Pip installable package

.gitignore

@@ -0,0 +1,160 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

NewsRoomEval.py

@@ -1,7 +1,7 @@
 import argparse
 import time
 import numpy as np
-from metrics.utils import *
+from disco_score.metrics.utils import *
 
 # We adapt the evaluation framework of BARTScore to evaluate DiscoScore on NewsRoom.
 # See https://github.com/neulab/BARTScore/blob/main/SUM/score.py
@@ -56,7 +56,7 @@ def score(self):
 
         if self.args.m == 'bart_score' or self.args.m == 'bart_score_cnn' or self.args.m == 'bart_score_para':
             """ Vanilla BARTScore, BARTScore-CNN, BARTScore-CNN-Para """
-            from metrics.bart_score import BARTScorer
+            from disco_score.metrics.bart_score import BARTScorer
             # Set up BARTScore
             if 'cnn' in self.args.m:
                 bart_scorer = BARTScorer(device=self.device, checkpoint='facebook/bart-large-cnn')
@@ -108,7 +108,7 @@ def score(self):
 
         elif self.args.m.startswith('DS_Focus') or self.args.m.startswith('DS_SENT'):
 
-            from metrics.metrics import Metrics
+            from disco_score.metrics.metrics import Metrics
 
             eval_metric = getattr(Metrics(args = self.args), self.args.m)                        
 

Newsroom/analysis_Newsroom.py

@@ -4,7 +4,7 @@
 import numpy as np
 import collections
 import pickle 
-from metrics.utils import *
+from disco_score.metrics.utils import *
 
 class SUMStat:
     """ A class used to get stats of SUM trained data """

README.md

@@ -10,6 +10,12 @@ DiscoScore is a parameterized, reference-based evaluation, with parameters on th
 
 # Usage
 
+Install from GitHub with pip:
+
+```bash
+pip install "git+https://github.com/AIPHES/DiscoScore.git"
+```
+
 Please run [SUMMEval.py](https://github.com/AIPHES/DiscoScore/blob/main/SUMMEval.py) to reproduce the results in our paper. 
 
 Note that DS_Focus uses [Conpono](https://drive.google.com/drive/folders/1FE2loCSfdBbYrYk_qHg6W_PTqvA9w46T?usp=sharing) (finetuned BERT on discourse corpora), and DS_SENT uses [BERT-NLI](https://drive.google.com/drive/folders/19-6TgHdfAVL6xzpqzMoTpxXKLkXOCBiO?usp=sharing) (finetuned BERT on NLI). Because this configuration performs best in our intial trials (see the paper for details). 
@@ -23,9 +29,12 @@ Note that if system and reference texts do not contain coherence phenomena (e.g.
 
 ```python
 from DiscoScorer import DiscoScorer
+
 disco_scorer = DiscoScorer(device='cuda:0', model_name='bert-base-uncased') # use Conpono for DS-Focus and BERT-NLI for DS-SENT for better results
+
 system = ["Paul Merson has restarted his row with andros townsend after the Tottenham midfielder was brought on with only seven minutes remaining in his team 's 0-0 draw with burnley. Townsend was brought on in the 83rd minute for Tottenham as they drew 0-0 against Burnley ."]
 references = [["Paul Merson has restarted his row with burnley on sunday. Townsend was brought on in the 83rd minute for tottenham. Andros Townsend scores england 's equaliser in their 1-1 friendly draw. Townsend hit a stunning equaliser for england against italy."]]
+
 for s, refs in zip(system, references):
    s = s.lower()
    refs = [r.lower() for r in refs]

SUMMEval.py

@@ -88,7 +88,7 @@ def generate_human_scores(self, num_instances, qa):
 
 
 import argparse
-from metrics.metrics import Metrics
+from disco_score.metrics.metrics import Metrics
 from scipy.stats import kendalltau                        
 
 parser = argparse.ArgumentParser()

disco_score/__init__.py

@@ -0,0 +1,3 @@
+__version__ = "0.1.0"
+
+from .scorer import DiscoScorer

metrics/discourse.py → disco_score/metrics/discourse.py

@@ -120,7 +120,7 @@ def _safe_divide(numerator, denominator):
 
 def DS_Focus(model, tokenizer, sys, ref, is_semantic_entity=False, we=None, threshold=0):
 
-    from metrics.focus_diff import get_embeddings
+    from disco_score.metrics.focus_diff import get_embeddings
 
     scores = []
     s_embedding, s_noun_positions = get_embeddings(model, tokenizer, sys, is_semantic_entity, 
@@ -147,7 +147,7 @@ def DS_Focus(model, tokenizer, sys, ref, is_semantic_entity=False, we=None, thre
 
 def DS_Sent(model, tokenizer, sys, ref, is_lexical_graph=False, we=None, threshold=0):
 
-    from metrics.sent_graph import get_embeddings
+    from disco_score.metrics.sent_graph import get_embeddings
 
     scores = []
     sys_u_a, num_sentences = EntityGraph(sys, is_lexical_graph, we, threshold)    

metrics/metrics.py → disco_score/metrics/metrics.py

@@ -1,5 +1,5 @@
-from metrics import discourse
-from metrics.word_embeddings import load_embeddings
+from disco_score.metrics import discourse
+from disco_score.metrics.word_embeddings import load_embeddings
 from transformers import *
 import torch
 

DiscoScorer.py → disco_score/scorer.py

@@ -1,5 +1,5 @@
-from metrics import discourse
-from metrics.word_embeddings import load_embeddings
+from disco_score.metrics import discourse
+from disco_score.metrics.word_embeddings import load_embeddings
 from transformers import BertConfig, BertTokenizer, BertModel
 import torch
 

setup.py

@@ -0,0 +1,19 @@
+import setuptools
+
+setuptools.setup(
+    name="disco_score",
+    version="0.1.0",
+    description="DiscoScore: Evaluating Text Generation with BERT and Discourse Coherence",
+    author="Wei Zhao",
+    url="https://github.com/AIPHES/DiscoScore",
+    python_requires=">=3.6",
+    packages=setuptools.find_packages(),
+    install_requires=[
+        "torch>=1.6.0",
+        "transformers>=4.6.1",
+        "nltk>=3.7.0",
+        "spacy>=3.3.1",
+        "spacy_udpipe>=1.0.0",
+        "networkx>=2.8.4",
+    ],
+)