#!/bin/bash
## subsetMBRStoSMALL -- used from generateElistFromHISTORY+MBRS -- Eugene Reimer 2007-Mar;
##
## note: Peggy sends 2 kinds, which i refer to as:  
##	BIG   -- all membership records since 2000-Jan     -- i give these filenames containing "BIG"
##	SMALL -- only the current In-Good-Standing members -- i give these filenames without BIG -- also without SMALL:-)
##	-- whereas her "small" contains only In-Good-Standing members,
##	   my SMALL constructed from a BIG, includes those delinquent by less than a year -- BUT ONLY early in the year (Jan..May);
##
## USAGE:
##	subsetMBRStoSMALL  BIG-htmfile  DEST-SMALL-htmfile	<-- only because generateElist script had both names handy:-)

BIG=$1;  DST=$2;  [ $# -ne 2 ] && { echo "usage subsetMBRStoSMALL  BIG-htmfile  DEST-SMALL-htmfile";  exit; }
##set -v

##YY=$(date +%y);  MM=$(date +%m)				##get current year YY, month MM, and PP=(YY-1)		(pre-07oct)
MDATE=${BIG/MBRS/}; MDATE=${MDATE/-BIG/}; MDATE=${MDATE%%.*}	##get MDATE from the input MBRS-db-name			(07oct)
YY=${MDATE:2:2};yy=10#$YY;  MM=${MDATE:4:2};mm=10#$MM		##get year YY and month MM from the input MBRS-db-name  (07oct)
((P=yy-1)); PP=$(printf "%02d" $P);pp=10#$PP			##get previous year PP = (YY-1)
echo -n "subsetMBRStoSMALL:  YY=$YY MM=$MM PP=$PP -- continue??";  read			##DEBUG  07oct

## COLUMNS: eg: to select those who paid dues for 2005 or 2006, test for columns 32 or 34 non-null;
##	Dues2004 is col 29
##	Dues2005 is col 32 -- since 2005, she's been adding exactly 2-columns per year (previously she added 3 per year)
##	Dues2006 is col 34
##	Dues2007 is col 36  (I hope it will be)
## DERIVE column-nbrs from the year:  Dues-YY is col 32+2*(YY-05)  <--EVEN BETTER WOULD BE to find the "Dues YYYY" heading in first row of table??
##
## ALGORITHM:  When MM < 06, then select based on Dues-YY-1 OR Dues-YY;  when MM>=06, use only Dues-YY;
##
((DuesYY= 32 + 2 * (yy-5) ))								##Dues-column-nbr for current-year YY
((DuesPP= 32 + 2 * (pp-5) ))								##Dues-column-nbr for previous-year PP

grep '<TABLE' $BIG |sed 's|.*<tr><td>||; s|<td>|~|g'  |tr '~' '\n'  >tmpColHeadings	##write ColHeadings temp-file, for checking...
hdgYY=$(head -n$DuesYY tmpColHeadings |tail -n1)					##Dues-column-contents for year YY from the ColHeadings
hdgPP=$(head -n$DuesPP tmpColHeadings |tail -n1)					##Dues-column-contents for year PP from the ColHeadings
if [[ $hdgYY != 20${YY}Dues* ]];then echo "subsetMBRStoSMALL:  column$DuesYY: EXPECTING:20${YY}DuesDate, FOUND:$hdgYY"; fi	##expect hdgYY to be 20${YY}DuesDate
if [[ $hdgPP != 20${PP}Dues* ]];then echo "subsetMBRStoSMALL:  column$DuesPP: EXPECTING:20${PP}DuesDate, FOUND:$hdgPP"; fi	##expect hdgPP to be 20${PP}DuesDate

if ((mm<6)); then									##--Month is Jan..May, select BadStanding as well as GoodStanding members
  cat $BIG						 			|
    sed 's| |~|g; s|<td| <td|g; s|<tr> |<tr>|g'			 		|	##field kludges
    awk '{ if(NF<34 || $'$DuesPP'!~"&nbsp;" || $'$DuesYY'!~"&nbsp;") print $0 }'|	##select if DuesPP or DuesYY NON-null
    sed 's|~| |g'	     		      		     	      	 	>$DST
else											##--Month is Jun..Dec, select GoodStanding members only
  cat $BIG						 			|
    sed 's| |~|g; s|<td| <td|g; s|<tr> |<tr>|g'			 		|	##field kludges
    awk '{ if(NF<34 || $'$DuesYY'!~"&nbsp;") print $0 }' 		 	|	##select if DuesYY NON-null
    sed 's|~| |g'	     		      				 	>$DST
fi
echo -n "subsetMBRStoSMALL:  Counting lines in Received-BIG:";     wc $BIG  		##DEBUG
echo -n "subsetMBRStoSMALL:  Counting lines in Subsetted-SMALL:";  wc $DST  		##DEBUG


exit
==might be nicer to call this from (or incorporate into) fixPeggyMBRS;  it can then do:
	(1) renaming to BIG if needed,  as well as
	(2) subsetting to SMALL by time-of-year rules (this scipt);

2007-Oct:
	time-of-year rules now based on DATE-of-Peggys-XLS database (NOT date of emailing);
	--may also revise the cutoff:  Early==Jan..APRIL  (rather than Jan..May)--??--

note:
	ONCE A YEAR, will send a special "renewal needed" reminder to BS-members, AND remove them from the Elist - say in April or May?
	=> early in the year, construct Elist from GS+BS mbrs;
	but after the "renewal needed" reminder, construct it from GS only;
	NOTE: in 06may, besides BadStanding cases to be removed, also had some long-Lapsed ex-members!  because I had started with 04feb list...
	did the 06may removal using output from classifyElistIDs, fed through grep+sed...
2007-Oct: have automated producing such ToBeRemoved list, in generateElistFromHISTORY+MBRS
2008-Feb: to get around the OCTAL problem, introduced yy,mm,Myy,Mmm vars having a leading "10#"  ==MAY NEED TESTING / DEBUGGING==??==