如何在斯坦福 CoreNLP 服务器上使用自定义 TokensRegex 规则注释器？

TokensRegex 规则颜色注释器 (stanford-corenlp-full-2016-10-31/tokensregex/color.rules.txt）通过命令行使用 CoreNLP 时加载成功，但对于 Web 服务器则失败java.lang.IllegalArgumentException: Unknown annotator: color.

Setup

# custom.properties
annotators=tokenize,ssplit,pos,lemma,ner,regexner,color
customAnnotatorClass.color = edu.stanford.nlp.pipeline.TokensRegexAnnotator
color.rules = tokensregex/color.rules.txt

命令行

$ java -cp "*" -Xmx2g edu.stanford.nlp.pipeline.StanfordCoreNLP -props custom.properties -file ./tokensregex/color.input.txt -outputFormat text
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Registering annotator color with class edu.stanford.nlp.pipeline.TokensRegexAnnotator
...
[main] INFO edu.stanford.nlp.pipeline.StanfordCoreNLP - Adding annotator color
[main] INFO edu.stanford.nlp.ling.tokensregex.CoreMapExpressionExtractor - Reading TokensRegex rules from tokensregex/color.rules.txt
[main] INFO edu.stanford.nlp.ling.tokensregex.CoreMapExpressionExtractor - Read 7 rules

# color.input.txt.output
Sentence #1 (9 tokens):
Both blue and light blue are nice colors.
[Text=Both CharacterOffsetBegin=0 CharacterOffsetEnd=4 PartOfSpeech=CC Lemma=both NamedEntityTag=O]
[Text=blue CharacterOffsetBegin=5 CharacterOffsetEnd=9 PartOfSpeech=JJ Lemma=blue NamedEntityTag=COLOR NormalizedNamedEntityTag=#0000FF]
...

Server

1. java -mx2g -cp "*" edu.stanford.nlp.pipeline.StanfordCoreNLPServer -c custom.properties

2. wget --post-data 'Both blue and light blue are nice colors.' 'localhost:9000/?properties={"annotators":"tokenize,ssplit,pos,lemma,ner,regexner,color","outputFormat":"json"}' -O -

   HTTP request sent, awaiting response... 500 Internal Server Error
       2016-11-05 14:41:27 ERROR 500: Internal Server Error.
   
   java.lang.IllegalArgumentException: Unknown annotator: color
       at edu.stanford.nlp.pipeline.StanfordCoreNLP.ensurePrerequisiteAnnotators(StanfordCoreNLP.java:304)
       at edu.stanford.nlp.pipeline.StanfordCoreNLPServer$CoreNLPHandler.getProperties(StanfordCoreNLPServer.java:713)
       at edu.stanford.nlp.pipeline.StanfordCoreNLPServer$CoreNLPHandler.handle(StanfordCoreNLPServer.java:540)
       at com.sun.net.httpserver.Filter$Chain.doFilter(Filter.java:79)
       at sun.net.httpserver.AuthFilter.doFilter(AuthFilter.java:83)
       at com.sun.net.httpserver.Filter$Chain.doFilter(Filter.java:82)
       at sun.net.httpserver.ServerImpl$Exchange$LinkHandler.handle(ServerImpl.java:675)
       at com.sun.net.httpserver.Filter$Chain.doFilter(Filter.java:79)
       at sun.net.httpserver.ServerImpl$Exchange.run(ServerImpl.java:647)
       at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
       at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
       at java.lang.Thread.run(Thread.java:745)
   

Solution

在请求中包含自定义注释器属性：wget --post-data 'Both blue and light blue are nice colors.' 'localhost:9000/?properties={"color.rules":"tokensregex/color.rules.txt","customAnnotatorClass.color":"edu.stanford.nlp.pipeline.TokensRegexAnnotator","annotators":"tokenize,ssplit,pos,lemma,ner,regexner,color","enforceRequirements":"false","outputFormat":"json"}' -O -

Add

"enforceRequirements":"false"

根据您的要求，这应该会阻止此错误！