Text should be segmented by some tags, even if they don't end in full stop. Currently these should include:
Later, when/if they're not removed, they should include:
This should be configurable, like removed tags.
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Segment by tags | mediawiki/extensions/Wikispeech | master | +284 -31 |
Segmenting on tags should be followed by the ordinary segmenter.
So:
<h1> Heading. The Story. </h1> Text.
should give:
[ Heading., The Story., Text.]
When T158954: Use XPath to get text nodes related to utterances is done, this should be doable by adding "segment break" objects during cleaning and cutting segments by these.
Implemented locally. Will wait for patches under review to be done before rebasing and review.
Worked on in Wikispeech (Sprint 2017-04-25):
To do in Wikispeech (Sprint 2017-05-10):
Change 356827 had a related patch set uploaded (by Sebastian Berlin (WMSE); owner: Sebastian Berlin (WMSE)):
[mediawiki/extensions/Wikispeech@master] Segment by tags
Decided to upload this even though T148623: Highlight recited word is still under review, since there is minimal overlap in the changed code. The rebasing should be relatively painless.
Worked on in Wikispeech (Sprint 2017-05-24):
To do in Wikispeech (Sprint 2017-06-07):
Change 356827 had a related patch set uploaded (by Sebastian Berlin (WMSE); owner: Sebastian Berlin (WMSE)):
[mediawiki/extensions/Wikispeech@master] Segment by tags
Change 356827 had a related patch set uploaded (by Sebastian Berlin (WMSE); owner: Sebastian Berlin (WMSE)):
[mediawiki/extensions/Wikispeech@master] Segment by tags
Worked on in Wikispeech (Sprint 2017-06-21):
To do in Wikispeech (Sprint 2017-07-05):
Change 356827 merged by jenkins-bot:
[mediawiki/extensions/Wikispeech@master] Segment by tags
Mentioned in SAL (#wikimedia-cloud) [2017-07-06T11:30:48Z] <Sebastian-WMSE> Deploy latest from Git master: c766368, 5abcb1f, 63af3be (T149091)