harisont / concept-alignment Goto Github PK
View Code? Open in Web Editor NEWSyntax-based Concept Alignment for Machine Translation
License: BSD 2-Clause "Simplified" License
Syntax-based Concept Alignment for Machine Translation
License: BSD 2-Clause "Simplified" License
e.g.
lt ||| rt...
instead of
lt|rt...
prAlignment
EvAlign
I guess this involves function getAlignments
in GG.
I mean instead of just manipulating strings.
See if it's possible to use them to replace divergence patterns expressed as functions to:
I'm running extract-concepts
on a conllu file that has sentences like the following, starting with a comment:
# <http://snomed.info/id/188462001>
1 Secondary secondary ADJ <*>.A.ABS.<SS> Degree=Pos 3 amod _ _
2 malignant malignant ADJ A.ABS Degree=Pos 3 amod _ _
3 neoplasm neoplasm NOUN N.NOM.SG Case=Nom|Number=Sing 0 root _ _
4 of of ADP PREP AdpType=Prep 5 case _ _
5 brain brain NOUN <Count>.N.NOM.SG Case=Nom|Number=Sing 3 nmod _ _
6 and and CCONJ CC.@CC _ 7 cc _ _
7 spinal cord spinal cord NOUN <Count>.N.NOM.SG Case=Nom|Number=Sing 5 conj _ _
In the result, the original comment has been replaced with new ones:
# sent_id = gfud1000002
# text = of brain and spinal cord
# reasons: {POS, UD} sentence IDs: {"missing sent_id"} correctness: _
1 of of ADP PREP AdpType=Prep 2 case _ ADJUSTED=True
2 brain brain NOUN <Count>.N.NOM.SG Case=Nom|Number=Sing 0 root _ ADJUSTED=True|ORIG_LABEL=nmod
3 and and CCONJ CC.@CC _ 4 cc _ ADJUSTED=True
4 spinal cord spinal cord NOUN <Count>.N.NOM.SG Case=Nom|Number=Sing 2 conj _ ADJUSTED=True
I would like to know where the fragment came from, so I'd like the original comment to be retained in the output, something like the following. Placement doesn't matter, just that it is there somewhere.
# sent_id = gfud1000002
# text = of brain and spinal cord
# <http://snomed.info/id/188462001>
…
Kinda alternative to #14.
Not necessarily to be implemented here, maybe just use someone else's code.
Use only the new Alignment
type (requires #17) and define the necessary show/prints.
i.e. read and write conllu format, so that judgement are useful for GG as well as debugging. TBD what the role of the "linearized format" is, if any.
evalign
+
or only =
alignment in CP/lexicon generation if the conllu files are annotatedSo to make it easier to search them in the synoptic CoNNL-U viewer.
sentIds
Generalize from pairs to lists.
This will take some time.
Kinda alternative to #15.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.