Skip to content

Utility functions

.nlp Utility functions detectLang Detect the language within a text findDates Find all the dates in a string findRegex Find regular expressions within a string findTimes Find all the times in a string getSentences Extract all sentences for a document loadTextFromDir Import all files in a director removeCustom Remove aspects of a string containing certain characters or expressions removeNonAscii Remove non-ASCII characters from a string removeReplace Replace individual characters in a string sentiment Calculate the sentiment of a sentence

The NLP library contains functions useful for in-depth document analysis. They extract elements of the text that can be applied to NLP algorithms, or that can help you with your analysis.

In the below examples, the parsedTab/parsedDict variable is the output from the .nlp.newParser example defined in the data-preprocessing section.


Language of a text

.nlp.detectLang text

Where text is a string, returns a symbol denoting its language.

q).nlp.detectLang "This is a string"
q).nlp.detectLang "Ein, zwei, drei, vier"

This function uses Python’s langdetect module.

Language codes


Find dates in a string

.nlp.findDates text

Where text is a string, potentially containing multiple dates, returns a general list:

  1. Start date of the range
  2. End date of the range
  3. Text of the range
  4. Start index of the date (long)
  5. Index after the end index (long)
q).nlp.findDates "I am going on holidays on the 12/04/2018 to New York and come back on the 18.04.2018"
2018.04.12 2018.04.12 "12/04/2018" 30 40
2018.04.18 2018.04.18 "18.04.2018" 74 84


Find regular expressions within a string



  • text is a string
  • expr is the expression type as a symbol to be searched for within the text

returns a dictionary, extracting the expression along with the indices for the expressions.

The expression types that can be sought within the text are:

`specialChars                 `year
`money                        `yearfull
`phoneNumber                  `am
`emailAddress                 `pm
`url                          `time12
`zipCode                      `time24
`postalCode                   `time
`postalOrZipCode              `yearmonthList
`dtsep (date separator)       `yearmonthdayList
`day                          `yearmonth
`month                        `yearmonthday
q)txt:"You can call the number 123 456 7890 or email us on in book an
   appoinment for January,February and March for £30.00"
phoneNumber  | ,(" 123 456 7890";23;36)
emailAddress | ,("";52;66)
yearmonthList| (("January";97;104);("February";105;113);("March";118;123);("30";129;131);("00";13..
money        | ,("\302\24330.00";128;134)


Find times in a string

.nlp.findTimes text

Where text is a string, returns a general list:

  1. Time
  2. Text of the time (string)
  3. Start index (long)
  4. Index after the end index (long)
q).nlp.findTimes "I went to work at 9:00am and had a coffee at 10:20"
09:00:00.000 "9:00am" 18 24
10:20:00.000 "10:20"  45 50


Extract sentences from a document

.nlp.getSentences parsedDict

Where parsedDict is a dictionary containing a single parsed text (as returned by .nlp.newParser) returns the sentences from the text as a list of strings.

// Finds the sentences in the first chapter of MobyDick
q).nlp.getSentences parsedDict
"CHAPTER 1\n\n  Loomings\n\n\n\nCall me Ishmael."
" Some years ago--never mind how long precisely-- having little or no money in my purse, and noth..
" It is a way I have of driving off the spleen and regulating the circulation."
"Whenever I find myself growing grim about the mouth; whenever it is a damp, drizzly November in ..
" This is my substitute for pistol and ball."
"With a philosophical flourish Cato throws himself upon his sword; I quietly take to the ship."


Import all files in a directory

.nlp.loadTextFromDir filepath

Where filepath is the directory’s filepath as a string, returns a table of filenames, paths and texts contained within the filepath.


fileName path                                           text                 ..
1.       :./datasets/maildir/skilling-j/_sent_mail/1.   "Message-ID: <1461010..
10.      :./datasets/maildir/skilling-j/_sent_mail/10.  "Message-ID: <1371054..
100.     :./datasets/maildir/skilling-j/_sent_mail/100. "Message-ID: <47397.1..
101.     :./datasets/maildir/skilling-j/_sent_mail/101. "Message-ID: <2486283..


Remove characters from a string



  • text is a string
  • char is a list of characters or expressions to be removed from the text

returns the string without defined characters or expressions.

q)rmvList   :("*\n*";"*?!*";"*,";"*&*";"*[0-9]*")

"Re:\n\nHow much to you have?!  SRS\n\n\n\n\nKevin Hannon @ ENRON COMMUNICATIONS on 04/20/2001 08..

"much to you  SRS\n\n\n\n\nKevin Hannon ENRON COMMUNICATIONS on  \n\n\nOK Sherri how much do you ..


Remove non-ASCII characters from a string


Where text is a string returns it with all non-ASCII characters removed.

q).nlp.removeNonAscii["This is ä senteñcê"]
"This is  sentec"


Remove and replace characters from a string



  • text is a string
  • char is a string of characters to be removed
  • replace is the characters or expressions which to replace the removed character/s

returns the string with the characters replaced.

"Re????????How much to you have????  SRS??????????Kevin Hannon ?? ENRON COMMUNICATIONS on 04??20?..


Sentiment of a sentence

.nlp.sentiment text

Where text is string, returns a dictionary containing the sentiment score divided between compound, positive, negative and neutral components.

A run of sentences from Moby Dick:

q).nlp.sentiment each ("Three cheers,men--all hearts alive!";"No,no! shame upon all cowards-shame upon them!")
compound   pos       neg       neu
0.7177249  0.5996797 0         0.4003203
-0.8802318 0         0.6910529 0.3089471
Back to top