torchx-nightly 2024.6.17__py3-none-any.whl → 2024.6.20__py3-none-any.whl

torchx/specs/api.py

@@ -10,6 +10,7 @@
 import copy
 import json
 import re
+import typing
 from dataclasses import asdict, dataclass, field
 from datetime import datetime
 from enum import Enum
@@ -189,8 +190,25 @@ class macros:
             role = copy.deepcopy(role)
             role.args = [self.substitute(arg) for arg in role.args]
             role.env = {key: self.substitute(arg) for key, arg in role.env.items()}
+            role.metadata = self._apply_nested(role.metadata)
+
             return role
 
+        def _apply_nested(self, d: typing.Dict[str, Any]) -> typing.Dict[str, Any]:
+            stack = [d]
+            while stack:
+                current_dict = stack.pop()
+                for k, v in current_dict.items():
+                    if isinstance(v, dict):
+                        stack.append(v)
+                    elif isinstance(v, str):
+                        current_dict[k] = self.substitute(v)
+                    elif isinstance(v, list):
+                        for i in range(len(v)):
+                            if isinstance(v[i], str):
+                                v[i] = self.substitute(v[i])
+            return d
+
         def substitute(self, arg: str) -> str:
             """
             substitute applies the values to the template arg.
@@ -219,11 +237,15 @@ class RetryPolicy(str, Enum):
                 application to deal with failed replica departures and
                 replacement replica admittance.
     2. APPLICATION: Restarts the entire application.
-
+    3. HOT_SPARE: Restarts the replicas for a role as long as quorum (min_replicas)
+                is not violated using extra hosts as spares. It does not really support
+                elasticity and just uses the delta between num_replicas and min_replicas
+                as spares (EXPERIMENTAL).
     """
 
     REPLICA = "REPLICA"
     APPLICATION = "APPLICATION"
+    HOT_SPARE = "HOT_SPARE"
 
 
 class MountType(str, Enum):
@@ -322,6 +344,8 @@ class Role:
                 and num_replicas depending on the cluster resources and
                 policies. If the scheduler doesn't support auto scaling this
                 field is ignored and the job size will be num_replicas.
+                EXPERIMENTAL: For HOT_SPARE restart policy this field is used to
+                indicate the quorum required for the job to run.
             max_retries: max number of retries before giving up
             retry_policy: retry behavior upon replica failures
             resource: Resource requirement for the role. The role should be scheduled

{torchx_nightly-2024.6.17.dist-info → torchx_nightly-2024.6.20.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: torchx-nightly
-Version: 2024.6.17
+Version: 2024.6.20
 Summary: TorchX SDK and Components
 Home-page: https://github.com/pytorch/torchx
 Author: TorchX Devs

{torchx_nightly-2024.6.17.dist-info → torchx_nightly-2024.6.20.dist-info}/RECORD

@@ -83,7 +83,7 @@ torchx/schedulers/ray/__init__.py,sha256=fE0IHi1JJpxsNVBNzWNee2thrNXFFRhY94c80Rx
 torchx/schedulers/ray/ray_common.py,sha256=pyNYFvTKVwdjDAeCBNbPwAWwVNmlLOJWExfn90XY8u8,610
 torchx/schedulers/ray/ray_driver.py,sha256=Wl-1jldL8veVKzmYDEeR2va3JSlAjZpFE1h8HWE9YVE,12286
 torchx/specs/__init__.py,sha256=vF-WUu_4NZP30lCtNYg0YVenY6wRQ8k7K36fOxqbOKc,5477
-torchx/specs/api.py,sha256=73jJOYQ-H-jX3Yvbc9hJjolavNQdF8bW9Q0oQnWzLL0,34885
+torchx/specs/api.py,sha256=Y5uT7a-qZ4pP4kSfS6yYP1jUzQWLPI4qLQeyuBv5uDQ,36085
 torchx/specs/builders.py,sha256=QDcQrnCO4bdSaiP0216XbCgTsnLutO_1_FW5jDiEIWI,9939
 torchx/specs/file_linter.py,sha256=IeiomB1BgHUlT-ZsvGxar3llY63NOupfLBrOrD_---A,11860
 torchx/specs/finder.py,sha256=MnwxG_UC4a-3X2wQ37ANEQR6D1TvriCLyuVYBh_-wuI,16249
@@ -113,9 +113,9 @@ torchx/workspace/__init__.py,sha256=FqN8AN4VhR1C_SBY10MggQvNZmyanbbuPuE-JCjkyUY,
 torchx/workspace/api.py,sha256=1heBmPgB-W5Zf9gwViM7NrqvHpZlVYeMN7jpY8Qkytc,5479
 torchx/workspace/dir_workspace.py,sha256=npNW_IjUZm_yS5r-8hrRkH46ndDd9a_eApT64m1S1T4,2268
 torchx/workspace/docker_workspace.py,sha256=PFu2KQNVC-0p2aKJ-W_BKA9ZOmXdCY2ABEkCExp3udQ,10269
-torchx_nightly-2024.6.17.dist-info/LICENSE,sha256=WVHfXhFC0Ia8LTKt_nJVYobdqTJVg_4J3Crrfm2A8KQ,1721
-torchx_nightly-2024.6.17.dist-info/METADATA,sha256=EwU_RjFfOJFS4IlszEZ1cDndcgmG44C7bNdTMe3mBqM,6184
-torchx_nightly-2024.6.17.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-torchx_nightly-2024.6.17.dist-info/entry_points.txt,sha256=3JYZFlX9aWzR-Gs_qsx1zq7mlqbFz6Mi9rQUULW8caI,170
-torchx_nightly-2024.6.17.dist-info/top_level.txt,sha256=pxew3bc2gsiViS0zADs0jb6kC5v8o_Yy_85fhHj_J1A,7
-torchx_nightly-2024.6.17.dist-info/RECORD,,
+torchx_nightly-2024.6.20.dist-info/LICENSE,sha256=WVHfXhFC0Ia8LTKt_nJVYobdqTJVg_4J3Crrfm2A8KQ,1721
+torchx_nightly-2024.6.20.dist-info/METADATA,sha256=KkAPJrmZf_3EXXqqfBmrIyZEVK3Alrl2xN02o5QTVYs,6184
+torchx_nightly-2024.6.20.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+torchx_nightly-2024.6.20.dist-info/entry_points.txt,sha256=3JYZFlX9aWzR-Gs_qsx1zq7mlqbFz6Mi9rQUULW8caI,170
+torchx_nightly-2024.6.20.dist-info/top_level.txt,sha256=pxew3bc2gsiViS0zADs0jb6kC5v8o_Yy_85fhHj_J1A,7
+torchx_nightly-2024.6.20.dist-info/RECORD,,