From 81f757e35f28225c0c6b86397c3177c39aaa394c Mon Sep 17 00:00:00 2001
From: kanoi <gitkanoi@k1k2.com>
Date: Sun, 10 Apr 2016 15:42:00 +1000
Subject: [PATCH] sql - a luck report

---
 sql/luck.sh | 318 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 318 insertions(+)
 create mode 100755 sql/luck.sh

diff --git a/sql/luck.sh b/sql/luck.sh
new file mode 100755
index 00000000..b4b2f609
--- /dev/null
+++ b/sql/luck.sh
@@ -0,0 +1,318 @@
+#!/bin/bash
+#
+tmp="/tmp/`basename $0`-$$"
+tmp0="$tmp.0.sql"
+tmp1="$tmp.1.sql"
+tmp2="$tmp.2.sql"
+tmp3="$tmp.3.sql"
+tmp4="$tmp.4.sql"
+msdd="1,000,000,000"
+msd="`echo "$msdd" | tr -d ,`"
+#
+usAge()
+{
+ echo "usAge: `basename $0` [-s sdiffX] [w] 0|[b startblock]|[s shift] endblock|z"
+ echo " run as root or change dosql1()"
+ echo
+ echo " -s sdiffX = all shares must have minsdiff <= sdiffX - if not then abort"
+ echo "    any commas are removed with \"tr -d ,\""
+ echo "    shares with sdiff below sdiffX are ignored. Default $msdd"
+ echo " w = worker stats - without w = user stats"
+ echo " 0 = from the first hi share"
+ echo " b startblock = start after the workinfoid of startblock (instead of 0)"
+ echo "                i.e. this only includes data AFTER 'startblock'"
+ echo "                'startblock' is the lowest block >= 'startblock'"
+ echo " s shift = start from the shift with the starting 5 char code 'shift'"
+ echo "           if there ever is a duplicate, it will use the lowest one"
+ echo " endblock = finish at the workinfoid of endblock (instead of z)"
+ echo "            this includes all of 'endblock'"
+ echo "            'endblock' is the highest block <= 'endblock'"
+ echo " z = to the last high share"
+ echo
+ echo " Basically would be '0 z' for all available share history or"
+ echo "  e.g. 'b NNN MMM' for a block range after NNN up to (including) MMM or"
+ echo "  e.g. 'b NNN z' starting after block NNN up to now"
+ echo " Put a 'w' in front, to get worker instead of user summarisation"
+ echo
+ echo " The share history determines the limits used"
+ echo " i.e. markersummaries before the first share, or after the last share,"
+ echo "  are not included since that would invalidate the results"
+ echo " Missing shares within the share data would invalidate the results"
+ exit 1
+}
+#
+dosql1()
+{
+ su - postgres << EOF
+echo "$sql" | psql ckdb
+EOF
+}
+#
+procsql()
+{
+ got=""
+ fin=""
+ prev=""
+ while true ; do
+	read line
+	if [ "$?" != "0" ] ; then
+		break
+	fi
+	if [ "$got" ] ; then
+		if [ -z "$fin" ] ; then
+			echo "$line"
+			if [ "${line:0:1}" = "(" ] ;then
+				fin="y"
+			fi
+		fi
+	else
+		if [ "${line:0:5}" = "-----" ] ; then
+			got="y"
+			echo "$prev"
+			echo "$line"
+		else
+			prev="$line"
+		fi
+	fi
+ done
+}
+#
+dosql()
+{
+ dosql1 2>&1 | procsql
+}
+#
+# reduce a large, 9 or more digit, comma number with 2 decimal places to nnG
+toG()
+{
+ sed -e "s/,\([0-9][0-9]\)[0-9],[0-9][0-9][0-9],[0-9][0-9][0-9]\.[0-9][0-9]*/.\1G/g" -e "s/\(\.[0-9][0-9]G\)   *|/\1 |/g" -e "s/  *$//"
+}
+#
+dedup()
+{
+ cut -d '|' -f 1-5,7-10,12
+}
+#
+dedupw()
+{
+ cut -d '|' -f 1-4,6-9,11
+}
+#
+p1="$((0x1d))"
+p1v="`echo "$p1 3 - 8 * p" | dc`"
+#
+# calculate the diff ratio for each diffacc, given bits
+diff1()
+{
+ read line
+ read line
+ while true ; do
+	read line
+	if [ "$?" != "0" ] ; then
+		break
+	fi
+	if [ "${line:0:1}" = "(" ] ; then
+		break
+	fi
+	bits="`echo "$line" | cut -d '|' -f2 | tr -d ' '`"
+	diffacc="`echo "$line" | cut -d '|' -f3 | tr -d ' '`"
+	po="$((0x${bits:0:2}))"
+	bi="$((0x${bits:2:6}))"
+	pd="`echo "$p1v $po 3 - 8 * - p" | dc`"
+	if [ "$pd" -lt "8" ] ; then
+		pd="8"
+	fi
+	per="`echo "8 k $diffacc 2 $pd ^ 65535 * $bi / / p" | dc`"
+	echo "$line | ${per}"
+ done
+}
+#
+diffvals()
+{
+ grep "^$uorw  *|" "$tmp4" | cut -d '|' -f4 | tr "\n" "+" | sed -e "s/+/ +/g"
+}
+#
+getdiff()
+{
+ echo "4 k 0 `diffvals` p" | dc
+}
+#
+# blocks are appended 3rd since if blocks exist, then at least 1 share must
+proc()
+{
+ read line
+ read line
+ while true ; do
+	read line
+	if [ "$?" != "0" ] ; then
+		break
+	fi
+	if [ "${line:0:1}" = "(" ] ; then
+		echo "$line"
+		break
+	fi
+	uorw="`echo "$line" | cut -d'|' -f1 | tr -d ' '`"
+	diffs="`getdiff` BDR"
+	if [ "${diffs:0:1}" = "." ] ; then
+		diffs="  0$diffs"
+	elif [ "${diffs:1:1}" = "." ] ; then
+		diffs="  $diffs"
+	elif [ "${diffs:2:1}" = "." ] ; then
+		diffs=" $diffs"
+	fi
+	stats="`grep "^$uorw  *|" "$tmp2"`"
+	bstats="`grep "^$uorw  *|" "$tmp3"`"
+	if [ "$stats" ] ; then
+		s=" | "
+	else
+		s=""
+	fi
+	if [ "$bstats" ] ; then
+		b=" | "
+	else
+		b=""
+	fi
+	echo "$line | $diffs$s$stats$b$bstats"
+ done
+}
+#
+procblk()
+{
+ while true ; do
+	read line
+	if [ "$?" != "0" ] ; then
+		break
+	fi
+	if [ "${line:0:1}" = "(" ] ; then
+		echo "$line"
+		break
+	fi
+	uorw="`echo "$line" | cut -d'|' -f1 | tr -d ' '`"
+	bstats="`grep "^$uorw  *|" "$tmp3"`"
+	echo "$line | $bstats"
+ done
+}
+#
+if [ "$1" = "-s" ] ; then
+ shift
+ msd="`echo "$1" | tr -d ,`"
+ shift
+fi
+#
+if [ -z "$1" ] ; then
+ usAge
+fi
+#
+if [ "$1" = "-?" -o "$1" = "-h" -o "$1" = "-help" -o "$1" = "--help" ] ; then
+ usAge
+fi
+#
+work=""
+if [ "$1" = "w" ] ; then
+ work="w"
+ shift
+fi
+#
+valsh="sdiff>=$msd"
+# Orphans included, Rejects not - since they can be below diff, or invalid, or too late or ...
+valblk="expirydate>'5555-05-05' and confirmed!='R'"
+#
+# START
+if [ "$1" = "0" ] ; then
+ shift
+ selwm1="ms.markerid>=(select min(markerid) from workmarkers where workinfoidstart>=(select min(workinfoid) from shares))"
+ selsh1=""
+ selblk1="and b.workinfoid>=(select min(workinfoid) from shares)"
+else
+ if [ "$1" = "s" ] ; then
+	shift
+	selwm1="ms.markerid>=(select min(markerid) from workmarkers where description like 'Shift fin: $1 %')"
+	selsh1="and workinfoid>=(select min(workinfoidstart) from workmarkers where description like 'Shift fin: $1 %')"
+	selblk1="$selsh1"
+	shift
+ else
+	if [ "$1" = "b" ] ; then
+		shift
+		selwm1="ms.markerid>=(select min(markerid) from workmarkers where workinfoidstart>(select min(workinfoid) from blocks where height>=$1))"
+		selsh1="and s.workinfoid>=(select min(workinfoid) from blocks where height>=$1)"
+		selblk1="and b.height>$1"
+		shift
+	else
+		echo "ERR: Unknown first option '$1'"
+		usAge
+	fi
+ fi
+fi
+#
+# END
+if [ "$1" = "z" ] ; then
+ selwm2=""
+ selsh2=""
+ selblk2=""
+else
+ if [ -z "$1" ] ; then
+	echo "ERR: missing ending block|z"
+	usAge
+ else
+	selwm2="and wm.markerid<=(select max(markerid) from workmarkers where workinfoidend<=(select max(workinfoid) from blocks where height<=$1))"
+	selsh2="and s.workinfoid<=(select max(workinfoid) from blocks where height<=$1)"
+	selblk2="and b.height<=$1"
+ fi
+fi
+#
+# check all shares have minsdiff > $msd
+sql="select count(*) from shares s where s.minsdiff>$msd $selsh1 $selsh2 ;"
+#
+dosql > "$tmp0"
+count="`head -n 3 "$tmp0" | tail -n 1`"
+if [ "$count" != "0" ] ; then
+ count
+ echo "ERROR: there were '$count' shares with minsdiff>$msd - that must be '0'"
+ usAge
+fi
+#
+# summarise markersummary into tmp1 for any user or worker with 'msd' diffacc or more
+if [ -z "$work" ] ; then
+ sql="select ms.userid,substring(max(ms.workername) from '(^[^\._]*)'),to_char(sum(ms.diffacc),'999G999G999G999G999.99'),to_char(sum(ms.sharecount),'999G999G999G999G999') from markersummary ms where $selwm1 $selwm2 group by ms.userid having sum(ms.diffacc)>$msd order by sum(ms.diffacc) desc;"
+else
+ sql="select ms.workername,to_char(sum(ms.diffacc),'999G999G999G999G999.99'),to_char(sum(ms.sharecount),'999G999G999G999G999.99') from markersummary ms where $selwm1 $selwm2 group by ms.workername having sum(ms.diffacc)>$msd order by sum(ms.diffacc) desc;"
+fi
+#
+dosql > "$tmp1"
+#
+# summarise the shares into tmp2
+if [ -z "$work" ] ; then
+ sql="select s.userid,count(*),to_char(min(s.sdiff),'999G999G999G999G999.99'),to_char(max(s.sdiff),'999G999G999G999G999.99'),to_char(avg(s.sdiff),'999G999G999G999G999.99') from shares s where $valsh $selsh1 $selsh2 group by s.userid;"
+else
+ sql="select s.workername,count(*),to_char(min(s.sdiff),'999G999G999G999G999.99'),to_char(max(s.sdiff),'999G999G999G999G999.99'),to_char(avg(s.sdiff),'999G999G999G999G999.99') from shares s where $valsh $selsh1 $selsh2 group by s.workername;"
+fi
+#
+dosql > "$tmp2"
+#
+# summarise the blocks into tmp3
+if [ -z "$work" ] ; then
+ sql="select b.userid,count(*)||' B' from blocks b where $valblk $selblk1 $selblk2 group by b.userid;"
+else
+ sql="select b.workername,count(*)||' B' from blocks b where $valblk $selblk1 $selblk2 group by b.workername;"
+fi
+#
+dosql > "$tmp3"
+#
+# summarise diff% into tmp4 (all ids will have a value)
+if [ -z "$work" ] ; then
+ sql0="select ms.userid,wi.bits,sum(ms.diffacc)"
+ sqlz="group by ms.userid,wi.bits;"
+else
+ sql0="select ms.workername,wi.bits,sum(ms.diffacc)"
+ sqlz="group by ms.userid,ms.workername,wi.bits;"
+fi
+#
+valwm="wm.status='p' and wm.expirydate>'5555-05-05'"
+sql="$sql0 from markersummary ms join workmarkers wm on ms.markerid=wm.markerid join workinfo wi on wm.workinfoidend=wi.workinfoid where $valwm and $selwm1 $selwm2 $sqlz"
+#
+dosql | diff1 > "$tmp4"
+#
+# join (append) to each tmp1 row, the matching row/data in tmp4, tmp2 and tmp3
+cat "$tmp1" | proc | toG | dedup$work
+#
+shred -uz "$tmp0" "$tmp1" "$tmp2" "$tmp3" "$tmp4" &> /dev/null