Skip to content

Instantly share code, notes, and snippets.

@WaxCylinderRevival

WaxCylinderRevival/analyze-text-for-date-patterns.xq

Last active October 30, 2017 04:53
Show Gist options
  • Save WaxCylinderRevival/677233e2baae9a5e414d57b8cdc2ed08 to your computer and use it in GitHub Desktop.
Save WaxCylinderRevival/677233e2baae9a5e414d57b8cdc2ed08 to your computer and use it in GitHub Desktop.
Download ZIP
analyze-text-for-date-patterns.xq
Raw
analyze-text-for-date-patterns.xq
(: declare namespace dp='https://history.state.gov/ns/xquery/date-processing' :)
declare variable $local:regexes :=
map {
"month-regex" : "(?:January|February|March|April|May|June|July|August|September|October|November|December)",
"month-regex-fr" : "(?:janvier|février|fevrier|mart|avril|mai|juin|juillet|août|aout|septembre|octobre|novembre|décembre|decembre)",
"month-regex-sp" : "(?:enero|febrero|marzo|abril|mayo|junio|julio|agosto|septiembre|setiembre|octubre|noviembre|diciembre)",
"day-regex" : "(?:\d{1,2})(?:st|d|nd|rd|th)?",
"day-range-regex" : "(?:\d{1,2})(?:st|d|nd|rd|th)?\s*[-–—]\s*(?:\d{1,2})(?:st|d|nd|rd|th)?",
"year-regex" : "(?:\d{4})",
"day-spelled-out-regex" : "(?:(?:\d{1,2})(?:st|d|nd|rd|th)?|(?:thirtieth|thirty|twentieth|twenty)?(?:-|–|\s+)?(?:nineteenth|eighteenth|seventeenth|sixteenth|fifteenth|fourteenth|thirteenth|twelfth|eleventh|tenth|ninth|eighth|seventh|sixth|fifth|fourth|third|second|first)?)",
"year-spelled-out-regex" : "one\s+thousand\s+(?:nine|eight)\s+hundred\s+(?:ninety|eighty|seventy|sixty|fifty|forty|thirty|twenty)?(?:-\s+)?(?:nineteen|eighteen|seventeen|sixteen|fifteen|fourteen|thirteen|twelve|eleven|ten|nine|eight|seven|six|five|four|three|two|one)?"
}
;
declare function local:get-dates($source) {
let $day-month-year-regex := $local:regexes?day-regex || '\s+' || $local:regexes?month-regex || ',?\s+' || $local:regexes?year-regex
let $month-day-year-regex := $local:regexes?month-regex || '\s+' || $local:regexes?day-regex || ',?\s+' || $local:regexes?year-regex
let $month-day-range-year-regex := $local:regexes?month-regex || '\s+' || $local:regexes?day-range-regex || ',?\s+' || $local:regexes?year-regex
let $day-month-year-official-regex := $local:regexes?day-spelled-out-regex || '\s+day\s+of\s+' || $local:regexes?month-regex || ',\s+in\s+the\s+year\s+of\s+(?:our|the)\s+lord\s+' || $local:regexes?year-spelled-out-regex
let $day-month-year-regex-fr := $local:regexes?day-regex || '\s+' || $local:regexes?month-regex-fr || ',?\s+' || $local:regexes?year-regex
let $day-month-year-regex-sp := $local:regexes?day-regex || '\s+(?:de\s+)?' || $local:regexes?month-regex-sp || ',?\s+(?:(?:de|del)\s+)?' || $local:regexes?year-regex
return
(
analyze-string($source, $month-day-year-regex, "i"),
analyze-string($source, $month-day-range-year-regex, "i"),
analyze-string($source, $day-month-year-regex, "i"),
analyze-string($source, $day-month-year-official-regex, 'i'),
analyze-string($source, $day-month-year-regex-fr, "i"),
analyze-string($source, $day-month-year-regex-sp, "i")
)
};
@WaxCylinderRevival
Copy link
Author

WaxCylinderRevival commented Oct 26, 2017 •
edited

Input:

let $input := <head>The first date is <hi>February 2d, 1865</hi>. The next date is march 1 2010.  The <strong>third</strong> date is <hi rend="italic">31st July</hi> 2015.<note>This is a note with a date: September 1-3rd, 1929.</note> An example of an official date is the 2d day of May, in the year of the lord one thousand nine hundred nineteen. Another example is the fifth day of June, in the year of our Lord one thousand eight hundred sixty-five. This is a French date with whitespace: le 6 août
                            1902. This is a Spanish date: 6 de Agosto de 1902. This is another: 8 de septiembre del 2010.</head>

let $dates := local:get-dates($input)/fn:match

return <results>{$dates}</results>

@WaxCylinderRevival
Copy link
Author

WaxCylinderRevival commented Oct 26, 2017 •
edited

Results:

<results>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">February 2d, 1865</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">march 1 2010</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">September 1-3rd, 1929</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">31st July 2015</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">2d day of May, in the year of the lord one thousand nine hundred nineteen</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions"> fifth day of June, in the year of our Lord one thousand eight hundred sixty</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">6 août
                            1902</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">6 de Agosto de 1902</fn:match>
  <fn:match xmlns:fn="http://www.w3.org/2005/xpath-functions">8 de septiembre del 2010</fn:match>
</results>

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment