much faster indexer (but will require maintenance via cron)

[shellwiki] / parsers / 40_indexer.sh

diff --git a/parsers/40_indexer.sh b/parsers/40_indexer.sh
index ec7be4a7f4a6d9fa17da3b19ed762c814ce9b78b..7190b0b84f02c6a62f3a91b83f8887b767edac2f 100755 (executable)
--- a/parsers/40_indexer.sh
+++ b/parsers/40_indexer.sh
@@ -34,29 +34,23 @@ mkdir -p "$I"
   printf '%s\n' "$line"
 done \
 | awk '
-  BEGIN { FS = "([] \t\n\r!\"#'\''()*+,./:;<=>?\\^_`{|}~[-]|%[0-9A-Fa-f]{2})+" }
+  BEGIN { # Field separator FS should include punctuation, including Unicode Block U+2000 - U+206F
+          if ( length("¡") == 1 )  # Utf-8 aware AWK
+          FS = "([] \\t\\n\\r!\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '[\342\200\200-\342\201\257]')"')+";
+          else                     # UTF-8 Hack
+          FS = "([] \\t\\n\\r!\"#'\''()*+,./:;<=>?\\\\^_`{|}~[-]|%[0-9A-Fa-f]{2}|'"$(printf '\342\200[\200-\277]|\342\201[\201-\257]')"')+";
+          fi
+        }
         { for (n = 1; n <= NF; n++) {
             if ( $n != "" && length($n) <= 128 ) {
               words[tolower($n)]++; total++;
         } } }
     END { for (w in words) printf "%i %i %f %s\n", words[w], total, words[w] / total, w; }
 ' \
-| while read num total freq word; do
+| while read -r num total freq word; do
   [ "$word" ] || continue
-  F="$I/$word"
-  L="$(STRING "$DOC")"
-
-  if LOCK "$F"; then
-    touch "$F"
-    { while read d l f n t; do
-      [ "$l" = "$L" ] \
-      || printf "%i    %s      %f      %i      %i\n" \
-                "$d" "$l" "$f" "$n" "$t"
-      done <"$F"
-      printf "%i       %s      %f      %i      %i\n" \
-             "$_DATE" "$L" "$freq" "$num" "$total"
-    } >"$F.$$"
-    mv -- "$F.$$" "$F"
-    RELEASE "$F"
-  fi
+  printf "%i   %s      %f      %i      %i\n" \
+         "$_DATE" "$(STRING "$DOC")" \
+         "$freq" "$num" "$total" \
+  >>"$I/$word"
 done