fal 1.27.1__tar.gz → 1.28.1__tar.gz

{fal-1.27.1/fal.egg-info → fal-1.28.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fal
-Version: 1.27.1
+Version: 1.28.1
 Summary: fal is an easy-to-use Serverless Python Framework
 Author: Features & Labels <support@fal.ai>
 Requires-Python: >=3.8

{fal-1.27.1 → fal-1.28.1/fal.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fal
-Version: 1.27.1
+Version: 1.28.1
 Summary: fal is an easy-to-use Serverless Python Framework
 Author: Features & Labels <support@fal.ai>
 Requires-Python: >=3.8

{fal-1.27.1 → fal-1.28.1}/src/fal/_fal_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '1.27.1'
-__version_tuple__ = version_tuple = (1, 27, 1)
+__version__ = version = '1.28.1'
+__version_tuple__ = version_tuple = (1, 28, 1)

{fal-1.27.1 → fal-1.28.1}/src/fal/api.py

@@ -953,6 +953,11 @@ def function(  # type: ignore
 ):
     if host is None:
         host = FalServerlessHost()
+
+    # NOTE: assuming kind="container" if image is provided
+    if config.get("image"):
+        kind = "container"
+
     options = host.parse_options(kind=kind, **config)
 
     def wrapper(func: Callable[ArgsT, ReturnT]):

{fal-1.27.1 → fal-1.28.1}/src/fal/app.py

@@ -411,28 +411,31 @@ class App(BaseServable):
                 )
             return response
 
-        @app.middleware("http")
-        async def set_global_object_preference(request, call_next):
-            try:
-                preference_dict = request_lifecycle_preference(request)
-                if preference_dict is not None:
-                    # This will not work properly for apps with multiplexing enabled
-                    # we may mix up the preferences between requests
-                    LIFECYCLE_PREFERENCE.set(preference_dict)
-            except Exception:
-                from fastapi.logger import logger
-
-                logger.exception(
-                    "Failed set a global lifecycle preference %s",
-                    self.__class__.__name__,
-                )
+        multiplexing = self.host_kwargs.get("max_multiplexing") or 1
+        if multiplexing == 1:
+            # just register the middleware if we are not multiplexing
+            @app.middleware("http")
+            async def set_global_object_preference(request, call_next):
+                try:
+                    preference_dict = request_lifecycle_preference(request)
+                    if preference_dict is not None:
+                        # This will not work properly for apps with multiplexing enabled
+                        # we may mix up the preferences between requests
+                        LIFECYCLE_PREFERENCE.set(preference_dict)
+                except Exception:
+                    from fastapi.logger import logger
+
+                    logger.exception(
+                        "Failed set a global lifecycle preference %s",
+                        self.__class__.__name__,
+                    )
 
-            try:
-                return await call_next(request)
-            finally:
-                # We may miss the global preference if there are operations
-                # being done in the background that go beyond the request
-                LIFECYCLE_PREFERENCE.set(None)
+                try:
+                    return await call_next(request)
+                finally:
+                    # We may miss the global preference if there are operations
+                    # being done in the background that go beyond the request
+                    LIFECYCLE_PREFERENCE.set(None)
 
         @app.middleware("http")
         async def set_request_id(request, call_next):

{fal-1.27.1 → fal-1.28.1}/src/fal/cli/_utils.py

@@ -12,7 +12,7 @@ def get_client(host: str, team: str | None = None):
 
 def is_app_name(app_ref: tuple[str, str | None]) -> bool:
     is_single_file = app_ref[1] is None
-    is_python_file = app_ref[0].endswith(".py")
+    is_python_file = app_ref[0] is None or app_ref[0].endswith(".py")
 
     return is_single_file and not is_python_file
 
@@ -42,9 +42,16 @@ def get_app_data_from_toml(app_name):
 
     app_auth = app_data.pop("auth", "private")
     app_deployment_strategy = app_data.pop("deployment_strategy", "recreate")
-    app_no_scale = app_data.pop("no_scale", False)
+
+    if "no_scale" in app_data:
+        # Deprecated
+        app_no_scale = app_data.pop("no_scale")
+        print("[WARNING] no_scale is deprecated, use app_scale_settings instead")
+        app_reset_scale = not app_no_scale
+    else:
+        app_reset_scale = app_data.pop("app_scale_settings", False)
 
     if len(app_data) > 0:
         raise ValueError(f"Found unexpected keys in pyproject.toml: {app_data}")
 
-    return app_ref, app_auth, app_deployment_strategy, app_no_scale
+    return app_ref, app_auth, app_deployment_strategy, app_reset_scale

{fal-1.27.1 → fal-1.28.1}/src/fal/cli/deploy.py

@@ -66,9 +66,9 @@ def _deploy_from_reference(
     app_ref: Tuple[Optional[Union[Path, str]], ...],
     app_name: str,
     args,
-    auth: Optional[Literal["public", "shared", "private"]] = None,
-    deployment_strategy: Optional[Literal["recreate", "rolling"]] = None,
-    scale: bool = True,
+    auth: Optional[Literal["public", "shared", "private"]],
+    deployment_strategy: Optional[Literal["recreate", "rolling"]],
+    scale: bool,
 ):
     from fal.api import FalServerlessError, FalServerlessHost
     from fal.utils import load_function_from
@@ -146,7 +146,7 @@ def _deploy(args):
             raise ValueError("Cannot use --app-name or --auth with app name reference.")
 
         app_name = args.app_ref[0]
-        app_ref, app_auth, app_deployment_strategy, app_no_scale = (
+        app_ref, app_auth, app_deployment_strategy, app_scale_settings = (
             get_app_data_from_toml(app_name)
         )
         file_path, func_name = RefAction.split_ref(app_ref)
@@ -157,7 +157,7 @@ def _deploy(args):
         app_name = args.app_name
         app_auth = args.auth
         app_deployment_strategy = args.strategy
-        app_no_scale = args.no_scale
+        app_scale_settings = args.app_scale_settings
 
     _deploy_from_reference(
         (file_path, func_name),
@@ -165,7 +165,7 @@ def _deploy(args):
         args,
         app_auth,
         app_deployment_strategy,
-        scale=not app_no_scale,
+        scale=app_scale_settings,
     )
 
 
@@ -233,12 +233,17 @@ def add_parser(main_subparsers, parents):
     )
     parser.add_argument(
         "--no-scale",
+        action="store_false",
+        dest="app_scale_settings",
+        default=False,
+        help="Use the previous deployment of the application for scale settings. "
+        "This is the default behavior.",
+    )
+    parser.add_argument(
+        "--reset-scale",
         action="store_true",
-        help=(
-            "Use min_concurrency/max_concurrency/concurrency_buffer/max_multiplexing "
-            "from previous deployment of application with this name, if exists. "
-            "Otherwise will use the values from the application code."
-        ),
+        dest="app_scale_settings",
+        help="Use the application code for scale settings.",
     )
 
     parser.set_defaults(func=_deploy)

{fal-1.27.1 → fal-1.28.1}/src/fal/cli/profile.py

@@ -68,12 +68,17 @@ def _unset(args, config: Config | None = None):
 def _key_set(args):
     config = Config(validate_profile=True)
 
+    args.console.print("Go to https://fal.ai/dashboard/keys to create a key.")
+    args.console.print(
+        "Enter the key in the format [cyan][bold]id:secret[/][/]. You can copy and paste it from the browser."  # noqa: E501
+    )
+
     while True:
         key = input("Enter the key: ")
         if ":" in key:
             break
         args.console.print(
-            "[red]Invalid key. The key must be in the format [bold]key:value[/].[/]"
+            "[yellow]Invalid key. The key must be in the format:[/] [cyan][bold]id:secret[/][/]."  # noqa: E501
         )
 
     with config.edit():
@@ -99,10 +104,10 @@ def _create(args):
     with config.edit() as config:
         config._config[args.PROFILE] = {}
 
-    args.console.print(f"Profile [cyan]{args.PROFILE}[/] created.")
     args.console.print(
-        f"Use [bold]fal profile set {args.PROFILE}[/] to set it as default."
+        f"Profile [cyan]{args.PROFILE}[/] created. Setting it as default profile..."
     )
+    _set(args)
 
 
 def _delete(args):
@@ -184,7 +189,7 @@ def add_parser(main_subparsers, parents):
     )
     host_set_parser.set_defaults(func=_host_set)
 
-    create_help = "Create a new profile. Use 'fal profile set <name>' to set it as default after creation."  # noqa: E501
+    create_help = "Create a new profile."
     create_parser = subparsers.add_parser(
         "create",
         description=create_help,

{fal-1.27.1 → fal-1.28.1}/src/fal/config.py

@@ -6,7 +6,9 @@ from typing import Dict, Iterator, List, Optional
 
 SETTINGS_SECTION = "__internal__"
 
-NO_PROFILE_ERROR = ValueError("No profile set.")
+NO_PROFILE_ERROR = ValueError(
+    "No profile set. Use command [bold]fal profile set[/] to set a profile."
+)
 
 
 class Config:

{fal-1.27.1 → fal-1.28.1}/src/fal/exceptions/auth.py

@@ -7,5 +7,5 @@ class UnauthenticatedException(FalServerlessException):
     def __init__(self) -> None:
         super().__init__(
             "You must be authenticated. "
-            "Login via `fal auth login` or make sure to setup fal keys correctly."
+            "Use [bold]fal auth login[/] or [bold]fal profile key[/] to set your fal key."  # noqa: E501
         )

{fal-1.27.1 → fal-1.28.1}/src/fal/toolkit/file/file.py

@@ -201,9 +201,11 @@ class File(BaseModel):
 
         fdata = FileData(data, content_type, file_name)
 
-        object_lifecycle_preference = (
-            request_lifecycle_preference(request) or LIFECYCLE_PREFERENCE.get()
-        )
+        if request:
+            object_lifecycle_preference = request_lifecycle_preference(request)
+        else:
+            object_lifecycle_preference = LIFECYCLE_PREFERENCE.get()
+
         save_kwargs.setdefault(
             "object_lifecycle_preference", object_lifecycle_preference
         )
@@ -250,9 +252,12 @@ class File(BaseModel):
         fallback_save_kwargs = fallback_save_kwargs or {}
 
         content_type = content_type or "application/octet-stream"
-        object_lifecycle_preference = (
-            request_lifecycle_preference(request) or LIFECYCLE_PREFERENCE.get()
-        )
+
+        if request:
+            object_lifecycle_preference = request_lifecycle_preference(request)
+        else:
+            object_lifecycle_preference = LIFECYCLE_PREFERENCE.get()
+
         save_kwargs.setdefault(
             "object_lifecycle_preference", object_lifecycle_preference
         )
@@ -332,12 +337,9 @@ class CompressedFile(File):
             shutil.rmtree(self.extract_dir)
 
 
-def request_lifecycle_preference(request: Optional[Request]) -> dict[str, str] | None:
+def request_lifecycle_preference(request: Request) -> dict[str, str] | None:
     import json
 
-    if request is None:
-        return None
-
     preference_str = request.headers.get(OBJECT_LIFECYCLE_PREFERENCE_KEY)
     if preference_str is None:
         return None

{fal-1.27.1 → fal-1.28.1}/tests/cli/test_deploy.py

@@ -39,7 +39,7 @@ def mock_args(
     app_name: Optional[str] = None,
     auth: Optional[str] = None,
     strategy: Optional[str] = None,
-    no_scale: bool = False,
+    reset_scale: bool = False,
 ):
     args = MagicMock()
 
@@ -47,8 +47,7 @@ def mock_args(
     args.app_name = app_name
     args.auth = auth
     args.strategy = strategy
-
-    args.no_scale = no_scale
+    args.app_scale_settings = reset_scale
 
     return args
 
@@ -75,7 +74,7 @@ def test_deploy_with_toml_success(
         args,
         "shared",
         "rolling",
-        scale=True,
+        scale=False,
     )
 
 
@@ -101,7 +100,7 @@ def test_deploy_with_toml_no_auth(
         args,
         "private",
         "recreate",
-        scale=True,
+        scale=False,
     )
 
 
@@ -212,7 +211,7 @@ def test_deploy_with_toml_deployment_strategy(
         args,
         "shared",
         "rolling",
-        scale=True,
+        scale=False,
     )
 
 
@@ -236,7 +235,7 @@ def test_deploy_with_toml_default_deployment_strategy(
         args,
         "private",
         "recreate",
-        scale=True,
+        scale=False,
     )
 
 
@@ -258,7 +257,7 @@ def test_deploy_with_cli_auth(
         args,
         "shared",
         None,
-        scale=True,
+        scale=False,
     )
 
 
@@ -280,19 +279,19 @@ def test_deploy_with_cli_deployment_strategy(
         args,
         None,
         "rolling",
-        scale=True,
+        scale=False,
     )
 
 
 @patch("fal.cli._utils.find_pyproject_toml", return_value="pyproject.toml")
 @patch("fal.cli._utils.parse_pyproject_toml")
 @patch("fal.cli.deploy._deploy_from_reference")
-def test_deploy_with_cli_no_scale(
+def test_deploy_with_cli_reset_scale(
     mock_deploy_ref, mock_parse_toml, mock_find_toml, mock_parse_pyproject_toml
 ):
     mock_parse_toml.return_value = mock_parse_pyproject_toml
 
-    args = mock_args(app_ref=("src/my_app/inference.py", "MyApp"), no_scale=True)
+    args = mock_args(app_ref=("src/my_app/inference.py", "MyApp"), reset_scale=True)
 
     _deploy(args)
 
@@ -302,7 +301,7 @@ def test_deploy_with_cli_no_scale(
         args,
         None,
         None,
-        scale=False,
+        scale=True,
     )
 
 
@@ -324,5 +323,5 @@ def test_deploy_with_cli_scale(
         args,
         None,
         None,
-        scale=True,
+        scale=False,
     )