cars 1.0.0rc3__cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

cars/orchestrator/cluster/dask_config/jobqueue.yaml

@@ -0,0 +1,26 @@
+jobqueue:
+  pbs:
+    name: cars-dask-worker      # Name of dask worker (set to PBS JOB NAME)
+
+    # Dask worker options
+    cores: null                 # Total number of cores per job
+    memory: null                # Total amount of memory per job
+    processes: null             # Number of Python processes per job
+
+    interface: null             # Network interface to use like eth0 or ib0
+    death-timeout: 3000         # Number of seconds to wait if a worker can not find a scheduler
+    local-directory: null       # Location of fast local storage like /scratch or $TMPDIR
+    extra: []                   # Additional arguments to pass to `dask-worker`
+
+    # PBS resource manager options
+    shebang: "#!/usr/bin/env bash"  # Path to desired interpreter for your batch submission script.
+    queue: null                     # PBS queue by default. Take the one specified in PBSCluster() first
+    project: null                   # PBS project name in ACCOUNT information
+    walltime: '00:59:00'            # PBS walltime by default.
+    env-extra: []                   # PBS environnements to pass to workers
+    resource-spec: null             # Resource string for PBS cpu and memory by default
+    job-extra: []                   # ?
+    log-directory: null             # Log directory by default
+
+    # Scheduler options
+    scheduler-options: {}           # Used to pass additional arguments to Dask Scheduler.

cars/orchestrator/cluster/dask_config/reference_confs/dask-schema.yaml

@@ -0,0 +1,137 @@
+properties:
+  temporary-directory:
+    type:
+    - string
+    - "null"
+    description: |
+      Temporary directory for local disk storage /tmp, /scratch,
+      or /local.  This directory is used during dask spill-to-disk operations.
+
+      When the value is "null" (default), dask will create a directory from
+      where dask was launched: `cwd/dask-worker-space`
+
+  tokenize:
+    type: object
+    properties:
+      ensure-deterministic:
+        type:
+        - boolean
+        description: |
+          If ``true``, tokenize will error instead of falling back to uuids
+          when a deterministic token cannot be generated. Defaults to
+          ``false``.
+
+  dataframe:
+    type: object
+    properties:
+
+      shuffle-compression:
+        type:
+        - string
+        - "null"
+        description: |
+          Compression algorithm used for on disk-shuffling. Partd, the library used
+          for compression supports ZLib, BZ2, SNAPPY, and BLOSC
+
+      parquet:
+        type: object
+        properties:
+
+          metadata-task-size-local:
+            type: integer
+            description: |
+              The number of files to handle within each metadata-processing
+              task when reading a parquet dataset from a LOCAL file system.
+              Specifying 0 will result in serial execution on the client.
+
+          metadata-task-size-remote:
+            type: integer
+            description: |
+              The number of files to handle within each metadata-processing
+              task when reading a parquet dataset from a REMOTE file system.
+              Specifying 0 will result in serial execution on the client.
+
+  array:
+    type: object
+    properties:
+
+      svg:
+        type: object
+        properties:
+
+          size:
+            type: integer
+            description: |
+              The size of pixels used when displaying a dask array as an SVG image.
+              This is used, for example, for nice rendering in a Jupyter notebook
+
+      slicing:
+        type: object
+        properties:
+          split-large-chunks:
+            type: [boolean, 'null']
+            description: |
+              How to handle large chunks created when slicing Arrays. By default a
+              warning is produced. Set to ``False`` to silence the warning
+              and allow large output chunks. Set to ``True`` to silence the
+              warning and avoid large output chunks.
+
+  optimization:
+    type: object
+    properties:
+
+      fuse:
+        type: object
+        description: Options for Dask's task fusion optimizations
+        properties:
+
+          active:
+            type: [boolean, 'null']
+            description: |
+              Turn task fusion on/off. This option refers to the fusion of a
+              fully-materialized task graph (not a high-Level graph). By default
+              (None), the active task-fusion option will be treated as ``False``
+              for Dask-Dataframe collections, and as ``True`` for all other graphs
+              (including Dask-Array collections).
+
+          ave-width:
+            type: number
+            minimum: 0
+            description:
+              Upper limit for width, where width = num_nodes / height, a good measure
+              of parallelizability
+
+          max-width:
+            type: [number, 'null']
+            minimum: 0
+            description:
+              Don't fuse if total width is greater than this. Set to null to dynamically
+              adjust to 1.5 + ave_width * log(ave_width + 1)
+
+          max-height:
+            type: number
+            minimum: 0
+            description: Don't fuse more than this many levels
+
+          max-depth-new-edges:
+            type: [number, 'null']
+            minimum: 0
+            description:
+              Don't fuse if new dependencies are added after this many levels.
+              Set to null to dynamically adjust to ave_width * 1.5.
+
+          subgraphs:
+            type: [boolean, 'null']
+            description: |
+              Set to True to fuse multiple tasks into SubgraphCallable objects. Set to
+              None to let the default optimizer of individual dask collections decide.
+              If no collection-specific default exists, None defaults to False.
+
+          rename-keys:
+            type: boolean
+            description:
+              Set to true to rename the fused keys with `default_fused_keys_renamer`.
+              Renaming fused keys can keep the graph more understandable and
+              comprehensible, but it comes at the cost of additional processing. If
+              False, then the top-most key will be used. For advanced usage, a function
+              to create the new name is also accepted.

cars/orchestrator/cluster/dask_config/reference_confs/dask.yaml

@@ -0,0 +1,26 @@
+temporary-directory: null  # Directory for local disk like /tmp, /scratch, or /local
+
+tokenize:
+  ensure-deterministic: false  # If true, tokenize will error instead of falling back to uuids
+
+dataframe:
+  shuffle-compression: null  # compression for on disk-shuffling. Partd supports ZLib, BZ2, SNAPPY, BLOSC
+  parquet:
+    metadata-task-size-local: 512  # Number of files per local metadata-processing task
+    metadata-task-size-remote: 16  # Number of files per remote metadata-processing task
+
+array:
+  svg:
+    size: 120  # pixels
+  slicing:
+    split-large-chunks: null  # How to handle large output chunks in slicing. Warns by default.
+
+optimization:
+  fuse:
+    active: null  # Treat as false for dask.dataframe, true for everything else
+    ave-width: 1
+    max-width: null  # 1.5 + ave_width * log(ave_width + 1)
+    max-height: .inf
+    max-depth-new-edges: null  # ave_width * 1.5
+    subgraphs: null  # true for dask.dataframe, false for everything else
+    rename-keys: true