cogcomp-nlp

cogcomp-nlp copied to clipboard

getview vs addview

2

@mssammon In `Annotator`, getView and addView functions seem to be doing the same thing (one lazy and one not-lazy). Is that really the difference, or anything else I'm missing?

danyaljj

The `mvn install` process fails on transliteration with error `error: unmappable character for encoding Cp1252` on https://github.com/CogComp/cogcomp-nlp/blob/master/transliteration/src/main/java/edu/illinois/cs/cogcomp/transliteration/Program.java#L508

Slash0BZ

How to pass parameter [on runtime] to a specific annotator of the AnnotatorService?

1

Imagine you want to pas some options to an annotator, through pipeline or any other annotator service. How can we do this? We discussed something similar to this, but when...

danyaljj

VerbSense classifier is using old version of SL

2

Currently [VerbSense is using an old version of SL](https://github.com/CogComp/cogcomp-nlp/blob/master/verbsense/pom.xml#L70-L72) which gets overwritten by a newer version when trying to access it in the pipeline. (hence verb-sense doesn't work in the...

danyaljj

Constituent hashCode collisions

2

We get a ton of collisions on the hashCode custom produced by the Constituent class.

cowchipkid

Why we never talk about the metrics in the similarity module? Like we should let users choose their metric e.g. cosine, etc. ...

danyaljj

In the similarity module add a table on the size of the vocabulary for each similarity resource.

danyaljj

Sentence annotator don't work when Period don't follow by a space.

1

``` 2013-08-05T22: 47: 53 China's Ministry of Public Security to crack down on the use of "pseudo-base station" to implement illegal and criminal activities arrested 217 suspects liuyizhan China's Ministry...

haowu40

NER not able to figure non-traditional numbers

2

This example: "This vulnerability happened in Firefox 1.x before 1.8, Safari 2.x before 2.8." not able to recognize that 1.x and 2.x is one single number.

danyaljj

Tokenization: Dot in the middle of a word.

3

@Slash0BZ do you think this has been affected by your change in tokenization?  "E.coli" should be one token, I think.

danyaljj