asyml · gpengzhi · Nov 18, 2019 · Nov 22, 2019 · Nov 22, 2019 · Nov 22, 2019
diff --git a/.pylintrc b/.pylintrc
@@ -13,7 +13,7 @@ ignore=CVS
 
 # Add files or directories matching the regex patterns to the blacklist. The
 # regex matches against base names, not paths.
-ignore-patterns=
+ignore-patterns=.*_test.py
 
 # Pickle collected data for later comparisons.
 persistent=yes
@@ -65,7 +65,20 @@ confidence=
 # --enable=similarities". If you want to run only the classes checker, but have
 # no Warning level messages displayed, use"--disable=all --enable=classes
 # --disable=W"
-disable=print-statement,parameter-unpacking,unpacking-in-except,old-raise-syntax,backtick,import-star-module-level,apply-builtin,basestring-builtin,buffer-builtin,cmp-builtin,coerce-builtin,execfile-builtin,file-builtin,long-builtin,raw_input-builtin,reduce-builtin,standarderror-builtin,unicode-builtin,xrange-builtin,coerce-method,delslice-method,getslice-method,setslice-method,no-absolute-import,old-division,dict-iter-method,dict-view-method,next-method-called,metaclass-assignment,indexing-exception,raising-string,reload-builtin,oct-method,hex-method,nonzero-method,cmp-method,input-builtin,round-builtin,intern-builtin,unichr-builtin,map-builtin-not-iterating,zip-builtin-not-iterating,range-builtin-not-iterating,filter-builtin-not-iterating,using-cmp-argument,long-suffix,old-ne-operator,old-octal-literal,suppressed-message,useless-suppression
+disable=invalid-name,too-many-branches,too-many-statements,too-many-arguments,  # Unnecessarily strict checks
+        too-many-instance-attributes,too-few-public-methods,too-many-locals,
+        too-many-lines,too-many-return-statements,too-many-boolean-expressions,
+        too-many-ancestors,
+        fixme,
+        no-else-return,no-else-raise,len-as-condition,unnecessary-pass,         # Not exactly good conventions
+        bad-continuation,inconsistent-return-statements,stop-iteration-return,
+        no-member,not-callable,invalid-unary-operand-type,arguments-differ,     # We have mypy for this
+        no-name-in-module,unsubscriptable-object,import-error,
+        access-member-before-definition,
+        redefined-builtin,abstract-method,missing-docstring,                    # Too many false positives
+        no-self-use,                                                            # (cannot ignore overridden methods)
+        unused-wildcard-import,                                                 # (https://github.com/rogalski/astroid/commit/82c6ef644a2efb77217a23d9b8a6cfb5caffb4ba)
+        duplicate-code,                                                         # (will be fixed in next release)
 
 
 [REPORTS]
@@ -82,7 +95,7 @@ output-format=text
 files-output=no
 
 # Tells whether to display a full report or only the messages
-reports=yes
+reports=no
 
 # Python expression which should return a note less than 10 (10 is the highest
 # note). You have access to the variables errors warning, statement which
@@ -253,7 +266,7 @@ ignore-comments=yes
 ignore-docstrings=yes
 
 # Ignore imports when computing similarities.
-ignore-imports=no
+ignore-imports=yes
 
 
 [VARIABLES]
@@ -263,7 +276,7 @@ init-import=no
 
 # A regular expression matching the name of dummy variables (i.e. expectedly
 # not used).
-dummy-variables-rgx=(_+[a-zA-Z0-9]*?$)|dummy
+dummy-variables-rgx=(_+[a-zA-Z0-9]*?$)|dummy|unused
 
 # List of additional names supposed to be defined in builtins. Remember that
 # you should avoid to define new builtins when possible.
@@ -291,7 +304,8 @@ logging-modules=logging
 max-line-length=80
 
 # Regexp for a line that is allowed to be longer than the limit.
-ignore-long-lines=^\s*(# )?<?https?://\S+>?$
+# This regex matches URLs and link anchors.
+ignore-long-lines=^\s*((# )?`?<?https?://\S+>?|\.\. _`.*`:|`.*`_)$
 
 # Allow the body of an if to be on the same line as the test if there is no
 # else.
@@ -314,7 +328,7 @@ indent-string='    '
 indent-after-paren=4
 
 # Expected format of line ending, e.g. empty (any line ending), LF or CRLF.
-expected-line-ending-format=
+expected-line-ending-format=LF
 
 
 [DESIGN]
@@ -367,7 +381,7 @@ valid-metaclass-classmethod-first-arg=mcs
 
 # List of member names, which should be excluded from the protected access
 # warning.
-exclude-protected=_asdict,_fields,_replace,_source,_make
+exclude-protected=_asdict,_fields,_replace,_source,_make,_get_name
 
 
 [IMPORTS]
@@ -399,6 +413,8 @@ known-third-party=enchant
 # only in one or another interpreter, leading to false positives when analysed.
 analyse-fallback-blocks=no
 
+allow-wildcard-with-all=yes
+
 
 [EXCEPTIONS]
 

diff --git a/.readthedocs.yml b/.readthedocs.yml
@@ -1,6 +1,6 @@
 version: 2
 
 python:
-  version: 3.6
+  version: 3.7
   install:
      - requirements: docs/requirements.txt
diff --git a/.travis.yml b/.travis.yml
@@ -10,26 +10,35 @@ python:
 install:
   - pip install --upgrade pip
   - pip install --progress-bar off .[tensorflow-cpu]
-  - pip install flake8==3.7.7
+  - pip install pylint==2.3.1 flake8==3.7.7
   - pip install pytest
+  - pip install coverage codecov
 
 script:
   # Linting
+  - pylint texar/ examples/
   - flake8 texar/ examples/
   # Unit tests
   - pytest
+  # Test coverage
+  - coverage run -m pytest
 
-jobs:
-  include:
-    - stage: docs
-      python: "3.7"
-      install:
-        - pip install --upgrade pip
-        - pip install --progress-bar off -r docs/requirements.txt
-      script:
-        - cd docs
-        # Build documents
-        - sphinx-build -b html -d _build/doctrees . _build/html
+after_success:
+  - codecov
+
+# jobs:
+#   include:
+#     - stage: docs
+#       python: "3.7"
+#       install:
+#         - pip install --upgrade pip
+#         - pip install --progress-bar off -r docs/requirements.txt
+#       script:
+#         - cd docs
+#           # Build documents
+#           - sphinx-build -W -b html -d _build/doctrees . _build/html
+#           # Check for typos
+#           - sphinx-build -W -b spelling -d _build/doctrees . _build/spelling
 
 notifications:
   email: false
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,12 +2,32 @@
 
 ### New features
 
-* Support only Python 3.6 and 3.7. Drop support of older Python versions.
+* Support TensorFlow 2.0.
 
 ### Feature improvements
 
 ### Fixes
 
+## [v0.2.4](https://github.com/asyml/texar/releases/tag/v0.2.4) (2019-11-18)
+
+### New features
+
+* Support only Python 3.6 and 3.7. Drop support of older Python versions. ([#211](https://github.com/asyml/texar/pull/211))
+* Add Tokenizers including tokenizers for pretrained models (`BERTTokenizer`, `XLNetTokenizer`, etc). ([#225](https://github.com/asyml/texar/pull/225))
+* Add GPT2 modules (`GPT2Encoder`, `GPT2Decoder`, `GPT2Classifier`, etc). ([#228](https://github.com/asyml/texar/pull/228))
+
+### Feature improvements
+
+* Update embedder modules `dropout_strategy=='item'` to support TensorFlow v1.15. ([#231](https://github.com/asyml/texar/pull/231))
+* Update `.gitignore` and add `.gitignore` files to all examples. ([#233](https://github.com/asyml/texar/pull/233)) 
+* Polish code style according to flake8. ([#234](https://github.com/asyml/texar/pull/234))
+* Add GPT2 XL pretrained checkpoint. ([#243](https://github.com/asyml/texar/pull/243))
+
+### Fixes
+
+* Fix `examples/transformer/scripts/wmt14_en_de.sh` to create output dir automatically. ([#238](https://github.com/asyml/texar/pull/238))
+* Fix variable scope issue in `texar.tf.modules.decoders.dynamic_decode`. ([#246](https://github.com/asyml/texar/pull/246))
+
 ## [v0.2.3](https://github.com/asyml/texar/releases/tag/v0.2.3) (2019-09-22)
 
 ### New features

diff --git a/LICENSE b/LICENSE
@@ -178,15 +178,15 @@
    APPENDIX: How to apply the Apache License to your work.
 
       To apply the Apache License to your work, attach the following
-      boilerplate notice, with the fields enclosed by brackets "{}"
+      boilerplate notice, with the fields enclosed by brackets "[]"
       replaced with your own identifying information. (Don't include
       the brackets!)  The text should be enclosed in the appropriate
       comment syntax for the file format. We also recommend that a
       file or class name and description of purpose be included on the
       same "printed page" as the copyright notice for easier
       identification within third-party archives.
 
-   Copyright {yyyy} {name of copyright owner}
+   Copyright [yyyy] [name of copyright owner]
 
    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.

diff --git a/README.md b/README.md
@@ -9,6 +9,8 @@
 [![Build Status](https://travis-ci.org/asyml/texar.svg?branch=master)](https://travis-ci.org/asyml/texar)
 [![Documentation Status](https://readthedocs.org/projects/texar/badge/?version=latest)](https://texar.readthedocs.io/en/latest/?badge=latest)
 [![License](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](https://github.com/asyml/texar/blob/master/LICENSE)
+
+# Texar TF2.0 is currently under incubation, please check back soon.
 
 
 **Texar** is a toolkit aiming to support a broad set of machine learning, especially natural language processing and text generation tasks. Texar provides a library of easy-to-use ML modules and functionalities for composing whatever models and algorithms. The tool is designed for both researchers and practitioners for fast prototyping and experimentation.
@@ -18,13 +20,14 @@
 * **Two Versions, (Mostly) Same Interfaces**. Texar-TensorFlow (this repo) and **[Texar-PyTorch](https://github.com/asyml/texar-pytorch)** have mostly the same interfaces. Both further combine the best design of TF and PyTorch:
   - Interfaces and variable sharing in *PyTorch convention*
   - Excellent factorization and rich functionalities in *TF convention*.
-* **Rich Pre-trained Models, Rich Usage with Uniform Interfaces**. BERT, GPT2, XLNet, etc, for encoding, classification, generation, and composing complex models with other Texar components!
-* **Fully Customizable** at multiple abstraction level -- both novice-friendly and expert-friendly. 
-  - Free to plug in whatever external modules, since Texar is fully compatible with the native TF/PyTorch APIs. 
-* **Versatile** to support broad tasks, models, algorithms, data processing, evaluation, etc. 
+* **Versatile** to support broad needs:
+   - data processing, model architectures, loss functions, training and inference algorithms, evaluation, ...
    - encoder(s) to decoder(s), sequential- and self-attentions, memory, hierarchical models, classifiers... 
    - maximum likelihood learning, reinforcement learning, adversarial learning, probabilistic modeling, ... 
+* **Fully Customizable** at multiple abstraction level -- both novice-friendly and expert-friendly. 
+  - Free to plug in whatever external modules, since Texar is fully compatible with the native TensorFlow/PyTorch APIs. 
 * **Modularized** for maximal re-use and clean APIs, based on principled decomposition of *Learning-Inference-Model Architecture*. 
+* **Rich Pre-trained Models, Rich Usage with Uniform Interfaces**. BERT, GPT2, XLNet, etc, for encoding, classification, generation, and composing complex models with other Texar components!
 * **Distributed** model training with multiple GPUs.
 * Clean, detailed [documentation](https://texar.readthedocs.io) and rich [examples](./examples).
 
@@ -33,8 +36,13 @@
    <img src="./docs/_static/img/texar_stack.png"><br><br>
 </div> 
 
+<div align="center">
+   <img src="./docs/_static/img/texar_modules_big.png"><br><br>
+</div> 
+
 ### Library API Example
 Builds an encoder-decoder model, with maximum likelihood learning:
+
 ```python
 import texar.tf as tx
 
@@ -72,6 +80,7 @@ outputs_bs, _, _ = tx.modules.beam_search_decode(
     end_token=data.target_vocab.eos_token_id)
 ```
 The same model, but with adversarial learning:
+
 ```python
 helper = tx.modules.GumbelSoftmaxTraingHelper( # Gumbel-softmax decoding
     start_tokens=[BOS]*batch_size, end_token=EOS, embedding=embedder)
@@ -85,6 +94,7 @@ G_loss, D_loss = tx.losses.binary_adversarial_losses(
     discriminator_fn=discriminator)
 ```
 The same model, but with RL policy gradient learning:
+
 ```python
 agent = tx.agents.SeqPGAgent(samples=outputs.sample_id,
                              logits=outputs.logits,
@@ -99,16 +109,18 @@ Many more examples are available [here](./examples)
 
 Texar requires:
 
-* `tensorflow >= 1.10.0 (but < 2.0.0)`. Follow the [tensorflow official instructions](https://www.tensorflow.org/install) to install the appropriate version
-* `tensorflow_probability >= 0.3.0 (but < 0.8.0)`. Follow the [tensorflow_probability official instractions](https://www.tensorflow.org/probability/install) to install.
+* `tensorflow >= 2.0.0`. Follow the [tensorflow official instructions](https://www.tensorflow.org/install) to install the appropriate version
+* `tensorflow_probability >= 0.3.0`. Follow the [tensorflow_probability official instractions](https://www.tensorflow.org/probability/install) to install.
 
 After `tensorflow` and `tensorflow_probability` are installed, install Texar from PyPI: 
+
 ```bash
 pip install texar
 ```
 
 To use cutting-edge features or develop locally, install from source: 
-```
+
+```bash
 git clone https://github.com/asyml/texar.git
 cd texar
 pip install .
@@ -120,6 +132,7 @@ pip install .
 
 ### Reference
 If you use Texar, please cite the [tech report](https://arxiv.org/abs/1809.00794) with the following BibTex entry:
+
 ```
 Texar: A Modularized, Versatile, and Extensible Toolkit for Text Generation
 Zhiting Hu, Haoran Shi, Bowen Tan, Wentao Wang, Zichao Yang, Tiancheng Zhao, Junxian He, Lianhui Qin, Di Wang, Xuezhe Ma, Zhengzhong Liu, Xiaodan Liang, Wanrong Zhu, Devendra Sachan and Eric Xing

diff --git a/docs/_static/img/texar_modules_big.png b/docs/_static/img/texar_modules_big.png
diff --git a/examples/README.md b/examples/README.md
@@ -3,107 +3,3 @@
 Rich examples are included to demonstrate the use of Texar. The implementations of cutting-edge models/algorithms also provide references for reproducibility and comparisons. 
 
 More examples are continuously added...
-
-## Examples by Models/Algorithms ##
-
-### RNN / Seq2seq ###
-
-* [language_model_ptb](./language_model_ptb): Basic RNN language model
-* [distributed_gpu](./distributed_gpu): Basic RNN language model with distributed training
-* [seq2seq_attn](./seq2seq_attn): Attentional seq2seq
-* [seq2seq_configs](./seq2seq_configs): Seq2seq implemented with Texar model template
-* [seq2seq_rl](./seq2seq_rl): Attentional seq2seq trained with policy gradient
-* [seq2seq_exposure_bias](./seq2seq_exposure_bias): Various algorithms tackling exposure bias in sequence generation
-* [hierarchical_dialog](./hierarchical_dialog): Hierarchical recurrent encoder-decoder model for conversation response generation
-* [torchtext](./torchtext): Use of torchtext data loader
-
-### Transformer (Self-attention) ###
-
-* [transformer](./transformer): Transformer for machine translation
-* [bert](./bert): Pre-trained BERT model for text representation
-* [gpt-2](./gpt-2): Pre-trained OpenAI GPT-2 language model
-* [vae_text](./vae_text): VAE with a transformer decoder for improved language modeling 
-
-### Variational Autoencoder (VAE) ###
-
-* [vae_text](./vae_text): VAE language model
-
-### GANs / Discriminiator-supervision ###
-
-* [seqGAN](./seqgan): GANs for text generation
-* [text_style_transfer](./text_style_transfer): Discriminator supervision for controlled text generation
-
-### Reinforcement Learning ###
-
-* [seq2seq_rl](./seq2seq_rl): Attentional seq2seq trained with policy gradient.
-* [seqGAN](./seqgan): Policy gradient for sequence generation
-* [rl_gym](./rl_gym): Various RL algoritms for games on OpenAI Gym
-
-### Memory Network ###
-
-* [memory_network_lm](./memory_network_lm): End-to-end memory network for language modeling
-
-### Classifier / Sequence Prediction ###  
-
-* [bert](./bert): Pre-trained BERT model for text representation
-* [sentence_classifier](./sentence_classifier): Basic CNN-based sentence classifier
-* [sequence_tagging](./sequence_tagging): BiLSTM-CNN model for Named Entity Recognition (NER)
-
-### Reward Augmented Maximum Likelihood (RAML) ###
-
-* [seq2seq_exposure_bias](./seq2seq_exposure_bias): RAML and other learning algorithms for sequence generation 
-
----
-
-## Examples by Tasks
-
-### Language Modeling ###
-
-* [gpt-2](./gpt-2): Pre-trained OpenAI GPT-2 language model
-* [language_model_ptb](./language_model_ptb): Basic RNN language model
-* [vae_text](./vae_text): VAE language model
-* [seqGAN](./seqgan): GAN + policy gradient
-* [memory_network_lm](./memory_network_lm): End-to-end memory network for language modeling
-
-### Machine Translation ###
-
-* [seq2seq_attn](./seq2seq_attn): Attentional seq2seq
-* [seq2seq_configs](./seq2seq_configs): Seq2seq implemented with Texar model template.
-* [seq2seq_rl](./seq2seq_rl): Attentional seq2seq trained with policy gradient.
-* [seq2seq_exposure_bias](./seq2seq_exposure_bias): Various algorithms tackling exposure bias in sequence generation (MT and summarization as examples).
-* [transformer](./transformer): Transformer for machine translation
-
-### Dialog ###
-
-* [hierarchical_dialog](./hierarchical_dialog): Hierarchical recurrent encoder-decoder model for conversation response generation.
-
-### Text Summarization ###
-
-* [seq2seq_exposure_bias](./seq2seq_exposure_bias): Various algorithms tackling exposure bias in sequence generation (MT and summarization as examples).
-
-### Text Style Transfer ###
-
-* [text_style_transfer](./text_style_transfer): Discriminator supervision for controlled text generation
-
-### Classification ###
-
-* [bert](./bert): Pre-trained BERT model for text representation
-* [sentence_classifier](./sentence_classifier): Basic CNN-based sentence classifier
-
-### Sequence Tagging ###
-
-* [sequence_tagging](./sequence_tagging): BiLSTM-CNN model for Named Entity Recognition (NER)
-
-### Games ###
-
-* [rl_gym](./rl_gym): Various RL algoritms for games on OpenAI Gym
-
----
-
-## MISC ##
-
-### Distributed training ###
-
-* [distributed_gpu](./distributed_gpu): Basic example of distributed training.
-* [bert](./bert): Distributed training of BERT.
-