token-regex.txt

## example token regular expression file
## Example: (parse "I like Teddy Grams on Tuesday")
## ({:entity-type "PRODUCT",
##   :token-range [2 4],
##   :sent-index 0,
##   :tok-re-ner-tag "PRODUCT",
##   :char-range [7 18],
##   :text "Teddy Grams"}
##  {:entity-type "DAY_OF_WEEK",
##   :token-range [5 6],
##   :sent-index 0,
##   :tok-re-ner-tag "DAY_OF_WEEK",
##   :char-range [22 29],
##   :text "Tuesday"})
##  ...
##    {:token-range [2 3],
##     :ner-tag "PERSON",
##     :pos-tag "NNP",
##     :lemma "Teddy",
##     :token-index 3,
##     :sent-index 0,
##     :tok-re-ner-features {:food-type "dessert"},
##     :tok-re-ner-tag "PRODUCT",
##     :tok-re-ner-item-id 497,
##     :char-range [7 12],
##     :text "Teddy"}
##  ...
## more examples:
## http://nlp.stanford.edu/software/tokensregex.html

# define ruleType to be over tokens
ENV.defaults["ruleType"] = "tokens"

# define keys for annotations when added
zstrner = {type:"CLASS",value:"zensols.stanford.nlp.TokenRegexAnnotations$NERAnnotation"}
zstrfeat = {type:"CLASS",value:"zensols.stanford.nlp.TokenRegexAnnotations$NERFeatureCreateAnnotation"}
zstriid = {type:"CLASS",value:"zensols.stanford.nlp.TokenRegexAnnotations$NERItemIDAnnotation"}

# Case insensitive pattern matching (see java.util.regex.Pattern flags)
#ENV.defaultStringPatternFlags = 2

# define rules
{pattern:(/(?iu)monday|tuesday|wednesday|thursday|friday|saturday|sunday/),action:(Annotate($0,zstrner,"DAY_OF_WEEK"))}
{pattern:([{word:"Teddy"}] [{lemma:"Grams"}]),action:(Annotate($0,zstrner,"PRODUCT"),Annotate($0,zstriid,"497"),Annotate($0,zstrfeat,"food-type={dessert}"))}