From: Paul Hänsch <paul@plutz.net>
Date: Tue, 21 Dec 2021 22:06:41 +0000 (+0100)
Subject: improved auto grouping
X-Git-Url: http://git.plutz.net/?p=serve0;a=commitdiff_plain;h=dad9cd471358d165f1de2ce8d5806703ac85f5bb

improved auto grouping
---

diff --git a/db_meta.sh b/db_meta.sh
index 9f427fc..d2d6999 100755
--- a/db_meta.sh
+++ b/db_meta.sh
@@ -28,15 +28,15 @@ read_meta() {
 
   # Global exports
   META_LENGTH='' META_WIDTH='' META_HEIGHT='' META_TAGS=''
-  META_COMMENT='' META_NAME='' META_GROUP=''
+  META_COMMENT='' META_NAME='' META_GROUP='' META_GROUPORDER=''
 
   if [ $# -eq 0 ]; then
     read -r META_LENGTH META_WIDTH META_HEIGHT META_TAGS META_COMMENT \
-            META_NAME META_GROUP
+            META_NAME META_GROUP META_GROUPORDER
   elif [ "$name" -a -f "$meta_db" -a -r "$meta_db" ]; then
     read -r META_LENGTH META_WIDTH META_HEIGHT META_TAGS META_COMMENT \
-            META_NAME META_GROUP <<-EOF
-	$(grep -F "	${name}${CR}" "$meta_db")
+            META_NAME META_GROUP META_GROUPORDER <<-EOF
+	$(grep -F "	${name}${CR}" "$meta_db" |dbmeta_autogroup)
 	EOF
   fi
   if [ "$META_NAME" ]; then
@@ -155,10 +155,38 @@ list_meta(){
     pfx="$(STRING "${pfx#/}")"
     [ "$pfx" = '\' ] && pfx='' || pfx="${pfx}/"
 
-    while read -r LENGTH WIDTH HEIGHT TAGS COMMENT NAME GROUP; do
+    dbmeta_autogroup "$meta" \
+    | while read -r LENGTH WIDTH HEIGHT TAGS COMMENT NAME GROUP; do
       printf '%i	%i	%i	tags=%s	comment=%s	%s%s\r	%s\n' \
              $LENGTH $WIDTH $HEIGHT "${TAGS#tags=}" \
              "${COMMENT#comment=}" "${pfx}" "${NAME%${CR}}" "${GROUP:-\\}"
-    done <"$meta"
+    done
   done
 }
+
+dbmeta_autogroup(){
+  sed -E '
+    # perform auto grouping if group id is missing or empty
+    /\r$/bAUTOGROUP; s;\r	\\$;\r;; tAUTOGROUP;
+  
+    # skip auto grouping if manual group id is present
+    /\r	.+$/b;
+  
+    :AUTOGROUP
+    # save original dataset
+    h;
+  
+    # strip all fields but the name
+    s;^[^\t]+	[^\t]+	[^\t]+	[^\t]+	[^\t]+	;;
+    # strip common suffixes of web video sites
+    s;-([0-9a-zA-Z_-]{11}|ph[0-9a-f]{13}|[0-9]{6,})\r$;-\r;;
+  
+    # replace all numeric parts and append numerals to an ordering field
+    # the group id will be made up of only the non-numeric character frame
+    # the ordering field will hold all numbers from the name
+    :X s;^([^\r0-9]*)([0-9]+)([^\r0-9]*)\r\t?([0-9]*)$;\1\r\3\r	\4\2;; tX;
+  
+    # append group id and ordering field to dataset
+    H; g; s;\n;\t;;
+  ' "$@"
+}
diff --git a/list.sh b/list.sh
index 65d929a..6245261 100755
--- a/list.sh
+++ b/list.sh
@@ -87,24 +87,14 @@ list_filter(){
 }
 
 list_order(){
-  local fm fn fn al length ln h w t c name buffer l
+  local fm fn fn al length ln h w t c name group o buffer l
 
   if [ $ORDER = Name ]; then
     sort -k6
   elif [ $ORDER = Group ]; then
-    sed -E '
-      /^[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\r]*\r$/bX;
-      /^[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\r]*\r\t\\$/bX;
-      s;^([^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t)([^\r]*)\r\t(.*)$;\1\r\3 \2\r\t;;
-      t;
-      :X
-      s;^([^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^ 0-9]*)-([0-9a-zA-Z_-]{11}\r|ph[0-9a-f]{13}\r|[0-9]{8}\r)(.*)$;\1-\r\3 \2;;
-      t;
-      s;^([^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^ 0-9]*)([0-9]+)(.*)$;\1\r\3 \2;;
-      tX;' \
-    | { sort -n -k7 -k8 -k9 -k10 -k11 |sort -s -k6,6 ; echo '0 0 0 tags= comment= _'; } \
-    | while read -r length w h t c name; do
-      if [ "${ln%% *}" = "${name%% *}" ]; then
+    { sort -n -k8 -k6,6 |sort -s -k7,7 ; echo '0 0 0 tags= comment= _'; } \
+    | while read -r length w h t c name group o; do
+      if [ "${ln%% *}" = "${group}" ]; then
         al=$((al + length))
         buffer="${buffer}${BR}$length	$w	$h	$t	$c	$name"
       else
@@ -114,13 +104,9 @@ list_order(){
         al="$length"
         buffer="$length	$w	$h	$t	$c	$name"
       fi
-      ln="$name"
+      ln="$group"
     done \
-    | sort -s -n -k1,1 |sed -E 's;^[0-9]+\t;;;' \
-    | sed -E '
-      s;^([^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t)\r([0-9A-Za-z:=]{16}) ([^ ]+)$;\1\3\r\t\2;; t;
-      :X s;^([^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\t]+\t[^\r]*)\r([^ ]*) ([^ ]+)( .*)?$;\1\3\2\4;; tX
-    '
+    | sort -s -n -k1,1 |sed -E 's;^[0-9]+\t;;;'
   elif [ $ORDER = Length ]; then
     sort -sn -k1
   elif [ $ORDER = Date ]; then