Skip to content

Commit df53e4c

Browse files
cthoytlvermuemberr
authoredJun 13, 2021
💯 🧼 Finalize PyKEEN 1.5.0 (pykeen#449)
* Fix entrypoints * Update translational_toy_example.rst * Update moves.py * Bump version: 1.4.1-dev → 1.5.0-dev * Update README * Update README.md * Add funding section * Add details box * Update header * Revert * Add pygments to doctests environment This is used by doctests for nice colored output * Update funding * Update author information in setup * Update README.md * Update setup.cfg * Update README.md * Improve grammar * Update conf.py * Update setup.cfg * Create .mailmap * Update CHANGELOG.rst * Final update Trigger CI * Update MANIFEST.in Trigger CI Co-authored-by: Laurent Vermue <[email protected]> Co-authored-by: Max Berrendorf <[email protected]>
1 parent 0d396bb commit df53e4c

File tree

12 files changed

+84
-24
lines changed

12 files changed

+84
-24
lines changed
 

‎.bumpversion.cfg

+1-1
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
[bumpversion]
2-
current_version = 1.4.1-dev
2+
current_version = 1.5.0-dev
33
commit = True
44
tag = False
55
parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(?:-(?P<release>[0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*))?(?:\+(?P<build>[0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*))?

‎.mailmap

+8
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
Laurent Vermue <lauve@dtu.dk> Laurent Vermue <l.vermue@me.com>
2+
Laurent Vermue <lauve@dtu.dk> Laurent Vermue <lvermue@users.noreply.github.com>
3+
Daniel Domingo-Fernández <danieldomingofernandez@hotmail.com> ddomingof <danieldomingofernandez@hotmail.com>
4+
Daniel Domingo-Fernández <danieldomingofernandez@hotmail.com> daniel.domingo.fernandez <danieldomingofernandez@hotmail.com>
5+
Mehdi Ali <ali-mehdi@live.de> mali <ali-mehdi@live.de>
6+
Mehdi Ali <ali-mehdi@live.de> Mehdi Ali <33023925+mali-git@users.noreply.github.com>
7+
Anna Breit <anna@breit-mail.at> anna <anna@breit-mail.at>
8+
Sahand Sharifzadeh <sahand.sharifzadeh@gmail.com>

‎CHANGELOG.rst

+37-6
Original file line numberDiff line numberDiff line change
@@ -5,8 +5,11 @@ All notable changes to this project will be documented in this file.
55
The format is based on `Keep a Changelog <http://keepachangelog.com/>`_
66
and this project adheres to `Semantic Versioning <http://semver.org/>`_
77

8-
`Unreleased <https://github.com/pykeen/pykeen/compare/v1.4.0...HEAD>`_
8+
`Unreleased <https://github.com/pykeen/pykeen/compare/v1.5.0...HEAD>`_
99
-----------------------------------------------------------------------
10+
11+
`1.5.0 <https://github.com/pykeen/pykeen/compare/v1.4.0...v1.5.0>`_ - 2021-06-13
12+
--------------------------------------------------------------------------------
1013
New Metrics
1114
~~~~~~~~~~~
1215
- Adjusted Arithmetic Mean Rank Index (https://github.com/pykeen/pykeen/pull/378)
@@ -22,32 +25,60 @@ New Models
2225
- QuatE (https://github.com/pykeen/pykeen/pull/367)
2326
- CompGCN (https://github.com/pykeen/pykeen/pull/382)
2427
- CrossE (https://github.com/pykeen/pykeen/pull/467)
28+
- Reimplementation of LiteralE with arbitrary combination (g) function (https://github.com/pykeen/pykeen/pull/245)
29+
30+
New Negative Samplers
31+
~~~~~~~~~~~~~~~~~~~~~
32+
- Pseudo-typed Negative Sampler (https://github.com/pykeen/pykeen/pull/412)
2533

2634
Datasets
2735
~~~~~~~~
28-
- Removed invalid datasets (OpenBioLink filtered sets; https://github.com/pykeen/pykeen/pull/#439)
36+
- Removed invalid datasets (OpenBioLink filtered sets; https://github.com/pykeen/pykeen/pull/https://github.com/pykeen/pykeen/pull/439)
37+
- Added WK3k-15K (https://github.com/pykeen/pykeen/pull/403)
38+
- Added WK3l-120K (https://github.com/pykeen/pykeen/pull/403)
39+
- Added CN3l (https://github.com/pykeen/pykeen/pull/403)
2940

3041
Added
3142
~~~~~
3243
- Documentation on using PyKEEN in Google Colab and Kaggle (https://github.com/pykeen/pykeen/pull/379,
3344
thanks `@jerryIsHere <https://github.com/jerryIsHere>`_)
34-
- Documentation on
3545
- Pass custom training loops to pipeline (https://github.com/pykeen/pykeen/pull/334)
3646
- Compatibility later for the fft module (https://github.com/pykeen/pykeen/pull/288)
3747
- Official Python 3.9 support, now that PyTorch has it (https://github.com/pykeen/pykeen/pull/223)
3848
- Utilities for dataset analysis (https://github.com/pykeen/pykeen/pull/16, https://github.com/pykeen/pykeen/pull/392)
49+
- Filtering of negative sampling now uses a bloom filter by default (https://github.com/pykeen/pykeen/pull/401)
50+
- Optional embedding dropout (https://github.com/pykeen/pykeen/pull/422)
51+
- Added more HPO suggestion methods and docs (https://github.com/pykeen/pykeen/pull/446)
52+
- Training callbacks (https://github.com/pykeen/pykeen/pull/429)
53+
- Class resolver for datasets (https://github.com/pykeen/pykeen/pull/473)
54+
3955
Updated
4056
~~~~~~~
4157
- R-GCN implementation now uses new-style models and is super idiomatic (https://github.com/pykeen/pykeen/pull/110)
42-
43-
Changed
44-
~~~~~~~
58+
- Enable passing of interaction function by string in base model class (https://github.com/pykeen/pykeen/pull/384,
59+
https://github.com/pykeen/pykeen/pull/387)
60+
- Bump scipy requirement to 1.5.0+
4561
- Updated interfaces of models and negative samplers to enforce kwargs (https://github.com/pykeen/pykeen/pull/445)
62+
- Reorganize filtering, negative sampling, and remove triples factory from most objects (
63+
https://github.com/pykeen/pykeen/pull/400, https://github.com/pykeen/pykeen/pull/405,
64+
https://github.com/pykeen/pykeen/pull/406, https://github.com/pykeen/pykeen/pull/409,
65+
https://github.com/pykeen/pykeen/pull/420)
66+
- Update automatic memory optimization (https://github.com/pykeen/pykeen/pull/404)
67+
- Flexibly define positive triples for filtering (https://github.com/pykeen/pykeen/pull/398)
68+
- Completely reimplemented negative sampling interface in training loops (https://github.com/pykeen/pykeen/pull/427)
69+
- Completely reimplemented loss function in training loops (https://github.com/pykeen/pykeen/pull/448)
70+
- Forward-compatibility of embeddings in old-style models and updated docs on
71+
how to use embeddings (https://github.com/pykeen/pykeen/pull/474)
4672

4773
Fixed
4874
~~~~~
4975
- Regularizer passing in the pipeline and HPO (https://github.com/pykeen/pykeen/pull/345)
5076
- Saving results when using multimodal models (https://github.com/pykeen/pykeen/pull/349)
77+
- Add missing diagonal constraint on MuRE Model (https://github.com/pykeen/pykeen/pull/353)
78+
- Fix early stopper handling (https://github.com/pykeen/pykeen/pull/419)
79+
- Fixed saving results from pipeline (https://github.com/pykeen/pykeen/pull/428, thanks @kantholtz)
80+
- Fix OOM issues with early stopper and AMO (https://github.com/pykeen/pykeen/pull/433)
81+
- Fix ER-MLP functional form (https://github.com/pykeen/pykeen/pull/444)
5182

5283
`1.4.0 <https://github.com/pykeen/pykeen/compare/v1.3.0...v1.4.0>`_ - 2021-03-04
5384
--------------------------------------------------------------------------------

‎MANIFEST.in

+1-1
Original file line numberDiff line numberDiff line change
@@ -15,5 +15,5 @@ include config/config.ini
1515

1616
global-exclude *.py[cod] __pycache__ *.so *.dylib .DS_Store *.gpickle
1717

18-
exclude .bumpversion.cfg .coveragerc .flake8 .travis.yml .readthedocs.yml tox.ini .pre-commit-config.yaml
18+
exclude .bumpversion.cfg .coveragerc .flake8 .travis.yml .readthedocs.yml tox.ini .pre-commit-config.yaml .mailmap
1919
include LICENSE *.md *.rst

‎README.md

+6-4
Original file line numberDiff line numberDiff line change
@@ -340,10 +340,12 @@ This project has been supported by several organizations (in alphabetical order)
340340

341341
The development of PyKEEN has been funded by the following grants:
342342

343-
| Funding Body | Program | Grant |
344-
|----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------|---------------|
345-
| DARPA | [Automating Scientific Knowledge Extraction (ASKE)](https://www.darpa.mil/program/automating-scientific-knowledge-extraction) | HR00111990009 |
346-
| German Federal Ministry of Education and Research (BMBF) | [Maschinelles Lernen mit Wissensgraphen (MLWin)](https://mlwin.de) | 01IS18050D |
343+
| Funding Body | Program | Grant |
344+
|----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------|-----------------|
345+
| DARPA | [Automating Scientific Knowledge Extraction (ASKE)](https://www.darpa.mil/program/automating-scientific-knowledge-extraction) | HR00111990009 |
346+
| German Federal Ministry of Education and Research (BMBF) | [Maschinelles Lernen mit Wissensgraphen (MLWin)](https://mlwin.de) | 01IS18050D |
347+
| German Federal Ministry of Education and Research (BMBF) | [Munich Center for Machine Learning (MCML)](https://mcml.ai) | 01IS18036A |
348+
| Innovation Fund Denmark (Innovationsfonden) | [Danish Center for Big Data Analytics driven Innovation (DABAI)](https://dabai.dk) | Grand Solutions |
347349

348350
### Logo
349351

‎docs/source/conf.py

+4-1
Original file line numberDiff line numberDiff line change
@@ -52,7 +52,7 @@
5252
author = 'PyKEEN Project Team'
5353

5454
# The full version, including alpha/beta/rc tags.
55-
release = '1.4.1-dev'
55+
release = '1.5.0-dev'
5656

5757
# The short X.Y version.
5858
parsed_version = re.match(
@@ -263,3 +263,6 @@
263263
}
264264

265265
autoclass_content = 'both'
266+
# autodoc_member_order = 'bysource'
267+
# autodoc_typehints = 'both' # TODO turn on after 4.1 release
268+
# autodoc_preserve_defaults = True

‎setup.cfg

+4-4
Original file line numberDiff line numberDiff line change
@@ -16,10 +16,10 @@ project_urls =
1616
Bug Tracker = https://github.com/pykeen/pykeen/issues
1717

1818
# Author information
19-
author = Mehdi Ali
20-
author_email = mehdi.ali@cs.uni-bonn.de
21-
maintainer = Mehdi Ali
22-
maintainer_email = mehdi.ali@cs.uni-bonn.de
19+
author = "Mehdi Ali, Max Berrendorf, Charles Tapley Hoyt, Laurent Vermue"
20+
author_email = "<mehdi.ali@cs.uni-bonn.de>, <max.berrendorf@gmail.com>, <cthoyt@gmail.com>, <lauve@dtu.dk>"
21+
maintainer = "Mehdi Ali, Max Berrendorf, Charles Tapley Hoyt, Laurent Vermue"
22+
maintainer_email = "<mehdi.ali@cs.uni-bonn.de>, <max.berrendorf@gmail.com>, <cthoyt@gmail.com>, <lauve@dtu.dk>"
2323

2424
# License information
2525
license = MIT

‎src/pykeen/losses.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -586,7 +586,7 @@ def __init__(self, margin: float = 9.0, adversarial_temperature: float = 1.0, re
586586
The name of the reduction operation to aggregate the individual loss values from a batch to a scalar loss
587587
value. From {'mean', 'sum'}.
588588
589-
.. note:: The default hyperparameters are based the experiments for FB15k-237 in [sun2019]_.
589+
.. note:: The default hyperparameters are based on the experiments for FB15k-237 in [sun2019]_.
590590
"""
591591
super().__init__(reduction=reduction)
592592
self.inverse_softmax_temperature = adversarial_temperature

‎src/pykeen/models/cli/builders.py

+11-1
Original file line numberDiff line numberDiff line change
@@ -9,12 +9,14 @@
99
from typing import Any, Mapping, Optional, Type, Union
1010

1111
import click
12+
from class_resolver import HintOrType
1213
from torch import nn
1314

1415
from . import options
1516
from .options import CLI_OPTIONS
1617
from ..base import Model
1718
from ...nn.message_passing import Decomposition
19+
from ...regularizers import Regularizer
1820
from ...triples import TriplesFactory
1921
from ...typing import Constrainer, Hint, Initializer, Normalizer
2022

@@ -44,6 +46,14 @@
4446
Optional[Mapping[str, Any]],
4547
Union[None, str, nn.Module],
4648
Union[None, str, Decomposition],
49+
50+
}
51+
_SKIP_HINTS = {
52+
Hint[Initializer],
53+
Hint[Constrainer],
54+
Hint[Normalizer],
55+
Hint[Regularizer],
56+
HintOrType[nn.Module],
4757
}
4858

4959

@@ -68,7 +78,7 @@ def _decorate_model_kwargs(command: click.Command) -> click.Command:
6878

6979
else:
7080
parameter = signature.parameters[name]
71-
if annotation in {Hint[Initializer], Hint[Constrainer], Hint[Normalizer]}: # type: ignore
81+
if annotation in _SKIP_HINTS:
7282
logger.debug('Unhandled hint: %s', annotation)
7383
continue
7484
if parameter.default is None:

‎src/pykeen/models/cli/options.py

+4
Original file line numberDiff line numberDiff line change
@@ -84,6 +84,10 @@ def _callback(_, __, value):
8484
'--clamp-score',
8585
type=float,
8686
),
87+
'combination_dropout': click.option(
88+
'--combination-dropout',
89+
type=float,
90+
),
8791
}
8892

8993
device_option = click.option(

‎src/pykeen/templates/README.md

+6-4
Original file line numberDiff line numberDiff line change
@@ -224,10 +224,12 @@ This project has been supported by several organizations (in alphabetical order)
224224

225225
The development of PyKEEN has been funded by the following grants:
226226

227-
| Funding Body | Program | Grant |
228-
|----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------|---------------|
229-
| DARPA | [Automating Scientific Knowledge Extraction (ASKE)](https://www.darpa.mil/program/automating-scientific-knowledge-extraction) | HR00111990009 |
230-
| German Federal Ministry of Education and Research (BMBF) | [Maschinelles Lernen mit Wissensgraphen (MLWin)](https://mlwin.de) | 01IS18050D |
227+
| Funding Body | Program | Grant |
228+
|----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------|-----------------|
229+
| DARPA | [Automating Scientific Knowledge Extraction (ASKE)](https://www.darpa.mil/program/automating-scientific-knowledge-extraction) | HR00111990009 |
230+
| German Federal Ministry of Education and Research (BMBF) | [Maschinelles Lernen mit Wissensgraphen (MLWin)](https://mlwin.de) | 01IS18050D |
231+
| German Federal Ministry of Education and Research (BMBF) | [Munich Center for Machine Learning (MCML)](https://mcml.ai) | 01IS18036A |
232+
| Innovation Fund Denmark (Innovationsfonden) | [Danish Center for Big Data Analytics driven Innovation (DABAI)](https://dabai.dk) | Grand Solutions |
231233

232234
### Logo
233235

‎src/pykeen/version.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -16,7 +16,7 @@
1616
'env',
1717
]
1818

19-
VERSION = '1.4.1-dev'
19+
VERSION = '1.5.0-dev'
2020

2121

2222
@lru_cache(maxsize=2)

0 commit comments

Comments
 (0)
Please sign in to comment.