@@ -151,21 +151,20 @@ def _post_connect(self, timezone: str = "UTC") -> None:
         with self.begin() as con:
             con.execute(f"ALTER SESSION SET TIME_ZONE = {timezone!r}")
 
-    def _from_url(self, url: ParseResult, **kwargs) -> BaseBackend:
-        """Construct an ibis backend from a URL."""
-        kwargs = {
-            "user": url.username,
-            "password": unquote_plus(url.password)
-            if url.password is not None
-            else None,
-            "schema": url.path[1:] or None,
-            "host": url.hostname,
-            "port": url.port,
-            **kwargs,
-        }
-
+    def _from_url(self, url: ParseResult, **kwarg_overrides) -> BaseBackend:
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if schema := url.path[1:]:
+            kwargs["schema"] = schema
+        if url.port:
+            kwargs["port"] = url.port
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
-
         return self.connect(**kwargs)
 
     @contextlib.contextmanager

ibis/backends/impala/__init__.py

@@ -55,36 +55,41 @@ class Backend(SQLBackend, CanCreateDatabase, NoExampleLoader):
     name = "impala"
     compiler = sc.impala.compiler
 
-    def _from_url(self, url: ParseResult, **kwargs: Any) -> Backend:
-        """Connect to a backend using a URL `url`.
-
-        Parameters
-        ----------
-        url
-            URL with which to connect to a backend.
-        kwargs
-            Additional keyword arguments passed to the `connect` method.
-
-        Returns
-        -------
-        BaseBackend
-            A backend instance
-
-        """
-        for name in ("username", "hostname", "port", "password"):
-            if value := (
-                getattr(url, name, None)
-                or os.environ.get(f"{self.name.upper()}_{name.upper()}")
-            ):
-                kwargs[name] = value
-
-        with contextlib.suppress(KeyError):
-            kwargs["host"] = kwargs.pop("hostname")
-
-        (database,) = url.path[1:].split("/", 1)
-        if database:
+    def _from_url(self, url: ParseResult, **kwarg_overrides: Any) -> Backend:
+        def _get_env(attr: str) -> str | None:
+            return os.environ.get(f"{self.name.upper()}_{attr.upper()}")
+
+        kwargs = {}
+        if (username := _get_env("username")) is not None:
+            kwargs["user"] = username
+        if url.username:
+            kwargs["user"] = url.username
+
+        if (password := _get_env("password")) is not None:
+            kwargs["password"] = password
+        if url.password:
+            kwargs["password"] = url.password
+
+        if (host := _get_env("hostname")) is not None:
+            kwargs["host"] = host
+        if (host := _get_env("host")) is not None:
+            kwargs["host"] = host
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if host := kwarg_overrides.get("hostname"):
+            kwargs["host"] = host
+
+        if (port := _get_env("port")) is not None:
+            kwargs["port"] = port
+        if url.port:
+            kwargs["port"] = url.port
+
+        if (database := _get_env("path")) is not None:
+            kwargs["database"] = database
+        if database := url.path[1:].split("/", 1)[0]:
             kwargs["database"] = database
 
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
         return self.connect(**kwargs)
 

ibis/backends/mssql/__init__.py

@@ -213,38 +213,20 @@ def _post_connect(self):
         with closing(self.con.cursor()) as cur:
             cur.execute("SET DATEFIRST 1")
 
-    def _from_url(self, url: ParseResult, **kwargs):
-        database, *_ = url.path[1:].split("/", 1)
-        kwargs.update(
-            {
-                "user": url.username,
-                "password": unquote_plus(url.password or ""),
-                "host": url.hostname,
-                "database": database or "",
-                "port": url.port or None,
-            }
-        )
-
+    def _from_url(self, url: ParseResult, **kwarg_overrides):
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if url.port:
+            kwargs["port"] = url.port
+        if database := url.path[1:].split("/")[0]:
+            kwargs["database"] = database
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
-
-        if "host" in kwargs and not kwargs["host"]:
-            del kwargs["host"]
-
-        if "user" in kwargs and not kwargs["user"]:
-            del kwargs["user"]
-
-        if "password" in kwargs and kwargs["password"] is None:
-            del kwargs["password"]
-
-        if "port" in kwargs and kwargs["port"] is None:
-            del kwargs["port"]
-
-        if "database" in kwargs and not kwargs["database"]:
-            del kwargs["database"]
-
-        if "driver" in kwargs and not kwargs["driver"]:
-            del kwargs["driver"]
-
         return self.connect(**kwargs)
 
     def get_schema(

ibis/backends/mysql/__init__.py

@@ -40,49 +40,20 @@ class Backend(SQLBackend, CanCreateDatabase, PyArrowExampleLoader):
     compiler = sc.mysql.compiler
     supports_create_or_replace = False
 
-    def _from_url(self, url: ParseResult, **kwargs):
-        """Connect to a backend using a URL `url`.
-
-        Parameters
-        ----------
-        url
-            URL with which to connect to a backend.
-        kwargs
-            Additional keyword arguments
-
-        Returns
-        -------
-        BaseBackend
-            A backend instance
-
-        """
-        database, *_ = url.path[1:].split("/", 1)
-        connect_args = {
-            "user": url.username,
-            "password": unquote_plus(url.password or ""),
-            "host": url.hostname,
-            "database": database or "",
-            "port": url.port or None,
-        }
-
-        kwargs.update(connect_args)
+    def _from_url(self, url: ParseResult, **kwarg_overrides):
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if database := url.path[1:].split("/", 1)[0]:
+            kwargs["database"] = database
+        if url.port:
+            kwargs["port"] = url.port
+        kwargs.update(kwarg_overrides)
         self._convert_kwargs(kwargs)
-
-        if "user" in kwargs and not kwargs["user"]:
-            del kwargs["user"]
-
-        if "host" in kwargs and not kwargs["host"]:
-            del kwargs["host"]
-
-        if "database" in kwargs and not kwargs["database"]:
-            del kwargs["database"]
-
-        if "password" in kwargs and kwargs["password"] is None:
-            del kwargs["password"]
-
-        if "port" in kwargs and kwargs["port"] is None:
-            del kwargs["port"]
-
         return self.connect(**kwargs)
 
     @cached_property

ibis/backends/oracle/__init__.py

@@ -206,16 +206,21 @@ def _post_connect(self) -> None:
         # Set to ensure decimals come back as decimals
         oracledb.defaults.fetch_decimals = True
 
-    def _from_url(self, url: ParseResult, **kwargs):
-        self.do_connect(
-            user=url.username,
-            password=unquote_plus(url.password) if url.password is not None else None,
-            database=url.path.removeprefix("/"),
-            port=url.port,
-            **kwargs,
-        )
-
-        return self
+    def _from_url(self, url: ParseResult, **kwarg_overrides):
+        kwargs = {}
+        if url.username:
+            kwargs["user"] = url.username
+        if url.password:
+            kwargs["password"] = unquote_plus(url.password)
+        if url.hostname:
+            kwargs["host"] = url.hostname
+        if database := url.path.removeprefix("/"):
+            kwargs["database"] = database
+        if url.port:
+            kwargs["port"] = url.port
+        kwargs.update(kwarg_overrides)
+        self._convert_kwargs(kwargs)
+        return self.connect(**kwargs)
 
     @property
     def current_catalog(self) -> str: