#!/bin/bash
## count number of Memberships in a Peggy db;  by Eugene Reimer  2006-Sep;
## her "Members-db" is really a mailing-list where one entry describes 2, 3, or more members living at one mailing-address;
## to get number of members, need to count as 2 any row having an "&" or the word "and" in FirstName column;
##
## checking for threesomes, foursomes, etc -- here are old examples:
##	eg:  FirstName:Lori & Ron/Helen			LastName:Drebnisky/ Vogt	<-- couple + mother-in-law  (weird use of slash??)
##	eg:  FirstName:Don, Barret, Lorressa, Sheena  	LastName:Batenchuk
##	eg:  FirstName:Terry, Jennifer & Vicki		LastName:Ewacha
##	eg:  FirstName:Terry, Richard, Michael		LastName:Zdan & Debra Lane	<-- NEED TO COUNT ampersands in LastName also--!!--
##	eg:  FirstName:Patty, Ed & 4 kids		LastName:Lotecki		<-- do not expect this style after Junior-mbrships...
## algorithm:
##	count the number of AMPERSAND|COMMA|SPACEandSPACE occurences, in FirstName+LastName,  and add that to NBRLINES;
##	-- AVOID COUNTING SLASH because of:  Froehlich/!Prairie!Originals  Orchid!Society!c/o!Robert!Kato
##	== COULD COUNT SLASH in FirstName ONLY--??--
##
## Not essential, but to also identify Group memberships:  look for "c/o" in LastName column??
##	(06nov: fails for Friends of Assiniboine Park Conservatory)
##	(BIG-db uses "Group" somewhere, in Hobbies?)
## USAGE:
##	countMBRS MBRS<date>.htm

I=$1
cat $I  |egrep -iv 'HTML>|>ID|<TABLE|TABLE>'  |
  sed 's| |!|g; s|<td| <td|g; s|<tr> |<tr>|g'  |
  arr 2 3  |													##just FirstName+LastName
  sed 's|!*&amp;!*|!\&!|g; s|&[a-z#][a-z0-9]*;|?|g; s|!and!|!\&!|g'  |						##stdize &amp; and "and" to ampersand
  sed 's|[&,]\+!*[&,]\+|\&|g'  |										##stdize sequence of [&,] to just &  (2011-01-10)
  sed 's|&!*Family||g'  |											##remove "& Family"
  cat  >tmpXXB													##into tmpXXB
quirks () { if eval $2 2>&1 >/dev/null;then echo "$1"; eval $2; echo ""; fi }					##function to suppress heading when zero examples

##irks "Lines with 2 or more member names:"		"grep '[&,]'       tmpXXB"				##--review the 2-or-more rows (normally yanked)
quirks "Lines with 3 or more member names:"		"grep '[&,].*[&,]' tmpXXB"				##--review the 3-or-more rows
quirks "Lines with digit or slash or hi-char in name:"	"grep '[0-9/?]'    tmpXXB"				##--review DIGIT(s) OR SLASH OR HI-CHAR in name
quirks "Lines with hi-char anywhere:"	"cat $I |sed 's|&amp;|\&|g; s|&nbsp;| |g' |grep '&[a-z#][a-z0-9]*;'"	##--review HI-CHAR anywhere

NBRLINES=$(    cat tmpXXB  |wc -l)
NBRTOTAL=$(    cat tmpXXB  |sed 's|[&,]|\n|g'  |wc -l)
##rm tmpXX?													##yank if tmpfiles wanted for DEBUG
echo "Nbr-of-Mailing-labels=NBRLINES=$NBRLINES"
echo "Nbr-of-Members(estimated)=NBR-LINES+AMPERSANDS+COMMAS+ANDs=$NBRTOTAL"


exit
2010-01-23:  added the quirks function for neater output;
	obsolete lines:
	echo "Lines with 3 or more member names:";             grep '[&,].*[&,]' tmpXXB
	echo "Lines with digit or slash or hi-char in name:";  grep '[0-9/?]'    tmpXXB
	echo "Lines with hi-char anywhere:";  cat $I |sed 's|&amp;|\&|g; s|&nbsp;| |g' |grep '&[a-z#][a-z0-9]*;'
2011-01-10:
	compared this Membership-count with Peggy's, hers being 161 at year-end, this one being 164;  one family was being miscounted and the correct total is 163;
	the troublesome one:  <td>Will,!Jose,!Aidan,!&!Emmett!  <td>Milne!	<--leads to 5; ought to be 4;
	fixed by revising sequence of [&,] to be just one;  now get 163 as wanted;