Question

I’m working on a short bash script to grab a JSON element from a curl response.

curl -H "api_key:[API_PASSWORD]" http://api.wordnik.com/v4/word.json/button/pronunciations?sourceDictionary=macmillan&typeFormat=IPA&useCanonical=false 

returns:

[{"id":0,"seq":0,"raw":"ˈbʌt(ə)n","rawType":"IPA"},{"id":0,"seq":0,"raw":"ˈbʌt(ə)n","rawType":"IPA"}]

I’m trying to extract the “ˈbʌt(ə)n” element.

Though I’m unfamiliar with regex, I think I should be using a substitution with this string:

/.*"(.*)",/

I’m trying to run the following command, but it doesn’t seem to work:

curl -H "api_key:[API_KEY]" http://api.wordnik.com/v4/word.json/button/pronunciations?sourceDictionary=macmillan&typeFormat=IPA&useCanonical=false | sed /.*"(.*)",\1/

I’m sure there are a few things I’m doing wrong, and after a few hours of searching and reading up on regex and bash I’m out of options.

I don’t need to be using sed, I am simply looking for a quick way of doing this in a bash command line so that I can implement it in a TextExpander script on the mac.

Answer 1

Use STRING : REGEXP to extract the value from the json string:

string=$(curl -H "api_key:[API_PASSWORD]" http://api.wordnik.com/v4/word.json/button/pronunciations?sourceDictionary=macmillan&typeFormat=IPA&useCanonical=false)
raw=$(expr "$string" : '.*"raw":"\([^"]*\)"')

echo $raw

See man expr:

   STRING : REGEXP
          anchored pattern match of REGEXP in STRING

   Pattern matches return the string matched between \(  and  \)  or null