move daily extraction to a separate script

2 files changed, 132 insertions, 121 deletions

diff --git a/bin/db-zugbildung-to-json b/bin/db-zugbildung-to-json
index af22818..3b90050 100755
--- a/bin/db-zugbildung-to-json
+++ b/bin/db-zugbildung-to-json
@@ -8,10 +8,7 @@ use warnings;
 use 5.020;
 use utf8;
 
-use DateTime;
-use DateTime::Format::Strptime;
 use Encode qw(decode);
-use File::Slurp qw(write_file);
 use IPC::Run3;
 use JSON;
 use List::Util qw(uniq);
@@ -27,38 +24,6 @@ sub show_usage {
 	exit $exit_code;
 }
 
-my $strp = DateTime::Format::Strptime->new(
-	pattern   => '%F',
-	time_zone => 'Europe/Berlin'
-);
-my $now = DateTime->now( time_zone => 'Europe/Berlin' );
-
-sub range_is_today {
-	my ($range) = @_;
-
-	if ( $range =~ m{^(.*)/(.*)$} ) {
-		my $dt1 = $strp->parse_datetime($1);
-		my $dt2 = $strp->parse_datetime($2);
-		if ( $dt1 and $dt2 and $dt1->epoch < $now->epoch < $dt2->epoch ) {
-			return 1;
-		}
-	}
-	else {
-		if ( my $dt = $strp->parse_datetime($range) ) {
-			if ( $dt->ymd eq $now->ymd ) {
-				return 1;
-			}
-		}
-	}
-	return;
-}
-
-sub weekday_is_today {
-	my ($weekdys) = @_;
-
-	return ( ${ $weekdys->[ $now->day_of_week - 1 ] } == 1 );
-}
-
 my @weekdays = (qw(Mo Di Mi Do Fr Sa So));
 my @months   = (qw(0 I II III IV V VI VII VIII IX X XI XII));
 my %weekday  = map { ( $weekdays[$_] => $_ ) } ( 0 .. $#weekdays );
@@ -761,90 +726,12 @@ for my $bork (qw(104 1588 1700 77689 78112 939 2310)) {
 	}
 }
 
-my %smap;
-for my $train_no ( keys %map ) {
-	if ( @{ $map{$train_no} } == 1 ) {
-		$smap{$train_no} = $map{$train_no}[0];
-	}
-	else {
-		my $latest_valid;
-		my $valid_count = 0;
-		for my $train ( @{ $map{$train_no} } ) {
-			my $is_valid = 0;
-			for my $schedule ( @{ $train->{schedules} // [] } ) {
-				for my $valid ( @{ $schedule->{valid} // [] } ) {
-					if ( range_is_today($valid) ) {
-						$is_valid = 1;
-					}
-				}
-
-				# invalid may override valid for certain days
-				for my $invalid ( @{ $schedule->{invalid} // [] } ) {
-					if ( range_is_today($invalid) ) {
-						$is_valid = 0;
-					}
-				}
-			}
-			if ($is_valid) {
-				$latest_valid = $train;
-				$valid_count++;
-			}
-		}
-		if ( $valid_count == 1 ) {
-			$smap{$train_no} = $latest_valid;
-		}
-		else {
-			$valid_count  = 0;
-			$latest_valid = undef;
-			for my $train ( @{ $map{$train_no} } ) {
-				my $is_valid = 0;
-				for my $schedule ( @{ $train->{schedules} // [] } ) {
-					for my $valid ( @{ $schedule->{valid} // [] } ) {
-						if (    range_is_today($valid)
-							and weekday_is_today( $schedule->{weekdays} ) )
-						{
-							$is_valid = 1;
-						}
-					}
-
-					# invalid may override valid for certain days
-					for my $invalid ( @{ $schedule->{invalid} // [] } ) {
-						if ( range_is_today($invalid) ) {
-							$is_valid = 0;
-						}
-					}
-				}
-				if ($is_valid) {
-					$latest_valid = $train;
-					$valid_count++;
-				}
-			}
-			if ( $valid_count == 1 ) {
-				$smap{$train_no} = $latest_valid;
-			}
-		}
-	}
-}
-
 # use canonical output (i.e., sort hash keys) to allow for easy diffing.
-if ( $ENV{WITH_VARIANTS} ) {
-	say JSON->new->utf8->canonical->encode(
-		{
-			deprecated     => \0,
-			source         => $wr_name,
-			train          => {%smap},
-			train_variants => {%map},
-			valid          => $valid,
-		}
-	);
-}
-else {
-	say JSON->new->utf8->canonical->encode(
-		{
-			deprecated => \0,
-			source     => $wr_name,
-			train      => {%smap},
-			valid      => $valid,
-		}
-	);
-}
+say JSON->new->utf8->canonical->encode(
+	{
+		deprecated     => \0,
+		source         => $wr_name,
+		train_variants => {%map},
+		valid          => $valid,
+	}
+);
diff --git a/bin/extract-zugbildung-for-today b/bin/extract-zugbildung-for-today
new file mode 100755
index 0000000..43922e4
--- /dev/null
+++ b/bin/extract-zugbildung-for-today
@@ -0,0 +1,124 @@
+#!/usr/bin/env perl
+# Copyright (C) 2020 Daniel Friesel
+#
+# SPDX-License-Identifier: BSD-2-Clause
+
+use strict;
+use warnings;
+use 5.020;
+use utf8;
+
+use DateTime;
+use DateTime::Format::Strptime;
+use File::Slurp qw(read_file);
+use JSON;
+
+my $strp = DateTime::Format::Strptime->new(
+	pattern   => '%F',
+	time_zone => 'Europe/Berlin'
+);
+my $now = DateTime->now( time_zone => 'Europe/Berlin' );
+
+sub range_is_today {
+	my ($range) = @_;
+
+	if ( $range =~ m{^(.*)/(.*)$} ) {
+		my $dt1 = $strp->parse_datetime($1);
+		my $dt2 = $strp->parse_datetime($2);
+		if ( $dt1 and $dt2 and $dt1->epoch < $now->epoch < $dt2->epoch ) {
+			return 1;
+		}
+	}
+	else {
+		if ( my $dt = $strp->parse_datetime($range) ) {
+			if ( $dt->ymd eq $now->ymd ) {
+				return 1;
+			}
+		}
+	}
+	return;
+}
+
+sub weekday_is_today {
+	my ($weekdys) = @_;
+
+	return ( ${ $weekdys->[ $now->day_of_week - 1 ] } == 1 );
+}
+
+if ( @ARGV == 0 ) {
+	exit 1;
+}
+
+my $data = JSON->new->utf8->decode( read_file( $ARGV[0] ) );
+
+my %map = %{ $data->{train_variants} };
+my %smap;
+
+for my $train_no ( keys %map ) {
+	if ( @{ $map{$train_no} } == 1 ) {
+		$smap{$train_no} = $map{$train_no}[0];
+	}
+	else {
+		my $latest_valid;
+		my $valid_count = 0;
+		for my $train ( @{ $map{$train_no} } ) {
+			my $is_valid = 0;
+			for my $schedule ( @{ $train->{schedules} // [] } ) {
+				for my $valid ( @{ $schedule->{valid} // [] } ) {
+					if ( range_is_today($valid) ) {
+						$is_valid = 1;
+					}
+				}
+
+				# invalid may override valid for certain days
+				for my $invalid ( @{ $schedule->{invalid} // [] } ) {
+					if ( range_is_today($invalid) ) {
+						$is_valid = 0;
+					}
+				}
+			}
+			if ($is_valid) {
+				$latest_valid = $train;
+				$valid_count++;
+			}
+		}
+		if ( $valid_count == 1 ) {
+			$smap{$train_no} = $latest_valid;
+		}
+		else {
+			$valid_count  = 0;
+			$latest_valid = undef;
+			for my $train ( @{ $map{$train_no} } ) {
+				my $is_valid = 0;
+				for my $schedule ( @{ $train->{schedules} // [] } ) {
+					for my $valid ( @{ $schedule->{valid} // [] } ) {
+						if (    range_is_today($valid)
+							and weekday_is_today( $schedule->{weekdays} ) )
+						{
+							$is_valid = 1;
+						}
+					}
+
+					# invalid may override valid for certain days
+					for my $invalid ( @{ $schedule->{invalid} // [] } ) {
+						if ( range_is_today($invalid) ) {
+							$is_valid = 0;
+						}
+					}
+				}
+				if ($is_valid) {
+					$latest_valid = $train;
+					$valid_count++;
+				}
+			}
+			if ( $valid_count == 1 ) {
+				$smap{$train_no} = $latest_valid;
+			}
+		}
+	}
+}
+
+delete $data->{train_variants};
+$data->{train} = {%smap};
+
+say JSON->new->utf8->canonical->encode($data);