Comments (7)
Ok, reproducible problem.
The problem is located in the NER module.
When running with --skip=n, everything runs smoothly.
Looking into it now.
from frog.
Ah great, I don't actually need NER, so just tried it with --skip=n and that works perfectly now! Thanks!
from frog.
OK, I nailed it down this simple example:
<?xml-stylesheet type="text/xsl" href="folia2html.xsl"?>
<FoLiA xmlns="http://ilk.uvt.nl/folia" xmlns:xlink="http://www.w3.org/1999/xlink" xml:id="bug" version="0.12.0" generator="manual">
<metadata type="native">
<annotations>
<entity-annotation set="set1" />
</annotations>
</metadata>
<text xml:id="bug.text">
<p xml:id="bug.p">
<s xml:id="bug.p.s">
<w xml:id="bug.p.s.w.1">
<t>A.W.A.</t>
</w>
<w xml:id="bug.p.s.w.2">
<t>Kemme|I|actor</t>
</w>
<w xml:id="bug.p.s.w.3">
<t>MA</t>
</w>
<w xml:id="bug.p.s.w.4">
<t>.</t>
</w>
<entities>
<entity class="actor">
<wref id="bug.p.s.w.1" />
<wref id="bug.p.s.w.2" />
</entity>
</entities>
</s>
</p>
</text>
</FoLiA>
The problem is that some entities are already there. That shouldn't crash of course.
UPDATE: The problem is that the NER module tries to generate an xml:id based on the id of the entities layer. Which has no ID. Will fix this to generate NO id when not possible.
On a side note: the original document is rather 'odd' as it already is tokenized (very old ilktok) and has declarations for pos-annotations and such.
from frog.
Ah ok, that makes sense. Yeah to be fair I don't actually need those pos declarations, I probably need to take those out, just need my own NER tagset. The reason it's already tokenised is because it has been manually tagged for named entities (converted from highlighted word documents), and I believe it needs to be tokenised to set the NER tags in folia?
from frog.
To add Named entities, you need words, but frog will tokenize on beforehand, if not already done.
Leaving your own tokenization in, can be good. For instances when (like in this case) you already have references to words in your own entities.
I am a bit confused though, as you said you didn't need the NER.
Thinking about what you need is always good. Do you need the parser, the MWU unit, the Morphological analysis etc? If not, the process can be sped up a lot using --skip
I would suggest --skip=cmp at least , skipping chunking, MWU and parsing.
from frog.
problem fixed in GIT
from frog.
Thanks for fixing the bug Ko!
Regarding NER, I meant that I don't need Frog to perform NER, as these docs have already been named entity tagged manually. Good tip on using the --skip option, that should save some time, thanks!
from frog.
Related Issues (20)
- Frog Chunker creates invalid FoLiA HOT 2
- released frog (0.29) depends on unreleased libfolia (2.15) HOT 2
- Building on Ubuntu 22.04 LTS Pop!_OS HOT 1
- Regression: Frog no longer outputs both FoLiA XML and columned output when requested HOT 2
- segmentation fault when invoked with a missing [[tokenizer]] section in the configuration HOT 5
- Server mode creates only 1 paragraph HOT 2
- Add JSON output as an alternative to 'tabbed' format HOT 3
- Frog breaks while processing large amount of txt data HOT 11
- Keep the deep_morph structure intact when resolving MWU's HOT 1
- Simplify option and configuration handling
- MWU output when no Parser is selected HOT 7
- Update debian package for v0.20
- Python Frog HOT 2
- Frog (through python-frog) accumulates a huge number of temporary files HOT 11
- Praktische vragen rondom grote datasets HOT 7
- Bug: frog server; frog-:connection lost unexpected : write to client failed HOT 2
- Segfault on FoLiA in to FoLiA out (speech data with events and utterances) HOT 7
- New release? HOT 3
- frog lemmatizer with --deep-morph misses a morpheme in FoLiA output
- [Docker] Initialization fails for nld-vnn and dum HOT 7
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from frog.