fal 1.2.3__py3-none-any.whl → 1.3.0__py3-none-any.whl

fal/_fal_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '1.2.3'
-__version_tuple__ = version_tuple = (1, 2, 3)
+__version__ = version = '1.3.0'
+__version_tuple__ = version_tuple = (1, 3, 0)

fal/api.py

@@ -44,7 +44,13 @@ from typing_extensions import Concatenate, ParamSpec
 import fal.flags as flags
 from fal._serialization import include_modules_from, patch_pickle
 from fal.container import ContainerImage
-from fal.exceptions import FalServerlessException
+from fal.exceptions import (
+    AppException,
+    CUDAOutOfMemoryException,
+    FalServerlessException,
+    FieldException,
+)
+from fal.exceptions._cuda import _is_cuda_oom_exception
 from fal.logging.isolate import IsolateLogPrinter
 from fal.sdk import (
     FAL_SERVERLESS_DEFAULT_KEEP_ALIVE,
@@ -1002,13 +1008,39 @@ class BaseServable:
                 # If it's not a generic 404, just return the original message.
                 return JSONResponse({"detail": exc.detail}, 404)
 
+        @_app.exception_handler(AppException)
+        async def app_exception_handler(request: Request, exc: AppException):
+            return JSONResponse({"detail": exc.message}, exc.status_code)
+
+        @_app.exception_handler(FieldException)
+        async def field_exception_handler(request: Request, exc: FieldException):
+            return JSONResponse(exc.to_pydantic_format(), exc.status_code)
+
+        @_app.exception_handler(CUDAOutOfMemoryException)
+        async def cuda_out_of_memory_exception_handler(
+            request: Request, exc: CUDAOutOfMemoryException
+        ):
+            return JSONResponse({"detail": exc.message}, exc.status_code)
+
         @_app.exception_handler(Exception)
         async def traceback_logging_exception_handler(request: Request, exc: Exception):
-            print(
-                json.dumps(
-                    {"traceback": "".join(traceback.format_exception(exc)[::-1])}  # type: ignore
+            _, MINOR, *_ = sys.version_info
+
+            # traceback.format_exception() has a different signature in Python >=3.10
+            if MINOR >= 10:
+                formatted_exception = traceback.format_exception(exc)  # type: ignore
+            else:
+                formatted_exception = traceback.format_exception(
+                    type(exc), exc, exc.__traceback__
                 )
-            )
+
+            print(json.dumps({"traceback": "".join(formatted_exception[::-1])}))
+
+            if _is_cuda_oom_exception(exc):
+                return await cuda_out_of_memory_exception_handler(
+                    request, CUDAOutOfMemoryException()
+                )
+
             return JSONResponse({"detail": "Internal Server Error"}, 500)
 
         routes = self.collect_routes()

fal/app.py

@@ -3,7 +3,9 @@ from __future__ import annotations
 import inspect
 import json
 import os
+import queue
 import re
+import threading
 import time
 import typing
 from contextlib import asynccontextmanager, contextmanager
@@ -72,17 +74,22 @@ def wrap_app(cls: type[App], **kwargs) -> fal.api.IsolatedFunction:
 
 
 class EndpointClient:
-    def __init__(self, url, endpoint, signature):
+    def __init__(self, url, endpoint, signature, timeout: int | None = None):
         self.url = url
         self.endpoint = endpoint
         self.signature = signature
+        self.timeout = timeout
 
         annotations = endpoint.__annotations__ or {}
         self.return_type = annotations.get("return") or None
 
     def __call__(self, data):
         with httpx.Client() as client:
-            resp = client.post(self.url + self.signature.path, json=dict(data))
+            resp = client.post(
+                self.url + self.signature.path,
+                json=data.dict() if hasattr(data, "dict") else dict(data),
+                timeout=self.timeout,
+            )
             resp.raise_for_status()
             resp_dict = resp.json()
 
@@ -93,7 +100,12 @@ class EndpointClient:
 
 
 class AppClient:
-    def __init__(self, cls, url):
+    def __init__(
+        self,
+        cls,
+        url,
+        timeout: int | None = None,
+    ):
         self.url = url
         self.cls = cls
 
@@ -101,29 +113,50 @@ class AppClient:
             signature = getattr(endpoint, "route_signature", None)
             if signature is None:
                 continue
-
-            setattr(self, name, EndpointClient(self.url, endpoint, signature))
+            endpoint_client = EndpointClient(
+                self.url,
+                endpoint,
+                signature,
+                timeout=timeout,
+            )
+            setattr(self, name, endpoint_client)
 
     @classmethod
     @contextmanager
     def connect(cls, app_cls):
         app = wrap_app(app_cls)
         info = app.spawn()
+        _shutdown_event = threading.Event()
+
+        def _print_logs():
+            while not _shutdown_event.is_set():
+                try:
+                    log = info.logs.get(timeout=0.1)
+                except queue.Empty:
+                    continue
+                print(log)
+
+        _log_printer = threading.Thread(target=_print_logs, daemon=True)
+        _log_printer.start()
+
         try:
             with httpx.Client() as client:
                 retries = 100
-                while retries:
+                for _ in range(retries):
                     resp = client.get(info.url + "/health")
+
                     if resp.is_success:
                         break
-                    elif resp.status_code != 500:
+                    elif resp.status_code not in (500, 404):
                         resp.raise_for_status()
                     time.sleep(0.1)
-                    retries -= 1
 
-            yield cls(app_cls, info.url)
+            client = cls(app_cls, info.url)
+            yield client
         finally:
             info.stream.cancel()
+            _shutdown_event.set()
+            _log_printer.join()
 
     def health(self):
         with httpx.Client() as client:

fal/exceptions/__init__.py

@@ -1,3 +1,4 @@
 from __future__ import annotations
 
-from ._base import FalServerlessException  # noqa: F401
+from ._base import AppException, FalServerlessException, FieldException  # noqa: F401
+from ._cuda import CUDAOutOfMemoryException  # noqa: F401

fal/exceptions/_base.py

@@ -1,7 +1,50 @@
 from __future__ import annotations
 
+from dataclasses import dataclass
+from typing import Sequence
+
 
 class FalServerlessException(Exception):
     """Base exception type for fal Serverless related flows and APIs."""
 
     pass
+
+
+@dataclass
+class AppException(FalServerlessException):
+    """
+    Base exception class for application-specific errors.
+
+    Attributes:
+        message: A descriptive message explaining the error.
+        status_code: The HTTP status code associated with the error.
+    """
+
+    message: str
+    status_code: int
+
+
+@dataclass
+class FieldException(FalServerlessException):
+    """Exception raised for errors related to specific fields.
+
+    Attributes:
+        field: The field that caused the error.
+        message: A descriptive message explaining the error.
+        status_code: The HTTP status code associated with the error. Defaults to 422
+        type: The type of error. Defaults to "value_error"
+    """
+
+    field: str
+    message: str
+    status_code: int = 422
+    type: str = "value_error"
+
+    def to_pydantic_format(self) -> Sequence[dict]:
+        return [
+            {
+                "loc": ["body", self.field],
+                "msg": self.message,
+                "type": self.type,
+            }
+        ]

fal/exceptions/_cuda.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+from ._base import AppException
+
+# PyTorch error message for out of memory
+_CUDA_OOM_MESSAGE = "CUDA error: out of memory"
+
+# Special status code for CUDA out of memory errors
+_CUDA_OOM_STATUS_CODE = 503
+
+
+@dataclass
+class CUDAOutOfMemoryException(AppException):
+    """Exception raised when a CUDA operation runs out of memory."""
+
+    message: str = _CUDA_OOM_MESSAGE
+    status_code: int = _CUDA_OOM_STATUS_CODE
+
+
+# based on https://github.com/Lightning-AI/pytorch-lightning/blob/37e04d075a5532c69b8ac7457795b4345cca30cc/src/lightning/pytorch/utilities/memory.py#L49
+def _is_cuda_oom_exception(exception: BaseException) -> bool:
+    return _is_cuda_out_of_memory(exception) or _is_cudnn_snafu(exception)
+
+
+# based on https://github.com/BlackHC/toma/blob/master/toma/torch_cuda_memory.py
+def _is_cuda_out_of_memory(exception: BaseException) -> bool:
+    return (
+        isinstance(exception, RuntimeError)
+        and len(exception.args) == 1
+        and "CUDA" in exception.args[0]
+        and "out of memory" in exception.args[0]
+    )
+
+
+# based on https://github.com/BlackHC/toma/blob/master/toma/torch_cuda_memory.py
+def _is_cudnn_snafu(exception: BaseException) -> bool:
+    # For/because of https://github.com/pytorch/pytorch/issues/4107
+    return (
+        isinstance(exception, RuntimeError)
+        and len(exception.args) == 1
+        and "cuDNN error: CUDNN_STATUS_NOT_SUPPORTED." in exception.args[0]
+    )

{fal-1.2.3.dist-info → fal-1.3.0.dist-info}/METADATA

@@ -1,50 +1,50 @@
 Metadata-Version: 2.1
 Name: fal
-Version: 1.2.3
+Version: 1.3.0
 Summary: fal is an easy-to-use Serverless Python Framework
 Author: Features & Labels <support@fal.ai>
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
-Requires-Dist: isolate[build] <1.14.0,>=0.13.0
-Requires-Dist: isolate-proto ==0.5.1
-Requires-Dist: grpcio ==1.64.0
-Requires-Dist: dill ==0.3.7
-Requires-Dist: cloudpickle ==3.0.0
-Requires-Dist: typing-extensions <5,>=4.7.1
-Requires-Dist: click <9,>=8.1.3
-Requires-Dist: structlog <23,>=22.3.0
-Requires-Dist: opentelemetry-api <2,>=1.15.0
-Requires-Dist: opentelemetry-sdk <2,>=1.15.0
-Requires-Dist: grpc-interceptor <1,>=0.15.0
-Requires-Dist: colorama <1,>=0.4.6
-Requires-Dist: portalocker <3,>=2.7.0
-Requires-Dist: rich <14,>=13.3.2
+Requires-Dist: isolate[build]<1.14.0,>=0.13.0
+Requires-Dist: isolate-proto==0.5.1
+Requires-Dist: grpcio==1.64.0
+Requires-Dist: dill==0.3.7
+Requires-Dist: cloudpickle==3.0.0
+Requires-Dist: typing-extensions<5,>=4.7.1
+Requires-Dist: click<9,>=8.1.3
+Requires-Dist: structlog<23,>=22.3.0
+Requires-Dist: opentelemetry-api<2,>=1.15.0
+Requires-Dist: opentelemetry-sdk<2,>=1.15.0
+Requires-Dist: grpc-interceptor<1,>=0.15.0
+Requires-Dist: colorama<1,>=0.4.6
+Requires-Dist: portalocker<3,>=2.7.0
+Requires-Dist: rich<14,>=13.3.2
 Requires-Dist: rich-argparse
-Requires-Dist: packaging >=21.3
-Requires-Dist: pathspec <1,>=0.11.1
-Requires-Dist: pydantic !=2.0.*,!=2.1.*,!=2.2.*,!=2.3.*,!=2.4.*,<3
-Requires-Dist: fastapi <1,>=0.99.1
-Requires-Dist: starlette-exporter >=0.21.0
-Requires-Dist: httpx >=0.15.4
-Requires-Dist: attrs >=21.3.0
-Requires-Dist: python-dateutil <3,>=2.8.0
-Requires-Dist: types-python-dateutil <3,>=2.8.0
-Requires-Dist: msgpack <2,>=1.0.7
-Requires-Dist: websockets <13,>=12.0
-Requires-Dist: pillow <11,>=10.2.0
-Requires-Dist: pyjwt[crypto] <3,>=2.8.0
-Requires-Dist: uvicorn <1,>=0.29.0
+Requires-Dist: packaging>=21.3
+Requires-Dist: pathspec<1,>=0.11.1
+Requires-Dist: pydantic!=2.0.*,!=2.1.*,!=2.2.*,!=2.3.*,!=2.4.*,<3
+Requires-Dist: fastapi<1,>=0.99.1
+Requires-Dist: starlette-exporter>=0.21.0
+Requires-Dist: httpx>=0.15.4
+Requires-Dist: attrs>=21.3.0
+Requires-Dist: python-dateutil<3,>=2.8.0
+Requires-Dist: types-python-dateutil<3,>=2.8.0
+Requires-Dist: msgpack<2,>=1.0.7
+Requires-Dist: websockets<13,>=12.0
+Requires-Dist: pillow<11,>=10.2.0
+Requires-Dist: pyjwt[crypto]<3,>=2.8.0
+Requires-Dist: uvicorn<1,>=0.29.0
 Requires-Dist: cookiecutter
-Requires-Dist: importlib-metadata >=4.4 ; python_version < "3.10"
+Requires-Dist: importlib-metadata>=4.4; python_version < "3.10"
 Provides-Extra: dev
-Requires-Dist: fal[test] ; extra == 'dev'
-Requires-Dist: openapi-python-client <1,>=0.14.1 ; extra == 'dev'
+Requires-Dist: fal[test]; extra == "dev"
+Requires-Dist: openapi-python-client<1,>=0.14.1; extra == "dev"
 Provides-Extra: test
-Requires-Dist: pytest <8 ; extra == 'test'
-Requires-Dist: pytest-asyncio ; extra == 'test'
-Requires-Dist: pytest-xdist ; extra == 'test'
-Requires-Dist: flaky ; extra == 'test'
-Requires-Dist: boto3 ; extra == 'test'
+Requires-Dist: pytest<8; extra == "test"
+Requires-Dist: pytest-asyncio; extra == "test"
+Requires-Dist: pytest-xdist; extra == "test"
+Requires-Dist: flaky; extra == "test"
+Requires-Dist: boto3; extra == "test"
 
 [![PyPI](https://img.shields.io/pypi/v/fal.svg?logo=PyPI)](https://pypi.org/project/fal)
 [![Tests](https://img.shields.io/github/actions/workflow/status/fal-ai/fal/integration_tests.yaml?label=Tests)](https://github.com/fal-ai/fal/actions)

{fal-1.2.3.dist-info → fal-1.3.0.dist-info}/RECORD

@@ -1,10 +1,10 @@
 fal/__init__.py,sha256=wXs1G0gSc7ZK60-bHe-B2m0l_sA6TrFk4BxY0tMoLe8,784
 fal/__main__.py,sha256=MSmt_5Xg84uHqzTN38JwgseJK8rsJn_11A8WD99VtEo,61
-fal/_fal_version.py,sha256=VUBVzEBuW57s195P22MF-rLGH7GWx3G_z5nV-l_8MBE,411
+fal/_fal_version.py,sha256=HGwtpza1HCPtlyqElUvIyH97K44TO13CYiYVZNezQ1M,411
 fal/_serialization.py,sha256=rD2YiSa8iuzCaZohZwN_MPEB-PpSKbWRDeaIDpTEjyY,7653
 fal/_version.py,sha256=EBGqrknaf1WygENX-H4fBefLvHryvJBBGtVJetaB0NY,266
-fal/api.py,sha256=LAPl5Hf6ZWzEjv4lFUtsisWgrnXH_qNUHdJrEHT_A5Y,40602
-fal/app.py,sha256=9HJGu_64ArtW8W91BC0U4Etr2gA31LXaHgR6HzoOops,15903
+fal/api.py,sha256=bOCxmOQSbdcR6h6VLPEuvsD4i0j_Mod6E2UNP07cAQo,41893
+fal/app.py,sha256=PGx-6Zr4evqe4Fzs4g4-MxKnaOp_7hW5G7vU1PpPvkI,16800
 fal/apps.py,sha256=FrKmaAUo8U9vE_fcva0GQvk4sCrzaTEr62lGtu3Ld5M,6825
 fal/container.py,sha256=V7riyyq8AZGwEX9QaqRQDZyDN_bUKeRKV1OOZArXjL0,622
 fal/flags.py,sha256=oWN_eidSUOcE9wdPK_77si3A1fpgOC0UEERPsvNLIMc,842
@@ -32,8 +32,9 @@ fal/cli/secrets.py,sha256=740msFm7d41HruudlcfqUXlFl53N-WmChsQP9B9M9Po,2572
 fal/console/__init__.py,sha256=ernZ4bzvvliQh5SmrEqQ7lA5eVcbw6Ra2jalKtA7dxg,132
 fal/console/icons.py,sha256=De9MfFaSkO2Lqfne13n3PrYfTXJVIzYZVqYn5BWsdrA,108
 fal/console/ux.py,sha256=KMQs3UHQvVHDxDQQqlot-WskVKoMQXOE3jiVkkfmIMY,356
-fal/exceptions/__init__.py,sha256=x3fp97qMr5zCQJghMq6k2ESXWSrkWumO1BZebh3pWsI,92
-fal/exceptions/_base.py,sha256=oF2XfitbiDGObmSF1IX50uAdV8IUvOfR-YsGmMQSE0A,161
+fal/exceptions/__init__.py,sha256=5b2rFnvLf2Q4fPGedcYN_JVLv-j6iOy5DNG7jB6FNUk,180
+fal/exceptions/_base.py,sha256=zFiiRceThgdnR8DEYtzpkq-IDO_D0EHGLd0oZjX5Gqk,1256
+fal/exceptions/_cuda.py,sha256=q5EPFYEb7Iyw03cHrQlRHnH5xOvjwTwQdM6a9N3GB8k,1494
 fal/exceptions/auth.py,sha256=gxRago5coI__vSIcdcsqhhq1lRPkvCnwPAueIaXTAdw,329
 fal/logging/__init__.py,sha256=snqprf7-sKw6oAATS_Yxklf-a3XhLg0vIHICPwLp6TM,1583
 fal/logging/isolate.py,sha256=jJSgDHkFg4sB0xElYSqCYF6IAxy6jEgSfjwFuKJIZbA,2305
@@ -122,8 +123,8 @@ openapi_fal_rest/models/workflow_node_type.py,sha256=-FzyeY2bxcNmizKbJI8joG7byRi
 openapi_fal_rest/models/workflow_schema.py,sha256=4K5gsv9u9pxx2ItkffoyHeNjBBYf6ur5bN4m_zePZNY,2019
 openapi_fal_rest/models/workflow_schema_input.py,sha256=2OkOXWHTNsCXHWS6EGDFzcJKkW5FIap-2gfO233EvZQ,1191
 openapi_fal_rest/models/workflow_schema_output.py,sha256=EblwSPAGfWfYVWw_WSSaBzQVju296is9o28rMBAd0mc,1196
-fal-1.2.3.dist-info/METADATA,sha256=tUsGGaLLvqfYpEePodE0fuc5nm_XSoYE7AWTrci2nEY,3805
-fal-1.2.3.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
-fal-1.2.3.dist-info/entry_points.txt,sha256=32zwTUC1U1E7nSTIGCoANQOQ3I7-qHG5wI6gsVz5pNU,37
-fal-1.2.3.dist-info/top_level.txt,sha256=r257X1L57oJL8_lM0tRrfGuXFwm66i1huwQygbpLmHw,21
-fal-1.2.3.dist-info/RECORD,,
+fal-1.3.0.dist-info/METADATA,sha256=sP1Mdzh8ciPGfXFrY_GzmRm8IH0Ew6gv9IUjBsHWCe4,3766
+fal-1.3.0.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
+fal-1.3.0.dist-info/entry_points.txt,sha256=32zwTUC1U1E7nSTIGCoANQOQ3I7-qHG5wI6gsVz5pNU,37
+fal-1.3.0.dist-info/top_level.txt,sha256=r257X1L57oJL8_lM0tRrfGuXFwm66i1huwQygbpLmHw,21
+fal-1.3.0.dist-info/RECORD,,