langs a-z

Task/Inverted-index/OCaml/inverted-index.ocaml

@@ -0,0 +1,82 @@
+TYPE_CONV_PATH "Inverted_index"
+
+type files = string array with sexp
+type inverted_index = (string * int list) list with sexp
+
+type t = files * inverted_index with sexp
+
+open Sexplib
+
+let data_file = "data.inv"
+let data_path = Filename.concat Filename.temp_dir_name data_file
+
+let get_inv_index() =
+  if Sys.file_exists data_path
+  then t_of_sexp(Sexp.load_sexp data_path)
+  else ([| |], [])
+
+let load_file f =
+  let ic = open_in f in
+  let n = in_channel_length ic in
+  let s = String.create n in
+  really_input ic s 0 n;
+  close_in ic;
+  (s)
+
+let array_push ar v =
+  let len = Array.length ar in
+  Array.init (succ len) (fun i ->
+    if i < len then Array.unsafe_get ar i else v), len
+
+let uniq lst =
+  let h = Hashtbl.create (List.length lst) in
+  List.iter (fun x -> Hashtbl.replace h x ()) lst;
+  Hashtbl.fold (fun x () xs -> x :: xs) h []
+
+let combine words i inv_index =
+  let h = Hashtbl.create (List.length inv_index) in
+  List.iter (fun (w, from) -> Hashtbl.replace h w from) inv_index;
+  List.iter (fun w ->
+    if Hashtbl.mem h w
+    then begin
+      let from = Hashtbl.find h w in
+      Hashtbl.replace h w (i::from)
+    end else
+      Hashtbl.add h w [i]
+  ) words;
+  Hashtbl.fold (fun w from acc -> (w, from) :: acc) h []
+
+let words_of_file in_file =
+  let str = load_file in_file in
+  let words = Str.split (Str.regexp "[ \r\n\t,;.?!:'/\034()]") str in
+  let words = uniq words in
+  (words)
+
+let index_file in_file =
+  let words = words_of_file in_file in
+  let files, inv_index = get_inv_index() in
+  let files, i = array_push files in_file in
+  let inv_index = combine words i inv_index in
+  let se = sexp_of_t (files, inv_index) in
+  Sexp.save data_path se
+
+let search_word word =
+  let files, inv_index = get_inv_index() in
+  try
+    let is_in = List.assoc word inv_index in
+    List.iter (fun i -> print_endline files.(i)) is_in
+  with Not_found ->
+    print_endline "# Not Found"
+
+let usage() =
+  Printf.printf "Usage: %s \
+    --index-file <file.txt> / \
+    --search-word <some-word>\n%!" Sys.argv.(0);
+  exit 1
+
+let () =
+  let cmd, arg = try (Sys.argv.(1), Sys.argv.(2)) with _ -> usage() in
+  match cmd, arg with
+  | "--index-file", in_file -> index_file in_file
+  | "--search-word", word -> search_word word
+  | _ -> usage()

Task/Inverted-index/Perl-6/inverted-index.pl6

@@ -0,0 +1,12 @@
+sub MAIN (*@files) {
+    (my %norm).push: do for @files -> $file {
+        $file X=> slurp($file).lc.words;
+    }
+    (my %inv).push: %norm.invert.uniq;
+
+    while prompt("Search terms: ").words -> @words {
+        for @words -> $word {
+            say "$word => %inv.{$word.lc}";
+        }
+    }
+}

Task/Inverted-index/TUSCRIPT/inverted-index.tuscript

@@ -0,0 +1,30 @@
+$$ MODE TUSCRIPT
+
+files="file1'file2'file3"
+LOOP file=files
+ERROR/STOP CREATE (file,seq-o,-std-)
+ENDLOOP
+
+content1="it is what it is"
+content2="what is it"
+content3="it is a banana"
+
+FILE/ERASE "file1" = content1
+FILE/ERASE "file2" = content2
+FILE/ERASE "file3" = content3
+
+ASK "search for": search=""
+IF (search=="") STOP
+
+BUILD R_TABLE/USER/AND search = *
+DATA  {search}
+
+LOOP/CLEAR file=files
+ ACCESS q: READ/RECORDS $file s.z/u,content,count
+  LOOP
+  COUNT/NEXT/EXIT q (-; search;-;-)
+  IF (count!=0) files=APPEND (files," ",file)
+  ENDLOOP
+ ENDACCESs q
+ENDLOOP
+PRINT "-> ",files

Task/Inverted-index/UNIX-Shell/inverted-index-1.sh

@@ -0,0 +1,25 @@
+#!/bin/ksh
+
+typeset -A INDEX
+
+function index {
+  typeset num=0
+  for file in "$@"; do
+    tr -s '[:punct:]' ' ' < "$file" | while read line; do
+      for token in $line; do
+        INDEX[$token][$num]=$file
+      done
+    done
+  ((++num))
+  done
+}
+
+function search {
+  for token in "$@"; do
+    for file in "${INDEX[$token][@]}"; do
+      echo "$file"
+    done
+  done | sort | uniq -c | while read count file; do
+    (( count == $# )) && echo $file
+  done
+}

Task/Inverted-index/UNIX-Shell/inverted-index-2.sh

@@ -0,0 +1,2 @@
+index *.txt
+search hello world

Task/Inverted-index/UNIX-Shell/inverted-index-3.sh

@@ -0,0 +1,49 @@
+#!/bin/sh
+# index.sh - create an inverted index
+
+unset IFS
+: ${INDEX:=index}
+
+# Prohibit '\n' in filenames (because '\n' is
+# the record separator for $INDEX/all.tab).
+for file in "$@"; do
+	# Use printf(1), not echo, because "$file" might start with
+	# a hyphen and become an option to echo.
+	test 0 -eq $(printf %s "$file" | wc -l) || {
+		printf '%s\n' "$file: newline in filename" >&2
+		exit 1
+	}
+done
+
+# Make a new directory for the index, or else
+# exit with the error message from mkdir(1).
+mkdir "$INDEX" || exit $?
+
+fi=1
+for file in "$@"; do
+	printf %s "Indexing $file." >&2
+
+	# all.tab maps $fi => $file
+	echo "$fi $file" >> "$INDEX/all.tab"
+
+	# Use punctuation ([:punct:]) and whitespace (IFS)
+	# to split tokens.
+	ti=1
+	tr -s '[:punct:]' ' ' < "$file" | while read line; do
+		for token in $line; do
+			# Index token by position ($fi, $ti). Ignore
+			# error from mkdir(1) if directory exists.
+			mkdir "$INDEX/$token" 2>/dev/null
+			echo $ti >> "$INDEX/$token/$fi"
+			: $((ti += 1))
+
+			# Show progress. Print a dot per 1000 tokens.
+			case "$ti" in
+			*000)	printf .
+			esac
+		done
+	done
+
+	echo >&2
+	: $((fi += 1))
+done

Task/Inverted-index/UNIX-Shell/inverted-index-4.sh

@@ -0,0 +1,33 @@
+#!/bin/sh
+# search.sh - search an inverted index
+
+unset IFS
+: ${INDEX:=index}
+
+want=sequence
+while getopts aos name; do
+	case "$name" in
+	a)	want=all;;
+	o)	want=one;;
+	s)	want=sequence;;
+	*)	exit 2;;
+	esac
+done
+shift $((OPTIND - 1))
+
+all() {
+	echo "TODO"
+	exit 2
+}
+
+one() {
+	echo "TODO"
+	exit 2
+}
+
+sequence() {
+	echo "TODO"
+	exit 2
+}
+
+$want "$@"