Add entity linking task #241

rmitsch · 2023-08-03T11:44:24Z

Description

Add entity linking task.

Corresponding documentation PR

explosion/spaCy#12988

Types of change

Checklist

I confirm that I have the right to submit this contribution under the project's MIT license.
I ran all tests in tests and usage_examples/tests, and all new and existing tests passed. This includes
- all external tests (i. e. pytest ran with --external)
- all tests requiring a GPU (i. e. pytest ran with --gpu)
My changes don't require a change to the documentation, or if they do, I've added all required information.

svlandeg

This will be great to have :-)

I have yet to play around with it more and I haven't reviewed the unit tests yet, but I figured I'd already send my first batch of review comments.

spacy_llm/tests/tasks/examples/entity_linking.json

spacy_llm/tasks/__init__.py

spacy_llm/tasks/entity_linking.py

spacy_llm/tasks/__init__.py

spacy_llm/tasks/entity_linking.py

# Conflicts: # .github/workflows/test_gpu.yml # spacy_llm/tasks/__init__.py # spacy_llm/tasks/builtin_task.py # spacy_llm/tasks/lemma/registry.py # spacy_llm/tasks/lemma/util.py # spacy_llm/tasks/ner/registry.py # spacy_llm/tasks/ner/util.py # spacy_llm/tasks/rel/registry.py # spacy_llm/tasks/rel/util.py # spacy_llm/tasks/sentiment/registry.py # spacy_llm/tasks/sentiment/util.py # spacy_llm/tasks/spancat/registry.py # spacy_llm/tasks/spancat/util.py # spacy_llm/tasks/summarization/registry.py # spacy_llm/tasks/summarization/util.py # spacy_llm/tasks/textcat/registry.py # spacy_llm/tasks/textcat/util.py # spacy_llm/ty.py

…ple.

svlandeg

I think the recent refactors have really improved the usability of this feature :-)

I'm not done reviewing yet, but need to break here for a bit, so sending my first batch of comments/suggestions.

usage_examples/el_openai/README.md

spacy_llm/tasks/entity_linker/registry.py

usage_examples/el_openai/fewshot.cfg

usage_examples/el_openai/zeroshot.cfg

spacy_llm/tasks/entity_linker/util.py

spacy_llm/tests/tasks/misc/el_kb_data.yml

usage_examples/el_openai/el_kb_data.yml

Co-authored-by: Sofie Van Landeghem <[email protected]>

usage_examples/el_openai/zeroshot.cfg

usage_examples/el_openai/fewshot.cfg

….yaml file.

svlandeg

Looking really good! This now works nicely with the NIL option, and the unit tests look great. I was able to get it to output different identifiers for the city of New York, the state, or using the string "New York" in a totally different context - it outputted NIL in that case. Nice!

Comments are only nitpicks.

One final issue to discuss is how to best support in-code experimenting with nlp.add_pipe("llm_entitylinker") and to add a unit test for that, too.

usage_examples/el_openai/README.md

spacy_llm/tasks/entity_linker/candidate_selector.py

spacy_llm/tasks/entity_linker/ty.py

spacy_llm/tasks/entity_linker/candidate_selector.py

spacy_llm/tasks/entity_linker/task.py

spacy_llm/tasks/entity_linker/util.py

spacy_llm/tests/pipeline/test_llm.py

spacy_llm/tests/tasks/test_entity_linker.py

Co-authored-by: Sofie Van Landeghem <[email protected]>

rmitsch · 2023-11-14T10:17:34Z

One final issue to discuss is how to best support in-code experimenting with nlp.add_pipe("llm_entitylinker") and to add a unit test for that, too.

Added a unit test in a631278.

spacy_llm/tests/tasks/test_entity_linker.py

svlandeg

Great! Let's get this merged 🎉

rmitsch added 6 commits August 1, 2023 15:34

Add foundation for EL task.

a224efc

Add foundation for EL task.

13aeb42

Continue building EL task.

629e5c7

Fix merge conflict.

d99429e

Fix merge conflicts. Finish zero-shot implementation.

114bfc7

Fix tests.

df371f8

rmitsch added feat/new New feature tests Everything related to the test suite feat/task Feature: tasks labels Aug 3, 2023

rmitsch self-assigned this Aug 3, 2023

rmitsch changed the base branch from main to develop August 3, 2023 11:44

rmitsch added 7 commits August 6, 2023 21:46

Create toy KB for EL tests.

1d86bf0

Fix LRU cache init.

1da5d1e

Fix remaining outdated paths.

6d50bcb

Fix Windows .csv issue (maybe).

428247a

Debug Windows .csv issue.

c00aeaa

Debug Windows .csv issue.

ab1c4cf

Clean up CSV parsing in CandidateSelector.

4c3ac2f

svlandeg reviewed Aug 21, 2023

View reviewed changes

rmitsch added 11 commits August 22, 2023 09:41

entity_linking -> entity_linker.

bfe7dd7

Migrate EL task to new task structure.

937c04d

Incorporate feedback.

0e7a28d

Incorporate feedback.

d6e347b

Use standardized entity candidate representation.

133e5da

Warn instead of raise if entity ID isn't in descriptions file.

c3d7a91

Ignore entities without KB IDs set instead of ignoring the whole exam…

e07f9c9

…ple.

Fix typing import.

770e985

Make entity description more robust.

0bacf74

Make entity description more robust.

b6c3da5

rmitsch added 7 commits October 13, 2023 11:40

Merge branch 'feat/el' of github.com:explosion/spacy-llm into feat/el

4566b71

Make embedding in EL data format optional.

bb54f3d

Fix processing of optional embeddings.

6c2277a

Merge branch 'develop' into feat/el

463506c

Merge branch 'develop' into feat/el

703b120

Revert docstring change.

d4c54e8

Merge branch 'develop' into feat/el

efdfd60

svlandeg reviewed Oct 19, 2023

View reviewed changes

Apply suggestions from code review

0b934c1

Co-authored-by: Sofie Van Landeghem <[email protected]>

rmitsch commented Oct 20, 2023

View reviewed changes

rmitsch added 3 commits October 20, 2023 09:28

Apply suggestions from code review

97a7bff

Rename KB loaders.

2e2895c

Change comment style. More informative error message on malformed KB …

f9991d5

….yaml file.

svlandeg reviewed Nov 9, 2023

View reviewed changes

rmitsch and others added 5 commits November 13, 2023 11:48

Add set_candidate_selector().

49b93da

Update spacy_llm/tests/pipeline/test_llm.py

5eb2167

Co-authored-by: Sofie Van Landeghem <[email protected]>

Update spacy_llm/tasks/entity_linker/util.py

2fb4465

Co-authored-by: Sofie Van Landeghem <[email protected]>

Update spacy_llm/tasks/entity_linker/task.py

c3b8931

Co-authored-by: Sofie Van Landeghem <[email protected]>

Apply suggestions from code review

a31624c

Co-authored-by: Sofie Van Landeghem <[email protected]>

rmitsch removed the Test external Run external tests label Nov 14, 2023

rmitsch added 2 commits November 14, 2023 09:35

Incorporate review feedback.

7ba003e

Merge branch 'develop' into feat/el

a117349

rmitsch added the Test external Run external tests label Nov 14, 2023

Add unit test for example with code only.

a631278

svlandeg reviewed Nov 14, 2023

View reviewed changes

spacy_llm/tests/tasks/test_entity_linker.py Outdated Show resolved Hide resolved

Check for candidate_selector being set in .initialize().

7770a13

svlandeg approved these changes Nov 16, 2023

View reviewed changes

svlandeg merged commit ea471bb into develop Nov 16, 2023

svlandeg deleted the feat/el branch November 16, 2023 21:09

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add entity linking task #241

Add entity linking task #241

rmitsch commented Aug 3, 2023 •

edited

Loading

svlandeg left a comment

svlandeg left a comment

svlandeg left a comment

rmitsch commented Nov 14, 2023

svlandeg left a comment

Add entity linking task #241

Add entity linking task #241

Conversation

rmitsch commented Aug 3, 2023 • edited Loading

Description

Corresponding documentation PR

Types of change

Checklist

svlandeg left a comment

Choose a reason for hiding this comment

svlandeg left a comment

Choose a reason for hiding this comment

svlandeg left a comment

Choose a reason for hiding this comment

rmitsch commented Nov 14, 2023

svlandeg left a comment

Choose a reason for hiding this comment

rmitsch commented Aug 3, 2023 •

edited

Loading