* feat: update docs to accomodate for DAD

* undo change

* Migrate PathwaysJob CRD to JobSet

This commit switches the XPK workload generation template for pathways workloads from using the PathwaysJob CRD directly over to a standard JobSet object.

- Updates PW_WORKLOAD_CREATE_YAML to use JobSet API

- Modifies component YAML generation in pathways.py to output Pod containers rather than custom PathwaysJob components

- Adds unit tests to workload_test.py to verify correct layout of pathways jobsets

* Fix args.worker_image attribute error in Pathways worker container string formatting

* Ensure Pathways JobSet parity with legacy PathwaysJob controller output

* Convert proxy and RM sidecars to initContainers with restartPolicy: Always.
* Ensure all container ports specify protocol: TCP.
* Enforce restartPolicy: OnFailure for the worker template.
* Inject required environment variables (JAX_PLATFORMS, JAX_BACKEND_TARGET, XCLOUD_ENVIRONMENT) natively into the primary user workload container.
* Add completionMode: Indexed to both head and worker replicated jobs.
* Set successPolicy, startupPolicy, and suspend fields to match legacy generated JobSet.

* Update pathways instance_type formatting in JobSet migration

* Added get_pathways_instance_type to core/pathways.py
* Format the gke_accelerator natively for the RM sidecar (e.g. tpuv6e:2x2)

* Update backoffLimit for Pathways worker to match legacy controller scaling

* Set worker backoffLimit to args.max_slice_restarts * 4 to replicate the legacy controller's logic.

* Remove PathwaysJob CRD installation from cluster creation

* Since XPK now deploys Pathways workloads using native JobSet API, we no longer need to install the PathwaysJob CRD when creating or adapting a cluster.

* Refactor append_custom_pathways_flags to remove magic number offsetting

* Changed prev_indentation to base_indentation to specify exactly how many spaces to prepend.
* Removed the convoluted len(indentation) - 2 calls in favor of passing base_indentation=16 directly.

* Refactor env variable injection to use regex

* Instead of manually checking for multiple variants of trailing whitespace after the 'env:' key, use the re module to robustly identify and inject the required Pathways environment variables.

* Remove unnecessary else block from env injection

* Since docker_container.py generates the container YAML template using `env: {env}\n`, the key will always be followed by an optional space and a newline, whether {env} is empty or not. Thus, the defensive else block was unreachable and unnecessary.

* Move regex import to top of pathways.py

* Moved the inline 'import re' from get_user_workload_for_pathways to the module level.

* Simplify environment variable injection logic

* Removed regex pattern matching in favor of a direct string replacement on 'env:' since 'env: {env}' always appears in the template even if {env} evaluates to an empty string. This ensures our injected block always gets prepended correctly before any user variables.

* Remove redundant 'env:' from env_injection string

* Since we always append the variables directly after the 'env:' key from the main container template, we can just define the injected block without the 'env:' header entirely and avoid having to strip it out.

* Add return type hint to get_user_workload_container

* Solves the 'Returning Any from function declared to return str' type checker error in pathways.py by explicitly defining the tuple return type of get_user_workload_container.

* Refactor test_workload_create_pathways_jobset_yaml assertions

* Replaced hardcoded strings and numbers (like 'test-pw-workload', 'test-docker', and 2) with formatted variables linked directly to the mocked args (e.g. {args.workload}, {args.num_slices}) to make the unit test more robust and maintainable.

* Expand test_workload_create_pathways_jobset_yaml assertions

* Added explicit assertions for the newly migrated JobSet fields including the coordinator block, network dns configurations, restart strategies, completion modes, and strict verification of the dynamically scaled backoffLimit calculation.

* Apply auto-formatting to pathways and workload files

* Fix inline worker_backoff_limit calculation inside format string

* Bypass CRD check during Pathways workload creation

* Pass system characteristics to dynamically calculate google.com/tpu limits

* Migrate PathwaysJob CRD to JobSet

This commit switches the XPK workload generation template for pathways workloads from using the PathwaysJob CRD directly over to a standard JobSet object.

- Updates PW_WORKLOAD_CREATE_YAML to use JobSet API

- Modifies component YAML generation in pathways.py to output Pod containers rather than custom PathwaysJob components

- Adds unit tests to workload_test.py to verify correct layout of pathways jobsets

* Fix args.worker_image attribute error in Pathways worker container string formatting

* Ensure Pathways JobSet parity with legacy PathwaysJob controller output

* Convert proxy and RM sidecars to initContainers with restartPolicy: Always.
* Ensure all container ports specify protocol: TCP.
* Enforce restartPolicy: OnFailure for the worker template.
* Inject required environment variables (JAX_PLATFORMS, JAX_BACKEND_TARGET, XCLOUD_ENVIRONMENT) natively into the primary user workload container.
* Add completionMode: Indexed to both head and worker replicated jobs.
* Set successPolicy, startupPolicy, and suspend fields to match legacy generated JobSet.

* Update pathways instance_type formatting in JobSet migration

* Added get_pathways_instance_type to core/pathways.py
* Format the gke_accelerator natively for the RM sidecar (e.g. tpuv6e:2x2)

* Update backoffLimit for Pathways worker to match legacy controller scaling

* Set worker backoffLimit to args.max_slice_restarts * 4 to replicate the legacy controller's logic.

* Remove PathwaysJob CRD installation from cluster creation

* Since XPK now deploys Pathways workloads using native JobSet API, we no longer need to install the PathwaysJob CRD when creating or adapting a cluster.

* Refactor append_custom_pathways_flags to remove magic number offsetting

* Changed prev_indentation to base_indentation to specify exactly how many spaces to prepend.
* Removed the convoluted len(indentation) - 2 calls in favor of passing base_indentation=16 directly.

* Refactor env variable injection to use regex

* Instead of manually checking for multiple variants of trailing whitespace after the 'env:' key, use the re module to robustly identify and inject the required Pathways environment variables.

* Remove unnecessary else block from env injection

* Since docker_container.py generates the container YAML template using `env: {env}\n`, the key will always be followed by an optional space and a newline, whether {env} is empty or not. Thus, the defensive else block was unreachable and unnecessary.

* Move regex import to top of pathways.py

* Moved the inline 'import re' from get_user_workload_for_pathways to the module level.

* Simplify environment variable injection logic

* Removed regex pattern matching in favor of a direct string replacement on 'env:' since 'env: {env}' always appears in the template even if {env} evaluates to an empty string. This ensures our injected block always gets prepended correctly before any user variables.

* Remove redundant 'env:' from env_injection string

* Since we always append the variables directly after the 'env:' key from the main container template, we can just define the injected block without the 'env:' header entirely and avoid having to strip it out.

* Add return type hint to get_user_workload_container

* Solves the 'Returning Any from function declared to return str' type checker error in pathways.py by explicitly defining the tuple return type of get_user_workload_container.

* Refactor test_workload_create_pathways_jobset_yaml assertions

* Replaced hardcoded strings and numbers (like 'test-pw-workload', 'test-docker', and 2) with formatted variables linked directly to the mocked args (e.g. {args.workload}, {args.num_slices}) to make the unit test more robust and maintainable.

* Expand test_workload_create_pathways_jobset_yaml assertions

* Added explicit assertions for the newly migrated JobSet fields including the coordinator block, network dns configurations, restart strategies, completion modes, and strict verification of the dynamically scaled backoffLimit calculation.

* Apply auto-formatting to pathways and workload files

* Fix inline worker_backoff_limit calculation inside format string

* Bypass CRD check during Pathways workload creation

* Pass system characteristics to dynamically calculate google.com/tpu limits

* Linter fix

* Fix unit tests

* fix linter

* Address PR feedback for Pathways JobSet migration

* Fixed get_pathways_instance_type to explicitly map GCE machine types to TPU versions.
* Fixed headless mode to properly inject proxy and RM into 'containers' instead of 'initContainers'.
* Fixed worker backoffLimit calculation to accurately scale by vms_per_slice.
* Respected --elastic-slices CLI argument by passing it to pathways-proxy.
* Added deprecation warning for the legacy colocate_head_with_workers deployment mode.
* Cleaned up dead targetReplicatedJob tracking and conditionally disabled successPolicy for headless workloads.
* Updated unit tests to mock and assert the new conditional formats and fields correctly.

* update goldens

* fix linter

* Handle unknown pathways machine types and fix tests

* Refactor Pathways TPU version handling to use SystemCharacteristics directly

* Added an optional  field directly to .
* Populated the  for all supported TPU generations (v4, v5e, v5p, v6e, tpu7, tpu7x) directly in their system definitions in .
* Removed the hardcoded  fallback mapping and its associated exception from  entirely, fetching the version dynamically from the system configuration instead.

* Enforce pathways_tpu_version as a mandatory keyword argument for TPU configurations

* Fix unit tests after changing get_tpu_system_characteristics_map signature

* some formatting changes

* golden updates

* cleanup

* Readme update

* Fix YAML indentation for pathways_head_containers injection

* Goldens update

* Refactoring workload list parsing to Python

* Fix goldens to use [0] instead of [*] in refactoring branch

* Remove _sum_counts from refactoring branch

* Update goldens

* Update goldens with PYTHONPATH=src

* Remove test-queued-job and move super_slicing test to wildcard

* Address refactoring feedback: fix row delimiter, restore headers, use case-insensitive matching

* Address refactoring feedback: add ROW_DELIMITER const, render <empty> row when empty

* Rename delimiter and fix cluster delete workloads parsing

* Migrate to JSON parsing and use _WorkloadListRow dataclass

* Fix Priority JSON path

* Add fallback to spec.priorityClassName

This reverts commit 989ba24.

Fix wildcard counting logic and priority for super-slicing

* Migrate PathwaysJob CRD to JobSet

This commit switches the XPK workload generation template for pathways workloads from using the PathwaysJob CRD directly over to a standard JobSet object.

- Updates PW_WORKLOAD_CREATE_YAML to use JobSet API

- Modifies component YAML generation in pathways.py to output Pod containers rather than custom PathwaysJob components

- Adds unit tests to workload_test.py to verify correct layout of pathways jobsets

* Fix args.worker_image attribute error in Pathways worker container string formatting

* Ensure Pathways JobSet parity with legacy PathwaysJob controller output

* Convert proxy and RM sidecars to initContainers with restartPolicy: Always.
* Ensure all container ports specify protocol: TCP.
* Enforce restartPolicy: OnFailure for the worker template.
* Inject required environment variables (JAX_PLATFORMS, JAX_BACKEND_TARGET, XCLOUD_ENVIRONMENT) natively into the primary user workload container.
* Add completionMode: Indexed to both head and worker replicated jobs.
* Set successPolicy, startupPolicy, and suspend fields to match legacy generated JobSet.

* Update pathways instance_type formatting in JobSet migration

* Added get_pathways_instance_type to core/pathways.py
* Format the gke_accelerator natively for the RM sidecar (e.g. tpuv6e:2x2)

* Update backoffLimit for Pathways worker to match legacy controller scaling

* Set worker backoffLimit to args.max_slice_restarts * 4 to replicate the legacy controller's logic.

* Remove PathwaysJob CRD installation from cluster creation

* Since XPK now deploys Pathways workloads using native JobSet API, we no longer need to install the PathwaysJob CRD when creating or adapting a cluster.

* Refactor append_custom_pathways_flags to remove magic number offsetting

* Changed prev_indentation to base_indentation to specify exactly how many spaces to prepend.
* Removed the convoluted len(indentation) - 2 calls in favor of passing base_indentation=16 directly.

* Refactor env variable injection to use regex

* Instead of manually checking for multiple variants of trailing whitespace after the 'env:' key, use the re module to robustly identify and inject the required Pathways environment variables.

* Remove unnecessary else block from env injection

* Since docker_container.py generates the container YAML template using `env: {env}\n`, the key will always be followed by an optional space and a newline, whether {env} is empty or not. Thus, the defensive else block was unreachable and unnecessary.

* Move regex import to top of pathways.py

* Moved the inline 'import re' from get_user_workload_for_pathways to the module level.

* Simplify environment variable injection logic

* Removed regex pattern matching in favor of a direct string replacement on 'env:' since 'env: {env}' always appears in the template even if {env} evaluates to an empty string. This ensures our injected block always gets prepended correctly before any user variables.

* Remove redundant 'env:' from env_injection string

* Since we always append the variables directly after the 'env:' key from the main container template, we can just define the injected block without the 'env:' header entirely and avoid having to strip it out.

* Add return type hint to get_user_workload_container

* Solves the 'Returning Any from function declared to return str' type checker error in pathways.py by explicitly defining the tuple return type of get_user_workload_container.

* Refactor test_workload_create_pathways_jobset_yaml assertions

* Replaced hardcoded strings and numbers (like 'test-pw-workload', 'test-docker', and 2) with formatted variables linked directly to the mocked args (e.g. {args.workload}, {args.num_slices}) to make the unit test more robust and maintainable.

* Expand test_workload_create_pathways_jobset_yaml assertions

* Added explicit assertions for the newly migrated JobSet fields including the coordinator block, network dns configurations, restart strategies, completion modes, and strict verification of the dynamically scaled backoffLimit calculation.

* Apply auto-formatting to pathways and workload files

* Fix inline worker_backoff_limit calculation inside format string

* Bypass CRD check during Pathways workload creation

* Pass system characteristics to dynamically calculate google.com/tpu limits

* Migrate PathwaysJob CRD to JobSet

This commit switches the XPK workload generation template for pathways workloads from using the PathwaysJob CRD directly over to a standard JobSet object.

- Updates PW_WORKLOAD_CREATE_YAML to use JobSet API

- Modifies component YAML generation in pathways.py to output Pod containers rather than custom PathwaysJob components

- Adds unit tests to workload_test.py to verify correct layout of pathways jobsets

* Fix args.worker_image attribute error in Pathways worker container string formatting

* Ensure Pathways JobSet parity with legacy PathwaysJob controller output

* Convert proxy and RM sidecars to initContainers with restartPolicy: Always.
* Ensure all container ports specify protocol: TCP.
* Enforce restartPolicy: OnFailure for the worker template.
* Inject required environment variables (JAX_PLATFORMS, JAX_BACKEND_TARGET, XCLOUD_ENVIRONMENT) natively into the primary user workload container.
* Add completionMode: Indexed to both head and worker replicated jobs.
* Set successPolicy, startupPolicy, and suspend fields to match legacy generated JobSet.

* Update pathways instance_type formatting in JobSet migration

* Added get_pathways_instance_type to core/pathways.py
* Format the gke_accelerator natively for the RM sidecar (e.g. tpuv6e:2x2)

* Update backoffLimit for Pathways worker to match legacy controller scaling

* Set worker backoffLimit to args.max_slice_restarts * 4 to replicate the legacy controller's logic.

* Remove PathwaysJob CRD installation from cluster creation

* Since XPK now deploys Pathways workloads using native JobSet API, we no longer need to install the PathwaysJob CRD when creating or adapting a cluster.

* Refactor append_custom_pathways_flags to remove magic number offsetting

* Changed prev_indentation to base_indentation to specify exactly how many spaces to prepend.
* Removed the convoluted len(indentation) - 2 calls in favor of passing base_indentation=16 directly.

* Refactor env variable injection to use regex

* Instead of manually checking for multiple variants of trailing whitespace after the 'env:' key, use the re module to robustly identify and inject the required Pathways environment variables.

* Remove unnecessary else block from env injection

* Since docker_container.py generates the container YAML template using `env: {env}\n`, the key will always be followed by an optional space and a newline, whether {env} is empty or not. Thus, the defensive else block was unreachable and unnecessary.

* Move regex import to top of pathways.py

* Moved the inline 'import re' from get_user_workload_for_pathways to the module level.

* Simplify environment variable injection logic

* Removed regex pattern matching in favor of a direct string replacement on 'env:' since 'env: {env}' always appears in the template even if {env} evaluates to an empty string. This ensures our injected block always gets prepended correctly before any user variables.

* Remove redundant 'env:' from env_injection string

* Since we always append the variables directly after the 'env:' key from the main container template, we can just define the injected block without the 'env:' header entirely and avoid having to strip it out.

* Add return type hint to get_user_workload_container

* Solves the 'Returning Any from function declared to return str' type checker error in pathways.py by explicitly defining the tuple return type of get_user_workload_container.

* Refactor test_workload_create_pathways_jobset_yaml assertions

* Replaced hardcoded strings and numbers (like 'test-pw-workload', 'test-docker', and 2) with formatted variables linked directly to the mocked args (e.g. {args.workload}, {args.num_slices}) to make the unit test more robust and maintainable.

* Expand test_workload_create_pathways_jobset_yaml assertions

* Added explicit assertions for the newly migrated JobSet fields including the coordinator block, network dns configurations, restart strategies, completion modes, and strict verification of the dynamically scaled backoffLimit calculation.

* Apply auto-formatting to pathways and workload files

* Fix inline worker_backoff_limit calculation inside format string

* Bypass CRD check during Pathways workload creation

* Pass system characteristics to dynamically calculate google.com/tpu limits

* Linter fix

* Fix unit tests

* fix linter

* Address PR feedback for Pathways JobSet migration

* Fixed get_pathways_instance_type to explicitly map GCE machine types to TPU versions.
* Fixed headless mode to properly inject proxy and RM into 'containers' instead of 'initContainers'.
* Fixed worker backoffLimit calculation to accurately scale by vms_per_slice.
* Respected --elastic-slices CLI argument by passing it to pathways-proxy.
* Added deprecation warning for the legacy colocate_head_with_workers deployment mode.
* Cleaned up dead targetReplicatedJob tracking and conditionally disabled successPolicy for headless workloads.
* Updated unit tests to mock and assert the new conditional formats and fields correctly.

* update goldens

* fix linter

* Handle unknown pathways machine types and fix tests

* Refactor Pathways TPU version handling to use SystemCharacteristics directly

* Added an optional  field directly to .
* Populated the  for all supported TPU generations (v4, v5e, v5p, v6e, tpu7, tpu7x) directly in their system definitions in .
* Removed the hardcoded  fallback mapping and its associated exception from  entirely, fetching the version dynamically from the system configuration instead.

* Enforce pathways_tpu_version as a mandatory keyword argument for TPU configurations

* Fix unit tests after changing get_tpu_system_characteristics_map signature

* some formatting changes

* golden updates

* cleanup

* Readme update

* Fix YAML indentation for pathways_head_containers injection

* Goldens update

* fix the identation problem

…1114)

fix aggregate reservation accelerator type matching for DWS Calendar reservations;

…1111)

fix: relax Kueue version check for custom image SHAs

* Refactor pathways workload scheduling to use Jinja template

* Refactor jinja injection and cleanup workload.py

…e created (#1125)

* fix: skip reservation capacity assessment when no nodepools need to be created

* test: add test for skipping reservation check

* test: improve unit test by mocking an empty reservation instead of patching get_reservations_list

* test: remove comments from test function

* Enable Crane. Keep a golden and integration test for Docker (to be cleaned up later)

* Remove non-crane tests