-
Notifications
You must be signed in to change notification settings - Fork 114
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add DisCoCirc extension to lambeq #179
Merged
Merged
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,30 @@ | ||
# DisCoCirc extension for lambeq | ||
|
||
Functionality to convert text into DisCoCirc string diagrams, using lambeq's grammar backend. | ||
|
||
## Installation | ||
|
||
Installing the experimental subpackage requires Python 3.10. | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. "Python 3.10 or higher", right? |
||
|
||
```bash | ||
git clone [email protected]:CQCL/lambeq.git | ||
cd lambeq | ||
pip install ".[experimental]" | ||
``` | ||
|
||
## Usage | ||
|
||
To get DisCoCirc diagrams using frames: | ||
|
||
```python | ||
from lambeq.experimental.discocirc import DisCoCircReader | ||
|
||
reader = DisCoCircReader() | ||
reader.text2circuit('Alice likes Bob. Bob likes Alice too.').draw() | ||
``` | ||
|
||
To get DisCoCirc diagrams with frames decomposed into multiple boxes: | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Need to add a reference here in one of the DisCoCirc papers that mentions Sandwich functor. |
||
|
||
```python | ||
reader.text2circuit('Alice likes Bob. Bob likes Alice too.', sandwich=True).draw() | ||
``` |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,29 @@ | ||
# Copyright 2021-2024 Cambridge Quantum Computing Ltd. | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
__all__ = ['DisCoCircReader', | ||
|
||
'CoreferenceResolver', | ||
'SpacyCoreferenceResolver', | ||
|
||
'TreeRewriter', | ||
'TreeRewriteRule'] | ||
|
||
from lambeq.experimental.discocirc.coref_resolver import ( | ||
CoreferenceResolver, | ||
SpacyCoreferenceResolver) | ||
from lambeq.experimental.discocirc.pregroup_tree_rewriter import ( | ||
TreeRewriter, | ||
TreeRewriteRule) | ||
from lambeq.experimental.discocirc.reader import DisCoCircReader |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,102 @@ | ||
# Copyright 2021-2024 Cambridge Quantum Computing Ltd. | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
from abc import ABC, abstractmethod | ||
|
||
import spacy | ||
|
||
|
||
class CoreferenceResolver(ABC): | ||
"""Class implementing corefence resolution.""" | ||
|
||
@abstractmethod | ||
def tokenise_and_coref( | ||
self, | ||
text: str | ||
) -> tuple[list[list[str]], list[list[list[int]]]]: | ||
"""Tokenise text and return its coreferences. | ||
|
||
Given a text consisting of possibly multiple sentences, | ||
return the sentences split into sentences and tokens. | ||
Additionally, return coreference information indicating tokens | ||
which correspond to the same entity. | ||
|
||
Parameters | ||
---------- | ||
text : str | ||
The text to tokenise. | ||
|
||
Returns | ||
------- | ||
list of list of str | ||
Each sentence in `text` as a list of tokens | ||
list of list of list of int | ||
Coreference information provided as a list for each | ||
coreferenced entity, consisting of a span for each sentence | ||
in `text`. | ||
|
||
""" | ||
|
||
def dict_from_corefs(self, | ||
corefs: list[list[list[int]]] | ||
) -> dict[tuple[int, int], tuple[int, int]]: | ||
"""Convert coreferences into a dict mapping each coreference to | ||
its first instance. | ||
|
||
Parameters | ||
---------- | ||
corefs : list[list[list[int]]] | ||
Coreferences as returned by `tokenise_and_coref` | ||
|
||
Returns | ||
------- | ||
dict[tuple[int, int], tuple[int, int]] | ||
Maps pairs of (sent index, tok index) to their first | ||
occurring coreference | ||
|
||
""" | ||
|
||
corefd = {} | ||
|
||
for coref in corefs: | ||
scorefs = [(i, scrf) for i, scoref in enumerate(coref) | ||
for scrf in scoref] | ||
|
||
for scoref in scorefs: | ||
corefd[scoref] = scorefs[0] | ||
|
||
return corefd | ||
|
||
|
||
class SpacyCoreferenceResolver(CoreferenceResolver): | ||
"""Corefence resolution and tokenisation based on spaCy.""" | ||
|
||
def __init__(self): | ||
self.nlp = spacy.load('en_coreference_web_trf', | ||
exclude=('span_resolver', 'span_cleaner')) | ||
|
||
def tokenise_and_coref(self, text): | ||
doc = self.nlp(text) | ||
coreferences = [] | ||
|
||
for cluster in doc.spans.values(): | ||
sent_clusters = [[] for _ in doc.sents] | ||
for span in cluster: | ||
for sent_cluster, sent in zip(sent_clusters, doc.sents): | ||
if sent.start <= span.start < sent.end: | ||
sent_cluster.append(span.start - sent.start) | ||
break | ||
coreferences.append(sent_clusters) | ||
|
||
return [[str(w) for w in s] for s in doc.sents], coreferences |
143 changes: 143 additions & 0 deletions
143
lambeq/experimental/discocirc/pregroup_tree_rewriter.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,143 @@ | ||
# Copyright 2021-2024 Cambridge Quantum Computing Ltd. | ||
# | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
from __future__ import annotations | ||
|
||
# __all__ = ['TreeRewriteRule', 'TreeRewriter'] | ||
|
||
from collections.abc import Iterable | ||
from dataclasses import replace | ||
|
||
from lambeq import AtomicType | ||
|
||
n = AtomicType.NOUN | ||
s = AtomicType.SENTENCE | ||
|
||
|
||
class TreeRewriteRule: | ||
"""General rewrite rule that merges tree nodes based on | ||
optional conditions.""" | ||
|
||
def __init__(self, | ||
match_type=False, | ||
match_words=None, | ||
max_depth=None, | ||
word_join='merge'): | ||
"""Instantiate a general rewrite rule""" | ||
self.match_type = match_type | ||
self.match_words = match_words | ||
self.max_depth = max_depth | ||
self.word_join = word_join | ||
|
||
def rewrite(self, node): | ||
return self.edit_tree(node)[0] | ||
|
||
def edit_tree(self, node): | ||
|
||
word_mergers = {'merge': lambda w1, w2: f'{w1} {w2}', | ||
'first': lambda w1, _: w1, | ||
'last': lambda _, w2: w2} | ||
|
||
if ((node.typ == self.match_type if self.match_type else True) | ||
and len(node.children) == 1 | ||
and node.children[0].typ == node.typ | ||
and (node.word.lower() in self.match_words | ||
if self.match_words else True)): | ||
# This node is one we want to contract with its child | ||
child, n_merges = self.edit_tree(node.children[0]) | ||
if self.max_depth is None or (n_merges < self.max_depth): | ||
return replace(child, | ||
word=word_mergers[self.word_join]( | ||
node.word, child.word) | ||
), n_merges + 1 | ||
# Not strictly necessary, but reduces eliminates recomputation | ||
return replace(node, children=[child]), n_merges | ||
|
||
return replace(node, | ||
children=[self.edit_tree(c)[0] | ||
for c in node.children]), 0 | ||
|
||
|
||
determiner_rule = TreeRewriteRule(match_type=n, | ||
match_words={'a', 'an', 'the'}, | ||
max_depth=1, | ||
word_join='last') | ||
|
||
auxiliary_rule = TreeRewriteRule(match_type=n.r@s, | ||
match_words={'has', 'had', 'have', | ||
'did', 'does', 'do'}, | ||
max_depth=1, | ||
word_join='last') | ||
|
||
|
||
noun_mod_rule = TreeRewriteRule(match_type=n, | ||
match_words=None, | ||
max_depth=None, | ||
word_join='merge') | ||
|
||
|
||
verb_mod_rule = TreeRewriteRule(match_type=n.r@s, | ||
match_words=None, | ||
max_depth=None, | ||
word_join='merge') | ||
|
||
sentence_mod_rule = TreeRewriteRule(match_type=s, | ||
match_words=None, | ||
max_depth=None, | ||
word_join='merge') | ||
|
||
|
||
class TreeRewriter: | ||
"""Class that rewrites a pregroup tree | ||
|
||
Comes with a set of default rules | ||
""" | ||
_default_rules = {'determiner': determiner_rule, | ||
'auxiliary': auxiliary_rule} | ||
|
||
_available_rules = {'determiner': determiner_rule, | ||
'auxiliary': auxiliary_rule, | ||
'noun_modification': noun_mod_rule, | ||
'verb_modification': verb_mod_rule, | ||
'sentence_modification': sentence_mod_rule} | ||
|
||
def __init__(self, | ||
rules: Iterable[TreeRewriteRule | str] | None = None | ||
) -> None: | ||
"""initialise a rewriter""" | ||
|
||
if rules is None: | ||
self.rules: list[TreeRewriteRule] = [*self._default_rules.values()] | ||
else: | ||
self.rules = [] | ||
self.add_rules(*rules) | ||
|
||
def add_rules(self, *rules: TreeRewriteRule | str) -> None: | ||
"""Add rules to this rewriter.""" | ||
for rule in rules: | ||
if isinstance(rule, TreeRewriteRule): | ||
self.rules.append(rule) | ||
else: | ||
try: | ||
self.rules.append(self._available_rules[rule]) | ||
except KeyError as e: | ||
raise ValueError( | ||
f'`{rule}` is not a valid rewrite rule.' | ||
) from e | ||
|
||
def __call__(self, node): | ||
"""Apply the rewrite rules to the given tree.""" | ||
for rule in self.rules: | ||
node = rule.rewrite(node) | ||
return node |
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reference to original paper.