Tradelog

tradelog.sh

#!/bin/bash
export LC_ALL=C
export POSIXLY_CORRECT=yes

printUsage() {
	cat <<-EOF
)"	Usage: $(basename "$0") [-h|--help] [FILTER] [COMMAND] [LOG [LOG2 [...]]
	  COMMAND:
	    list-tick    list all tickers mentioned in the logs
	    profit       final product of all entries (negative for a net loss)
	    pos          currently held positions ordered asc. by value
	    last-price   last
	    hist-ord     histogram of transactions grouped by tickers
	    graph-pos    graph positions for each ticker
	    [ommited]    entries are echoed according to the provided filters (if any)

	  FILTER:
	    -a DATETIME  only records AFTER (exclusive) this date are considered
	    -b DATETIME  only records BEFORE (exclusive) this date are considered
	       format: YYYY-MM-DD HH:MM:SS   or in POSIX: %Y-%m-%d %H:%M:%S
	    -t TICKER    count records for this TICKER, can be repeated
	    -w WIDTH     sets the width (in columns) of the longest line for commands
	      hist-ord   - if ommited, one block = one transaction
	      graph-pos  - if ommited, one block = 1000 units of currency

	      Note: When printing blocks of the graph, division products are floored
	      (rounded towards zero). So for \`graph-pos -w 6\` with 1234 (equal
	      to 6 blocks) as the maximum value, 1233.99 will be shown as #####.

	  Format:
	  	DATUM A CAS;TICKER;TYP TRANSAKCE;JEDNOTKOVA CENA;MENA;OBJEM;ID

	  Exit codes:
	    0            Success.
	    1            Invalid date provided for the -a or -b filter.
	    2            Duplicated parameter.
	EOF
	exit 1
}

ARG_DATE_AFTER="0000-01-01 00:00:00"
ARG_DATE_BEFORE="9999-12-31 23:59:59"
ARG_TICKER=""
ARG_WIDTH="0"
ARG_COMMAND="echo"
# needs spaces around all quotes
ACCEPTED_COMMANDS=" list-tick profit pos last-price hist-ord graph-pos echo "

# validateDateFormat (date)
# Returns 0 if the provided string follows YYYY-MM-DD HH:MM:SS.
# Note: This is a veeeeeeeeeeery basic check, allows e. g. "2020-12-24 15"
function validateDateFormat {
	# accepts 60 for leap seconds
	awkOut="$(echo "$1" | awk -v FS='[-: ]' -v RS='\0' ' { 
		print (match($0, /^[[:digit:]]{4,}-[[:digit:]]{2,}-[[:digit:]]{2,} [[:digit:]]{2,}:[[:digit:]]{2,}:[[:digit:]]{2,}\n?$/)) ? "good" : "bad"
	}')"
	[ "$awkOut" = "good" ];
}

# processFile (file)
function processFile {
	catcmd="cat"
	if [ "${1: -3}" == ".gz" ]; then
		catcmd="zcat"
	fi

	aftercmd="cat"
	awkProg=""

	# it could be done in awk, but so far we managed without GNU extensions,
	# hence we'll bite our tongue and just do this
	case "$ARG_COMMAND" in
		echo )
			awkProg="$(cat <<'EOF'
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);
}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{

	print $0
	
}

END {
	
}
EOF
)"			;;
		list-tick )
			aftercmd="sort -dk1"
			awkProg="$(cat <<'EOF'
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);
}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{
	if ($2 in unique_tickers == 0) {
		unique_tickers[$2] = ""
	}
}

END {
	for (i in unique_tickers) {
		print i
	}
}
EOF
)"			;;
		profit )
			awkProg="$(cat <<'EOF'
function sumItUp(units, cpu) {
	# avoid floating point arithmetic AT ALL COST!
	partsLen = split(cpu, parts, ".")
	if (partsLen == 1) {
		cost = parts[1] * 100
		return cost * units
	}

	dec = parts[2]
	dec = substr(dec, 0, 2)
	for (i = length(dec); i < 2; i++) {
		dec = dec"0"
	}

	cost = parts[1] * 100 + dec
	return cost * units
}
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);

	profit = 0
}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{
	if ($3 == "buy") {
		profit -= sumItUp($6, $4)
	} else if ($3 == "sell") {
		profit += sumItUp($6, $4)
	}
}

END {
	# taking the lazy way here, not cool!
	profit /= 100
	printf "%.2f\n", profit
}
EOF
)"			;;
		pos )
			awkProg="$(cat <<'EOF'
function sumItUp(units, cpu) {
	# avoid floating point arithmetic AT ALL COST!
	partsLen = split(cpu, parts, ".")
	if (partsLen == 1) {
		cost = parts[1] * 100
		return cost * units
	}

	dec = parts[2]
	dec = substr(dec, 0, 2)
	for (i = length(dec); i < 2; i++) {
		dec = dec"0"
	}

	cost = parts[1] * 100 + dec
	return cost * units
}
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);

}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{
	if ($3 == "buy") {
		profit[$2] += $6
	} else if ($3 == "sell") {
		profit[$2] -= $6
	}

	lastPrice[$2] = $4
}

END {
	widestNum = 0
	for (tic in profit) {
		profit[tic] = sumItUp(profit[tic], lastPrice[tic])

		profit[tic] /= 100
		str = sprintf("%.2f", profit[tic])
		strLen = length(str)
		if (strLen > widestNum) {
			widestNum = strLen
		}
		
	}

	for (tic in profit) {
		printf "%-9s : %"widestNum".2f\n", tic, profit[tic]
	}
}
EOF
)"			aftercmd="sort -nk3 -r"
			;;
		last-price )
			awkProg="$(cat <<'EOF'
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);

}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{
	lastPrice[$2] = $4
}

END {
	widestNum = 0
	for (tic in lastPrice) {
		str = sprintf("%.2f", lastPrice[tic])
		strLen = length(str)
		if (strLen > widestNum) {
			widestNum = strLen
		}
	}

	for (tic in lastPrice) {
		printf "%-9s : %"widestNum".2f\n", tic, lastPrice[tic]
	}
}
EOF
)"			aftercmd="sort -dk1"
			;;
		hist-ord )
			awkProg="$(cat <<'EOF'
BEGIN {
	RS="\n"
	FS=";"

	split(ARG_TICKER, _tmp_tickers, ";")
	for (i in _tmp_tickers) {
		tickers[_tmp_tickers[i]] = ""
		hasTickers = 1
	}

	gsub(/[-: ]/, "", ARG_DATE_AFTER);
	gsub(/[-: ]/, "", ARG_DATE_BEFORE);

}

(hasTickers == 1) && ($2 in tickers == 0) { next }

{
	fieldDate=$1
	gsub(/[-: ]/, "", fieldDate);

	if (ARG_DATE_AFTER != "" && ARG_DATE_AFTER > fieldDate) { next }
	if (ARG_DATE_BEFORE != "" && ARG_DATE_BEFORE < fieldDate) { next }
}

{
	totalTrans[$2]++
}

END {

	mostTrans = 0

	for (tic in totalTrans) {
		if (mostTrans < totalTrans[tic]) {
			mostTrans = totalTrans[tic]
		}
	}

	if (ARG_WIDTH < 1) {
		ARG_WIDTH = 1
	}

	for (tic in totalTrans) {
		if (ARG_WIDTH == 1) {
			cols = totalTrans[tic]
		} else {
			cols = (totalTrans[tic] / mostTrans) * ARG_WIDTH
		}
		
		row = ""
		for (i = 1; i <= cols; i++) {
			row = row"#"
		}

		printf "%-9s : %s\n", tic, row
	}
}
EOF
)"			aftercmd="sort -dk1"
			;;
	esac

	# dis betr
	# -f "$ARG_COMMAND.awk" \
		

	"$catcmd" "$1" | awk \
		-v ARG_TICKER="$ARG_TICKER" \
		-v ARG_DATE_AFTER="$ARG_DATE_AFTER" \
		-v ARG_DATE_BEFORE="$ARG_DATE_BEFORE" \
		-v ARG_WIDTH="$ARG_WIDTH" \
		"$awkProg" \
		-  | $aftercmd

	# Note: Oh wouldn't it be great to use the @include keyword
	# from GNU awk and avoid all that duplicated code. Oh well.
}

# Transform long options into short ones
for arg in "$@"; do
  shift
  case "$arg" in
    "--help") set -- "$@" "-h" ;;
    *)        set -- "$@" "$arg"
  esac
done

_hasSpecifiedWidth=0

if [ ! $1 ]; then
	cat -
	exit 0
fi

while getopts "ha:b:t:w:" o; do
	case "$o" in
		a)
			ARG_DATE_AFTER="$OPTARG"
			if ! validateDateFormat "$OPTARG"; then
				echo "Invalid after date." 1>&2
				exit 1
			fi
			;;
		b)
			ARG_DATE_BEFORE="$OPTARG"
			if ! validateDateFormat "$OPTARG"; then
				echo "Invalid before date." 1>&2
				exit 1
			fi
			p="$OPTARG"
			;;
		t)
			ARG_TICKER="$ARG_TICKER;$OPTARG"
			# if [ ! "$OPTARG" ]; then echo "Missing argument for -t" ; exit 2 ; fi
			;;
		w)
			ARG_WIDTH="$OPTARG"
			if [ $_hasSpecifiedWidth = "1" ]; then
				echo "Width specified twice!" 1>&2
				exit 2
			fi
			_hasSpecifiedWidth=1
			;;
		h)
			printUsage
			;;
		*)
			;;
	esac
done
shift "$((OPTIND-1))"

# otherwise fails with "binary operator expected" ?!?!?!
_tmpArgs="$@"
if [ -z "$_tmpArgs" ]; then
	processFile -
fi

for arg in "$@"; do
	if [ -z "${ACCEPTED_COMMANDS##* $arg *}" ]; then
		ARG_COMMAND="$arg"
		shift
	fi
done

for arg in "$@"; do
	processFile "$arg"
done