Quote Extraction And Attribution
Table of contents
Description
Deterministically picks out quotes from a text. All top-level quotes, are supplied by the top level annotation for a text. If a QuotationAnnotation corresponds to a quote that contains embedded quotes, these quotes will appear as embedded QuotationAnnotations that can be accessed from the QuotationAnnotation that they are embedded in. The QuoteAnnotator can handle multi-line and cross-paragraph quotes, but any embedded quotes must be delimited by a different kind of quotation mark than its parents.
Allows for embedded quotations so long as they are either directed unicode quotes or are of a different type of quote than the outer quotations (e.g. “‘Gadzooks’ is what he said to me” is legal whereas “They called me “Danger” when I was…” is illegal). Uses regular-expression-like rules to find quotes and does not depend on the tokenizer, which allows quotes like ‘‘Tis true!’ to be correctly identified.
Considers regular ascii (“”, ‘’, ``’’, and `’) as well as “smart” and international quotation marks as follows: “”,‘’, «», ‹›, 「」, 『』, „”, and ‚’.
Note: extracts everything within these pairs as a whole quote segment, which may or may not be the desired behaviour for texts that use different formatting styles than standard english ones.
| Property name | Annotator class name | Generated Annotation |
|---|---|---|
| quote | QuoteAnnotator | QuotationsAnnotation, CharacterOffsetBeginAnnotation, CharacterOffsetEndAnnotation, QuotationIndexAnnotation, QuotationsAnnotation (if embedded quotes), TokensAnnotation (if tokenize present), TokenBeginAnnotation (if tokenize present), TokenEndAnnotation (if tokenize present), SentenceBeginAnnotation (if ssplit present), SentenceEndAnnotation (if ssplit present) |
Quote Attribution
As of version 3.9.1 of Stanford CoreNLP, the QuoteAnnotator will also try to attribute the quotes by default.
This can be deactivated by setting the quote.attributeQuotes property to false.
Sample Command Line
If you run this command:
java edu.stanford.nlp.pipeline.StanfordCoreNLP -annotators tokenize,pos,lemma,ner,depparse,coref,quote -file quote-example.txt -outputFormat text
On this text quote-example.txt:
In the summer Joe Smith decided to go on vacation. He said, "I'm going to Hawaii." That July, vacationer Joe went to Hawaii.
You should get this output for the quote in the text:
Extracted quotes:
[QuotationIndex=0, CharacterOffsetBegin=61, Text="I'm going to Hawaii.", Speaker=Joe Smith]
If you didn’t want to run quote attribution, you would add -quote.attributeQuotes false to your command.
Options
- quote.singleQuotes: “true” or “false”, indicating whether or not to consider ‘ tokens (single quotes) to be quotation marks (default=false).
- quote.maxLength: maximum character length of quotes to consider (default=-1).
- quote.asciiQuotes: “true” or “false”, indicating whether or not to convert all quotes to ascii quotes before processing (can help when there are errors in quote directionality) (default=false).
- quote.allowEmbeddedSame: “true” or “false” indicating whether or not to allow smart/directed (everything except “ and ‘) quotes of the same kind to be embedded within one another (default=false).