Fix is_http for segmentation model to use token-based detection#669
Draft
de-code wants to merge 1 commit into
Draft
Fix is_http for segmentation model to use token-based detection#669de-code wants to merge 1 commit into
de-code wants to merge 1 commit into
Conversation
part of eLifePathways/ScienceBeam2.0#88 The segmentation model's FeatureDef called get_str_is_http(), which reads self.is_http. However, SegmentationLineFeaturesProvider reuses a single SegmentationLineFeatures instance and never sets is_http in its line loop, so the feature always returned '0'. Switched to get_str_is_http_token_based() (token_text.lower().startswith ('http')), which matches GROBID's Pattern.compile("http(s)?").find() check on the first token of each line. This affects 11 docs with 287 total Δlabels in the segmentation model, 142 of which cascade into header detection and downstream affiliation extraction.
ScienceBeam Parser EvaluationOverall (280 docs across 6 corpora)grobid 0.9.0-crf (default): 280 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 280 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 280 docs
biorxiv (47 docs)grobid 0.9.0-crf (default): 47 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 47 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 47 docs
ore (49 docs)grobid 0.9.0-crf (default): 49 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 49 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 49 docs
pkp (50 docs)grobid 0.9.0-crf (default): 50 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 50 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 50 docs
scielo_br (50 docs)grobid 0.9.0-crf (default): 50 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 50 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 50 docs
scielo_mx (34 docs)grobid 0.9.0-crf (default): 34 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 34 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 34 docs
scielo_preprints-jats (50 docs)grobid 0.9.0-crf (default): 50 docs | sciencebeam-parser:main-dc866983-20260616.1044 (grobid_crf): 50 docs | sciencebeam-parser:pr-669-fd92fdfd-20260616.1814 (grobid_crf): 50 docs
|
Collaborator
Author
|
For some reasons this actually caused more regressions |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
part of https://github.com/eLifePathways/ScienceBeam2.0/issues/102
The segmentation model's FeatureDef called get_str_is_http(), which reads self.is_http. However, SegmentationLineFeaturesProvider reuses a single SegmentationLineFeatures instance and never sets is_http in its line loop, so the feature always returned '0'.
Switched to get_str_is_http_token_based() (token_text.lower().startswith ('http')), which matches GROBID's Pattern.compile("http(s)?").find() check on the first token of each line. This affects 11 docs with 287 total Δlabels in the segmentation model, 142 of which cascade into header detection and downstream affiliation extraction.