Basic CluProcessor annotation fails for relatively simple operation

[reynoldsm88]

Description

I would like to use CluProcessor to go through a list of words and lemmatize them. In the past I used the code similar to the code snippet I've provided, however, now this operation is causing an exception related to something with WordEmbeddings during a sanity check of the document.

I'm not sure if this is something I'm doing wrong or if this is a bug.

Versions

• java : 1.8.0_202
• scala : 2.12.7
• clulab-processors-main : 8.4.2

Code

The following code results in an exception in my test case...

val processor = new CluProcessor()
val doc = processor.mkDocument( "counties" )
processor.lemmatize( doc )

Stacktrace

[info]   at org.clulab.processors.clu.CluProcessor.basicSanityCheck(CluProcessor.scala:520)
[info]   at org.clulab.processors.clu.CluProcessor.lemmatize(CluProcessor.scala:340)
[info]   at com.acme.MyTestSuite.$anonfun$new$1(MyTestSuite.scala:21)
[info]   at org.scalatest.OutcomeOf.outcomeOf(OutcomeOf.scala:85)
[info]   at org.scalatest.OutcomeOf.outcomeOf$(OutcomeOf.scala:83)
[info]   at org.scalatest.OutcomeOf$.outcomeOf(OutcomeOf.scala:104)
[info]   at org.scalatest.Transformer.apply(Transformer.scala:22)
[info]   at org.scalatest.Transformer.apply(Transformer.scala:20)
[info]   at org.scalatest.flatspec.AnyFlatSpecLike$$anon$5.apply(AnyFlatSpecLike.scala:1683)
[info]   at org.scalatest.TestSuite.withFixture(TestSuite.scala:196)

[MihaiSurdeanu]

Thanks @reynoldsm88 !

It seems to me that one dependency (the one that handles word embeddings) is not configured correctly. Because of its size, we store it on a local artifactory install, which must be added to the list of repos in the build file. Here is one build file that exemplifies this:

https://github.com/clulab/habitus/blob/main/build.sbt

I don't know if @kwalcock wants to add anything to this.

[BeckySharp]

@MihaiSurdeanu do we not need proc.tagPartsOfSpeech(doc) first?

[MihaiSurdeanu]

Not for the English CluProcessor, afaik.

…

On Wed, Sep 8, 2021 at 13:25 Becky Sharp ***@***.***> wrote: @MihaiSurdeanu <https://github.com/MihaiSurdeanu> do we not need proc.tagPartsOfSpeech(doc) first? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#548 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAI75TVQ2RS4MJTVRJ47MEDUA7BCHANCNFSM5DVS4EGA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[kwalcock]

This seems to work. YMMV. It may be that the sanity check is overreacting if only lemmatize is going to be run.

package org.clulab.processors.clu

import org.clulab.dynet.Utils

object LemmatizeApp extends App {
  Utils.initializeDyNet()

  val processor = new CluProcessor()
  val doc = processor.mkDocument("counties")

  processor.mkConstEmbeddings(doc)
  processor.lemmatize(doc)
}

[reynoldsm88]

Thanks for the quick response @kwalcock @MihaiSurdeanu and @BeckySharp!

Keith's fix did the trick, everything is working as usual again. Do you want me to mark the issue as resolved? Otherwise, feel free to close it as resolved.

[MihaiSurdeanu]

Thanks @reynoldsm88 !