doc: firts steps towards a structured API documentation

.gitignore

@@ -2,6 +2,7 @@
 /slurm-*.out
 /wandb
 /doc/_build
+/doc/generated/
 
 ############################  Auto Generated ############################
 

dmlcloud/__init__.py

@@ -1,11 +1,40 @@
 """
+Hello world
 """
 
+from dmlcloud.pipeline import TrainingPipeline
 from dmlcloud.stage import Stage, TrainValStage
+from dmlcloud.util.distributed import *
 
 __version__ = "0.3.3"
 
 __all__ = [
+    'TrainingPipeline',
     'Stage',
     'TrainValStage',
 ]
+
+# Ditributed helpers
+__all__ += [
+    'has_slurm',
+    'has_environment',
+    'has_mpi',
+    'is_root',
+    'root_only',
+    'root_first',
+    'rank',
+    'world_size',
+    'local_rank',
+    'local_world_size',
+    'local_node',
+    'print_worker',
+    'print_root',
+    'all_gather_object',
+    'gather_object',
+    'broadcast_object',
+    'init_process_group_dummy',
+    'init_process_group_slurm',
+    'init_process_group_MPI',
+    'init_process_group_auto',
+    'deinitialize_torch_distributed',
+]

dmlcloud/util/distributed.py

@@ -6,6 +6,30 @@
 
 from .tcp import find_free_port, get_local_ips
 
+__all__ = [
+    'has_slurm',
+    'has_environment',
+    'has_mpi',
+    'is_root',
+    'root_only',
+    'root_first',
+    'rank',
+    'world_size',
+    'local_rank',
+    'local_world_size',
+    'local_node',
+    'print_worker',
+    'print_root',
+    'all_gather_object',
+    'gather_object',
+    'broadcast_object',
+    'init_process_group_dummy',
+    'init_process_group_slurm',
+    'init_process_group_MPI',
+    'init_process_group_auto',
+    'deinitialize_torch_distributed',   
+]
+
 
 DEFAULT_PORT = os.environ.get('DMLCLOUD_PORT', 41312)  # dml
 

doc/conf.py

@@ -12,22 +12,27 @@
 release = 'v0.3.3'
 
 # -- General configuration ---------------------------------------------------
-# https://www.sphinx-doc.org/en/master/usage/configuration.html#general-configuration
-
 extensions = [
     'sphinx.ext.duration',
     'sphinx.ext.autodoc',
     'sphinx.ext.autosummary',
     'sphinx.ext.napoleon',
+    'sphinx.ext.coverage',
 ]
 
 templates_path = ['_templates']
 exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
 
 
 # -- Options for HTML output -------------------------------------------------
-# https://www.sphinx-doc.org/en/master/usage/configuration.html#options-for-html-output
-
 html_theme = 'sphinx_rtd_theme'
 html_static_path = ['_static']
 html_logo = '../misc/logo/dmlcloud_light.png'
+
+
+# Napoleon configs
+napoleon_google_docstring = True
+napoleon_numpy_docstring = False
+napoleon_use_param = False
+napoleon_use_rtype = False
+napoleon_preprocess_types = True

doc/reference.rst

@@ -1,9 +1,46 @@
+.. currentmodule:: dmlcloud
+
 API Reference
-=========
+=============
+
+This the API reference for the dmlcloud package.
+
+.. autosummary::
+    :toctree: generated
+
+    TrainingPipeline
+    Stage
+
+
+Initialization
+--------------
+.. autosummary::
+    :toctree: generated
+
+    init_process_group_dummy
+    init_process_group_slurm
+    init_process_group_MPI
+    init_process_group_auto
 
-This is an automatically generated API reference for the dmlcloud package.
 
+Distributed Helpers
+-------------------
 .. autosummary::
    :toctree: generated
 
-   dmlcloud
+    has_slurm
+    has_environment
+    has_mpi
+    is_root
+    root_only
+    root_first
+    rank
+    world_size
+    local_rank
+    local_world_size
+    local_node
+    print_worker
+    print_root
+    all_gather_object
+    gather_object
+    broadcast_object