JQ Capitalize first letter of each word-CodePudding

I have a large JSON file that I am using JQ to pair down to only those elements I need. I have that working but there are some values that are string in all caps. Unfortunately, while jq has ascii_downcase and ascii_upcase, it does not have a built in function for uppercasing only the first letter of each word.

I need to only perform this on brand_name and generic_name, while ensure that the manufacturer name is also first letter capitalized with the exception of things like LLC which should remain capitalized.

Here's my current jq statement:

jq '.results[] | select(.openfda.brand_name != null or .openfda.generic_name != null or .openfda.rxcui != null) | select(.openfda|has("rxcui")) | {brand_name: .openfda.brand_name[0], generic_name: .openfda.generic_name[0], manufacturer: .openfda.manufacturer_name[0], rxcui: .openfda.rxcui[0]}' filename.json > newfile.json

This is a sample output:

{
  "brand_name": "VELTIN",
  "generic_name": "CLINDAMYCIN PHOSPHATE AND TRETINOIN",
  "manufacturer": "Almirall, LLC",
  "rxcui": "882548"
}

I need the output to be:

{
  "brand_name": "Veltin",
  "generic_name": "Clindamycin Phosphate And Tretinoin",
  "manufacturer": "Almirall, LLC",
  "rxcui": "882548"
}

CodePudding user response：

Suppose we are given an array of words that are to be left as is, e.g.:

def exceptions: ["LLC", "USA"];

We can then define a capitalization function as follows:

# Capitalize all the words in the input string other than those specified by exceptions:
def capitalize:
  INDEX(exceptions[]; .) as $e
  | [splits("\\b") | select(length>0)]
  | map(if $e[.] then . else (.[:1]|ascii_upcase)   (.[1:] |ascii_downcase) end)
  | join("");

For example, given "abc-DEF ghi USA" as input, the result would be "Abc-Def Ghi USA".

CodePudding user response：

Split at space characters to get an array of words, then split again at the empty string to get an array of characters. For the inner array, use ascii_downcase on all elements but the first, then put all back together using add on the inner and join with a space character on the outer array.

(.brand_name, .generic_name) |= (
  (. / " ") | map(. / "" | .[1:] |= map(ascii_downcase) | add) | join(" ")
)

{
  "brand_name": "Veltin",
  "generic_name": "Clindamycin Phosphate And Tretinoin",
  "manufacturer": "Almirall, LLC",
  "rxcui": "882548"
}

Demo

To ignore certain words from being processed, capture them with an if condition:

map_values((. / " ") | map(
  if IN("LLC", "AND") then .
  else . / "" | .[1:] |= map(ascii_downcase) | add end
) | join(" "))

{
  "brand_name": "Veltin",
  "generic_name": "Clindamycin Phosphate AND Tretinoin",
  "manufacturer": "Almirall, LLC",
  "rxcui": "882548"
}

Demo