Merge 0.2.0 into main (PaddlePaddle#143)

* data adapter supports paddle 2.3.1 * update module from models.dl.paddlepaddle.* to models.forecasting.dl.* * Add some new functions and fix some bugs * Add some new functions and fix some bugs * Modify the detail of callbacks * Add some new functions and fix some bugs (PaddlePaddle#114) * thirdparty model integration. * Backtest feature upgrade * Modify the details of models (lstnet, mlp, tcn, transformer) * thirdparty transform integration * Add informer model * backtest unitest upgrade * remove paddlepaddle version check in paddle data adapter. * Add TS2Vec representaion model * fix __init__.py * Dataset splitter and Cross validation (PaddlePaddle#121) * Dataset splitter and Cross validation * Dataset splitter and Cross validation * Update logger.py (PaddlePaddle#122) * probability forecasting & deepar * probability forecasting & deepar * Pipeline recursive predict optimize (PaddlePaddle#125) * change version to 0.2.0 * representation dl data adapter. * Add unitest of informer and ts2vec * representation dl data adapter. (PaddlePaddle#128) * Add some necessary files * Add some necessary files * add AutoTS(autots、optimize_runner、search_space_configer、searcher） (PaddlePaddle#130) * Update logger.py * add AutoTS(autots、optimize_runner、search_space_configer、searcher） * setup autots * Update test_search_space_configer.py * fix bug in nhits (PaddlePaddle#131) * Update logger.py * add AutoTS(autots、optimize_runner、search_space_configer、searcher） * setup autots * Update test_search_space_configer.py * fix bug in nhits * paddlets 0.2.0 docs. * paddlets.__version__ == 0.2.0 * add autots docs. * update docker image in run_on_gpu.rst * sklearn<=1.0.2 * update index, add autots api docs. * add proba forecast docs. * add proba forecast docs. * update FFT.png in get_started.rst * add req/autots.txt into .readthedocs.yaml * supplement tsdataset api zh docs * supplement tsdataset api zh docs (PaddlePaddle#135) * add representation docs * add representation docs (PaddlePaddle#136) * supplement tsdataset api zh docs * add representation docs * modify docs utils.utils (PaddlePaddle#138) * update ml model warpper unittest. (PaddlePaddle#137) * fix doc bug (PaddlePaddle#139) * Update logger.py * add AutoTS(autots、optimize_runner、search_space_configer、searcher） * setup autots * Update test_search_space_configer.py * fix bug in nhits * fix doc bug * modify requirements * fix zh docs bug in repr * deadlink in model doc (PaddlePaddle#142) Co-authored-by: a10210532 <[email protected]> Co-authored-by: Annnnnnnnnnnnn <[email protected]> Co-authored-by: QGN123 <[email protected]> Co-authored-by: QGN123 <[email protected]> Co-authored-by: LinWencong <[email protected]> Co-authored-by: bianchuanxin <[email protected]>
Lfan-ke · Sep 19, 2022 · 8062e5c · 8062e5c
1 parent 514e6d6
commit 8062e5c
Show file tree

Hide file tree

Showing 283 changed files with 25,659 additions and 5,853 deletions.
diff --git a/.readthedocs.yaml b/.readthedocs.yaml
@@ -24,5 +24,6 @@ python:
     - requirements: requirements/core.txt
     - requirements: requirements/docs.txt
     - requirements: requirements/paddle.txt
+    - requirements: requirements/autots.txt
 
   system_packages: true
diff --git a/README_cn.md b/README_cn.md
@@ -23,27 +23,33 @@ PaddleTS 是一个易用的深度时序建模的Python库，它基于飞桨深
 * 内置业界领先的深度学习模型，如NBEATS、NHiTS、LSTNet、TCN、Transformer等
 * 内置多样化的数据转换算子，支持数据处理与转换，包括缺失值填充、异常值处理、归一化、时间相关的协变量提取等
 * 内置经典的数据分析算子，帮助开发者便捷实现数据探索，包括数据统计量信息及数据摘要等功能
+* 自动超参寻优
+* 第三方机器学习模型及数据转换模块自动集成
+* 支持在GPU设备上运行基于PaddlePaddle的时序模型
+* 新增DeepAR模型，支持时序概率预测
+* 新增TS2Vec模型，支持时序表征学习
 
 未来，更多的高级特性会进一步发布，包括但不限于：
-* 自动超参寻优
-* 时序表征模型
-* 概率预测模型
+* 支持时序异常检测
+* 更多时序表征模型
+* 更多概率预测模型
 * 场景化Pipeline，支持端到端真实场景解决方案
 
 
 ## 关于 PaddleTS
 
 具体来说，PaddleTS 时序库包含以下子模块：
 
-| 模块                                                                                                            | 简述                                              |
-|---------------------------------------------------------------------------------------------------------------|-------------------------------------------------|
-| [**paddlets.datasets**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/datasets/overview.html)   | 时序数据模块，统一的时序数据结构和预定义的数据处理方法           |
-| [**paddlets.transform**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/transform/overview.html) | 数据转换模块，提供数据预处理和特征工程相关能力 |
-| [**paddlets.models**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/models/overview.html)       | 时序模型模块，基于飞桨深度学习框架PaddlePaddle的时序模型     |
-| [**paddlets.pipeline**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/pipeline/overview.html)   | 建模任务流模块，支持特征工程、模型训练、模型评估的任务流实现    |
-| [**paddlets.metrics**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/metrics/overview.html)     | 效果评估模块，提供多维度模型评估能力                         |
-| [**paddlets.analysis**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/analysis/overview.html)   | 数据分析模块，提供高效的时序特色数据分析能力                      |
-| [**paddlets.utils**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/backtest/overview.html)      | 工具集模块，提供回测等基础功能                           |
+| 模块                                                                                                            | 简述                                 |
+|---------------------------------------------------------------------------------------------------------------|------------------------------------|
+| [**paddlets.datasets**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/datasets/overview.html)   | 时序数据模块，统一的时序数据结构和预定义的数据处理方法        |
+| [**paddlets.autots**](https://paddlets.readthedocs.io/en/latest/source/modules/autots/overview.html)          | 自动超参寻优                             |
+| [**paddlets.transform**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/transform/overview.html) | 数据转换模块，提供数据预处理和特征工程相关能力            |
+| [**paddlets.models**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/models/overview.html)       | 时序模型模块，基于飞桨深度学习框架PaddlePaddle的时序模型 |
+| [**paddlets.pipeline**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/pipeline/overview.html)   | 建模任务流模块，支持特征工程、模型训练、模型评估的任务流实现     |
+| [**paddlets.metrics**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/metrics/overview.html)     | 效果评估模块，提供多维度模型评估能力                 |
+| [**paddlets.analysis**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/analysis/overview.html)   | 数据分析模块，提供高效的时序特色数据分析能力             |
+| [**paddlets.utils**](https://paddlets.readthedocs.io/zh_CN/latest/source/modules/backtest/overview.html)      | 工具集模块，提供回测等基础功能                    |
 
 
 ## 安装

diff --git a/README_en.md b/README_en.md
@@ -35,12 +35,17 @@ PaddleTS is an easy to use Python library for deep time series modeling,
 * A set of transformation operators for data preprocessing (e.g. missing values/outliers handling, 
     one-hot encoding, normalization, and automatic date/time-related covariate generation, etc.);
 * A set of analysis operators for quick data exploration (e.g. basic statistics and summary).
+* Automatic hyper-parameter tuning;
+* Third-party ML models & data transformations integration (e.g. scikit-learn).
+* Run models on GPU devices;
+* TS2Vec, a time series representation learning model;
+* DeepAR, a deep time series model for probabilistic forecasting.
 
 In future, more advanced features will be coming, including:
 
-* Automatic hyper-parameter tuning;
-* Time series representation learning models；
-* Add support for probabilistic forecasting;
+* Time series anomaly detection;
+* More time series representation learning models;
+* More probabilistic forecasting models;
 * Scenario-specific pipelines which aim to provide an end-to-end solution for solving real-world business problems;
 * And more.
 
@@ -53,6 +58,7 @@ Specifically, PaddleTS consists of the following modules:
 | Module                                                                                                     | Description                                                                                   |
 |------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------|
 | [**paddlets.datasets**](https://paddlets.readthedocs.io/en/latest/source/modules/datasets/overview.html)   | Unified time series representation (TSDataset) and data repository with pre-built TSDatasets. |
+| [**paddlets.autots**](https://paddlets.readthedocs.io/en/latest/source/modules/autots/overview.html)       | Automatic hyper-parameter tuning.                                                             |
 | [**paddlets.transform**](https://paddlets.readthedocs.io/en/latest/source/modules/transform/overview.html) | Data preprocessing and data transformations.                                                  |
 | [**paddlets.models**](https://paddlets.readthedocs.io/en/latest/source/modules/models/overview.html)       | PaddlePaddle-based deep neural network models for time series modeling.                       |
 | [**paddlets.pipeline**](https://paddlets.readthedocs.io/en/latest/source/modules/pipeline/overview.html)   | Pipeline for building time series analysis and modeling workflows.                            |

diff --git a/docs/conf.py b/docs/conf.py
@@ -40,9 +40,9 @@
 
 # Localization of Documentation.
 # https://docs.readthedocs.io/en/stable/localization.html
-locale_dirs = ['locale/']
+locale_dirs = ["locale/"]
 gettext_compact = False
-language = 'en'
+language = "en"
 add_module_names = False
 gettext_uuid = True
 
@@ -56,5 +56,13 @@
 # -- Options for EPUB output
 # epub_show_urls = "footnote"
 
+html_theme_options = {
+    # "collapse_navigation": True,
+    "display_version": True,
+    # "navigation_depth": 5,
+    # "navigation_with_keys": True,
+    "body_max_width": "80%"
+}
+
 # Html logo in drawer.
 html_logo = "static/images/logo/paddlets-rtd-logo.png"
diff --git a/docs/index.rst b/docs/index.rst
@@ -16,11 +16,21 @@ and excellent user experiences for practitioners and professionals. It’s featu
 
 * A set of analysis operators for quick data exploration (e.g. basic statistics and summary).
 
+* Automatic hyper-parameter tuning.
+
+* Third-party ML models & data transformations integration (e.g., scikit-learn).
+
+* Run models on GPU devices.
+
+* TS2Vec, a time series representation learning model.
+
+* DeepAR, a deep time series model for probabilistic forecasting.
+
 In future, more advanced features will be coming, including:
 
-* Automatic hyper-parameter tuning;
-* Time series representation learning models；
-* Add support for probabilistic forecasting;
+* Time series anomaly detection;
+* More time series representation learning models；
+* More probabilistic forecasting models;
 * Scenario-specific pipelines which aim to provide an end-to-end solution for solving real-world business problems;
 * And more.
 
@@ -33,6 +43,7 @@ Project GitHub: https://github.com/PaddlePaddle/PaddleTS
     :caption: Get Started
 
     Get Started <source/get_started/get_started.rst>
+    Run On GPU <source/get_started/run_on_gpu.rst>
 
 .. toctree::
     :maxdepth: 1
@@ -52,12 +63,16 @@ Project GitHub: https://github.com/PaddlePaddle/PaddleTS
     :caption: Transform
 
     Transform <source/modules/transform/overview.md>
+    Third-Party And User-Define Transform <source/modules/transform/thirdparty_userdefine.rst>
 
 .. toctree::
     :maxdepth: 1
     :caption: Models
 
-    Models <source/modules/models/overview.md>
+    Overview <source/modules/models/overview.rst>
+    Third-party Model <source/modules/models/thirdparty.rst>
+    Probability Forecasting <source/modules/models/probability_forecasting.rst>
+    Representation  <source/modules/models/representation.rst>
 
 
 .. toctree::
@@ -85,14 +100,22 @@ Project GitHub: https://github.com/PaddlePaddle/PaddleTS
 
     Backtest <source/modules/backtest/overview.md>
 
+.. toctree::
+    :maxdepth: 1
+    :caption: AutoTS
+
+    AutoTS <source/modules/autots/overview.rst>
+
 .. toctree::
     :maxdepth: 1
     :caption: API
 
     paddlets.analysis <source/api/paddlets.analysis.rst>
+    paddlets.automl <source/api/paddlets.automl.rst>
     paddlets.datasets <source/api/paddlets.datasets.rst>
     paddlets.metrics <source/api/paddlets.metrics.rst>
     paddlets.models <source/api/paddlets.models.rst>
     paddlets.pipeline <source/api/paddlets.pipeline.rst>
     paddlets.transform <source/api/paddlets.transform.rst>
     paddlets.utils <source/api/paddlets.utils.rst>
+