Skip to content

Multiple improvements to results extraction #5

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

Merged
merged 49 commits into from
Apr 28, 2020
Merged

Multiple improvements to results extraction #5

merged 49 commits into from
Apr 28, 2020

Conversation

mkardas
Copy link
Collaborator

@mkardas mkardas commented Apr 28, 2020

No description provided.

mkardas and others added 30 commits December 10, 2019 16:58
* downgrade elasticsearch-dsl version due to compatibility issues
* make elastic connection setup optional
* make CRF loading optional
* disable multiprocessing in tokenization to avoid forking in celery
workers
* add evidences for all datasets and metrics found in taxonomy
* refactor evidence finding to EvidenceFinder
* fix manual evidences for semantic segmentation metrics
Remove stop words from automatically generated evidences.
Marcin Kardas added 19 commits February 17, 2020 17:35
* compute probabilities for each axis (tasks, datasets and metrics)
independently
* fix metric score extraction and conversion
* add abstract context
* add table contexts consinsting of paragraphs referencing a given table
* count evidences with repetitions
* stop adding abbreviations as datasets evidences
* fix passing paper context as string
* add confidence to non-filtered linking metrics to analyze results
* add threshold_map to make it easier to choose stable regions
Add a proposal with the best task, dataset and metric being choosen
independently.
Chrome DOM dumping was introduced as a cleaning method when htlatex
was used for LaTeX to html conversion. latexml gives much better
results and html5lib is 3 times faster than running chrome image.
Due to previous mismatch between version of elastic search instance
(6.7.1) and elasticsearch-dsl version (7.0.0) and the fact that
saving a document in non-existent index creates an index inferred from
the document, there's a mismatch between mapping type name. This commit
makes it possible to downgrade elasticsearch-dsl to version compatible
with elastic search 6.7.1 without a need for recreating the indexes.
@mkardas mkardas merged commit f442bd3 into master Apr 28, 2020
@mkardas mkardas deleted the push-api branch May 7, 2020 15:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant