Text completion: Difference between revisions

Content added Content deleted

Inline

@@ Line 446: / Line 446: @@
 Process finished with exit code 0
+</pre>
+=={{header|jq}}==
+{{works with|jq}}
+This entry uses unixdict.txt, the dictionary used frequently on other RosettaCode pages.
+Since the title of this particular page is "Text completion", the
+solution offered here includes both a straightforward lookup of the dictionary to
+check for possible completions, and a check for equally good matches
+based on the Levenshtein distance.
+The Levenshtein module used is the collection of function definitions on
+the RC page https://rosettacode.org/wiki/Levenshtein_distance
+and is thus not repeated here.
+The "percentage" reported is computed as described in the comments.
+The "debug" statements for showing the number of words under consideration for best Levenshtein matches is retained.
+<lang jq>
+include "levenshtein-distance" {search: "."};  # https://rosettacode.org/wiki/Levenshtein_distance#jq
+# input: a dictionary
+# output an array of {word, d, p} objects showing the best matches as measured
+# by the Levenshtein distance, d; p is an indicator of similarity expressed as
+# 100 * (max(0, ($word|length) - d) / ($word|length) ) rounded to the nearest integer
+def closest($word):
+  if .[$word] then {$word, d:0, percentage: 100}
+  else
+  "Levenshtein-closest words to \($word):",
+  (($word|length) as $length
+   | (keys_unsorted | map(select(length | (. > $length-3) and (. < $length + 3)))) as $candidates
+   | $candidates
+   | (length|debug) as $debug
+   | map( {word: ., d: levenshteinDistance($word; .)} )
+   | minima(.d)
+   | map( .p = (100 * ([0, ($length - .d) ] | max / $length) | round )))
+  end ;
+# Input: a dictionary
+# Output: an array of possible completions of $word
+def completion($word):
+  "Possible completions of \($word):",
+    (keys_unsorted | map(select(startswith(word))));
+def task:
+  INDEX(inputs; .)  # the dictionary
+  | completion("compli"),
+    closest("complition"),
+    closest("compxxxxtion")
+;
+task
+</lang>
+{{out}}
+Invocation: jq -nR -f text-completion.jq unixdict.txt
+This assumed levenshtein-distance.jq is in the pwd.
+See the comments above.
+<pre>
+Possible completions of compli:
+[
+  "compliant",
+  "complicate",
+  "complicity",
+  "compliment",
+  "complimentary",
+  "compline"
+]
+Levenshtein-closest words to complition:
+["DEBUG:",10396]
+[
+  {
+    "word": "completion",
+    "d": 1,
+    "p": 90
+  }
+]
+"Levenshtein-closest words to compxxxxtion:"
+["DEBUG:",4082]
+[
+  {
+    "word": "competition",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "compilation",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "completion",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "complexion",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "composition",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "compunction",
+    "d": 4,
+    "p": 67
+  },
+  {
+    "word": "computation",
+    "d": 4,
+    "p": 67
+  }
+]
 </pre>