Update copyright: 2023 -> 2022-2024. (#560)

* Update copyright: 2023 -> 2022-2024. * Update .ci/FILE_HEADER.
mosaicml · Jan 12, 2024 · 1bae380 · 1bae380
1 parent aead15a
commit 1bae380
Show file tree

Hide file tree

Showing 172 changed files with 173 additions and 173 deletions.
diff --git a/.ci/FILE_HEADER b/.ci/FILE_HEADER
@@ -1,2 +1,2 @@
-Copyright 2023 MosaicML Streaming authors
+Copyright 2022-2024 MosaicML Streaming authors
 SPDX-License-Identifier: Apache-2.0
diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Configuration file for the Sphinx documentation builder.
@@ -63,7 +63,7 @@
 # -- Project information -----------------------------------------------------
 
 project = 'Streaming'
-copyright = '2023, MosaicML, Inc.'
+copyright = '2022, MosaicML, Inc.'
 author = 'MosaicML'
 
 # -- General configuration ---------------------------------------------------

diff --git a/docs/source/doctest_cleanup.py b/docs/source/doctest_cleanup.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Cleanup script that is executed at the end of each doctest."""

diff --git a/docs/source/doctest_fixtures.py b/docs/source/doctest_fixtures.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 # disabling general type issues because of monkeypatching

diff --git a/regression/iterate_data.py b/regression/iterate_data.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Create a streaming dataset from toy data with various options for regression testing."""

diff --git a/regression/synthetic_dataset.py b/regression/synthetic_dataset.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Create a dataset and save it to a directory."""

diff --git a/regression/utils.py b/regression/utils.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Utility and helper functions for regression testing."""

diff --git a/scripts/compression/bench.py b/scripts/compression/bench.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Script to Benchmark compression algorithms."""

diff --git a/scripts/compression/plot.py b/scripts/compression/plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Utility and helper functions to plot compression information."""

diff --git a/scripts/epoch/bench.py b/scripts/epoch/bench.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Benchmark time taken to generate the epoch for a given dataset."""

diff --git a/scripts/hashing/bench.py b/scripts/hashing/bench.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Script to benchmark various hashing algorithms."""

diff --git a/scripts/hashing/plot.py b/scripts/hashing/plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Utility and helper functions to plot hashing information."""

diff --git a/scripts/partition/bench.py b/scripts/partition/bench.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Command-line tool to compare StreamingDataset sample space partitioning performance."""

diff --git a/scripts/partition/diff.py b/scripts/partition/diff.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Collect data on the dataset padding needed by partitioning."""

diff --git a/scripts/partition/plot.py b/scripts/partition/plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Plot results of comparing streaming dataset partitioning algorithms."""

diff --git a/scripts/partition/txt.py b/scripts/partition/txt.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Command-line tool to visualize StreamingDataset sample space partitioning."""

diff --git a/scripts/partition/web.py b/scripts/partition/web.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Web app to visualize StreamingDataset sample space partitioning.

diff --git a/scripts/samples/bench_and_plot.py b/scripts/samples/bench_and_plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Benchmark and plot sample access times across kinds of data and formats."""

diff --git a/scripts/serialization/compare.py b/scripts/serialization/compare.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Compare dataset serialization methods.

diff --git a/scripts/serialization/survey_fixed_decimals.py b/scripts/serialization/survey_fixed_decimals.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Survey the space of practical fixed-size decimal encodings."""

diff --git a/scripts/shuffle/bench.py b/scripts/shuffle/bench.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Command-line tool to benchmark StreamingDataset shuffling performance."""

diff --git a/scripts/shuffle/plot.py b/scripts/shuffle/plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Plot results of comparing streaming dataset shuffling algorithms."""

diff --git a/scripts/shuffle/vis.py b/scripts/shuffle/vis.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Visualize different shuffling algorithms on fake shards."""

diff --git a/scripts/webvid/bench_inside.py b/scripts/webvid/bench_inside.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Benchmark iterating an MP4-inside MDS dataset."""

diff --git a/scripts/webvid/bench_outside_dt.py b/scripts/webvid/bench_outside_dt.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Benchmark iterating an MP4-outside MDS dataset."""

diff --git a/scripts/webvid/bench_outside_gi.py b/scripts/webvid/bench_outside_gi.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Benchmark iterating an MP4-outside MDS dataset."""

diff --git a/scripts/webvid/plot.py b/scripts/webvid/plot.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Plot download performance comparison."""

diff --git a/setup.py b/setup.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Streaming package setup."""

diff --git a/simulation/__init__.py b/simulation/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Streaming simulation for throughput, network downloads, and shuffle quality."""
diff --git a/simulation/core/__init__.py b/simulation/core/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Core classes and functions for streaming simulation."""
diff --git a/simulation/core/create_index.py b/simulation/core/create_index.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Create a dataset index file from input parameters."""

diff --git a/simulation/core/last_used_ordered_set.py b/simulation/core/last_used_ordered_set.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """An ordered set that can be used as an LRU cache."""

diff --git a/simulation/core/main.py b/simulation/core/main.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Main simulation function, simulating bytes downloaded and time taken each training step."""

diff --git a/simulation/core/node_tracker.py b/simulation/core/node_tracker.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Class for tracking node information during simulation."""

diff --git a/simulation/core/shard_downloads.py b/simulation/core/shard_downloads.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Functions for simulating shard downloads and calculating needed cache limit for downloads."""

diff --git a/simulation/core/shuffle_quality.py b/simulation/core/shuffle_quality.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Determine shuffle quality of a run over a fixed number of samples."""

diff --git a/simulation/core/sim_dataset.py b/simulation/core/sim_dataset.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Near replica of StreamingDataset for simulation purposes."""

diff --git a/simulation/core/sim_spanner.py b/simulation/core/sim_spanner.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Mapping of global sample index to shard index for simulation purposes."""

diff --git a/simulation/core/sim_time.py b/simulation/core/sim_time.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Time classes ported from MosaicML composer.

diff --git a/simulation/core/sim_world.py b/simulation/core/sim_world.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Contains info about the nodes, ranks, and workers of the run for simulation purposes."""

diff --git a/simulation/core/utils.py b/simulation/core/utils.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Peripheral functions for simulation functionality."""

diff --git a/simulation/core/yaml_processing.py b/simulation/core/yaml_processing.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Ingest yaml and create SimulationDataset."""

diff --git a/simulation/interfaces/__init__.py b/simulation/interfaces/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Interfaces for streaming simulation."""
diff --git a/simulation/interfaces/interface_utils.py b/simulation/interfaces/interface_utils.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Peripheral functions for interface functionality."""

diff --git a/simulation/interfaces/sim_cli.py b/simulation/interfaces/sim_cli.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """sim_cli: simulate your training yaml from the command line."""

diff --git a/simulation/interfaces/sim_script.py b/simulation/interfaces/sim_script.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Script for simulating training downloads and throughput, and displaying results."""

diff --git a/simulation/interfaces/sim_ui.py b/simulation/interfaces/sim_ui.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Simulator web UI using streamlit."""

diff --git a/simulation/interfaces/widgets.py b/simulation/interfaces/widgets.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Streamlit widgets for simulation web UI."""

diff --git a/simulation/launcher.py b/simulation/launcher.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Launch simulator UI from command line when this package is installed."""

diff --git a/simulation/testing/wandb_testing.py b/simulation/testing/wandb_testing.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Test simulation results against run results from a wandb project."""

diff --git a/streaming/__init__.py b/streaming/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """MosaicML Streaming Datasets for cloud-native model training."""

diff --git a/streaming/_version.py b/streaming/_version.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """The Streaming Version."""

diff --git a/streaming/base/__init__.py b/streaming/base/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """MosaicML Streaming Datasets for cloud-native model training."""

diff --git a/streaming/base/array.py b/streaming/base/array.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """A read-only list of items that can be fancy indexed like a numpy array."""

diff --git a/streaming/base/batching/__init__.py b/streaming/base/batching/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Apportion shards/samples to nodes/ranks/workers for elastically deterministic sample order."""

diff --git a/streaming/base/batching/per_stream.py b/streaming/base/batching/per_stream.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Apportion shards/samples such that batches have samples only from a single stream."""

diff --git a/streaming/base/batching/random.py b/streaming/base/batching/random.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Apportion shards/samples such that batches have samples randomly selected from streams."""

diff --git a/streaming/base/batching/stratified.py b/streaming/base/batching/stratified.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Apportion shards/samples so each batch has the same amount of samples from each stream."""

diff --git a/streaming/base/compression.py b/streaming/base/compression.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """List of Compression and Decompression algorithms."""

diff --git a/streaming/base/constant.py b/streaming/base/constant.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Constants."""

diff --git a/streaming/base/converters/__init__.py b/streaming/base/converters/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Utility function for converting spark dataframe to MDS dataset."""

diff --git a/streaming/base/converters/dataframe_to_mds.py b/streaming/base/converters/dataframe_to_mds.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """A utility to convert spark dataframe to MDS."""

diff --git a/streaming/base/dataloader.py b/streaming/base/dataloader.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Streaming DataLoader."""

diff --git a/streaming/base/dataset.py b/streaming/base/dataset.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """A mid-epoch-resumable streaming/caching pytorch IterableDataset."""

diff --git a/streaming/base/distributed.py b/streaming/base/distributed.py
@@ -1,4 +1,4 @@
-# Copyright 2023 MosaicML Streaming authors
+# Copyright 2022-2024 MosaicML Streaming authors
 # SPDX-License-Identifier: Apache-2.0
 
 """Helper methods to get the distributed attributes."""