Text completion: Difference between revisions

Line 479:

=={{header|Phix}}==

uses levenshtein() from [[Levenshtein_distance#Phix]] and [http://wiki.puzzlers.org/pub/wordlists/unixdict.txt this unixdict.txt]

⚫

uses levenshtein() from [[Levenshtein_distance#Phix]] and [http://wiki.puzzlers.org/pub/wordlists/unixdict.txt unixdict] ~~(not the same as the one Julia uses - I did just check and indeed it does ''not'' have "collection" in it!)~~

<lang Phix>string word = "complition"

sequence words = get_text(join_path({"demo","unixdict.txt"}),GT_LF_STRIPPED)

sequence words = get_text(join_path({"demo","unixdict.txt"}),GT_LF_STRIPPED),

leven = apply(words,levenshtein,word)

function lt(string w, integer n) return ~~levenshtein(w,word)~~=n end function

function ln(string /*w*/, integer i, n) return leven[i]=n end function

for n=1 to 4 do

printf(1,"Words at Levenshtein distance of %d (%g%% similarity) from \"%s\": \n%s\n",

{n,100-round(100*n/length(word)),word,join_by(filter(words,lt,n),1,6)})

{n,100-round(100*n/length(word)),word,join_by(filter(words,ln,n),1,6)})

end for</lang>

Note the parameters of the filter routine lt() can be quite flexible, for instance you could instead do this (and get the same results)

<lang Phix>function lt(string w, sequence nw)

{integer n, string word} = nw

...

{n,100-round(100*n/length(word)),word,join_by(filter(words,lt,{n,word}),1,6)})</lang>

(matches Delphi/Go/Wren)

<pre>

Words at Levenshtein distance of 1 (90% similarity) from "complition":