add importer for anarchisitisches radio

5 files changed, 203 insertions, 10 deletions

diff --git a/INFO b/INFO
index 1ebf2bf..a8f585d 100644
--- a/INFO
+++ b/INFO
@@ -132,3 +132,12 @@
    Upload: ca. 1 Woche danach
    Such-Kriterium: cba:broadcastDate wenn gesetzt sonst cba:productionDate
    RH: 2wöchentlich, Woche 1/3, Samstag, 08:00-10:00
+
+- (ar) Anarchistisches Radio
+   https://www.a-radio.net/feed/mp3
+   (Feed enthält nur Titel und keine Beschreibungstexte)
+
+   Original: wöchentlich Sonntags
+   Upload: meist am selben Tag - manchaml sehr viel später
+   Such-Kriterium: Datum im Titel, Format (dd/mm/YYYY)
+   RH: wöchentlich, Mittwoch, 22:00-13:00
diff --git a/Makefile b/Makefile
index 72ba286..3464884 100644
--- a/Makefile
+++ b/Makefile
@@ -26,7 +26,7 @@ endif
 .PHONY: clean distclean
 
 EXECUTABLE := rhautoimport
-IMPORTER := ba bo btl cb cr dm dn ek eu gd lr oi po rs ra tr
+IMPORTER := ar ba bo btl cb cr dm dn ek eu gd lr oi po rs ra tr
 EXECUTABLES := $(IMPORTER:%=$(EXECUTABLE)-%)
 
 all: $(EXECUTABLE)
diff --git a/TODO b/TODO
index dd10967..4cbe217 100644
--- a/TODO
+++ b/TODO
@@ -1,12 +1,3 @@
-- (ar) Anarchistisches Radio
-   https://www.a-radio.net/feed/mp3
-   (Feed enthält nur Titel und keine Beschreibungstexte)
-
-   Original: wöchentlich Sonntags
-   Upload: meist 1-2 Tage danach
-   Such-Kriterium: Datum im Titel, Format (dd/mm/YYYY)
-   RH: wöchentlich, Donnerstag, 16:00-17:00
-
 - (ja) Jazz 'round midnight
    https://cba.fro.at/series/jazz-round-midnight/feed
    (Beschreibungstexte sind meist leer)
diff --git a/debian/changelog b/debian/changelog
index 11a37d6..6ee4068 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,9 @@
+rhautoimport (71-1) unstable; urgency=medium
+
+  * add import for Anarchistisches Radio
+
+ -- Christian Pointner <equinox@helsinki.at>  Fri, 18 Feb 2022 11:09:27 +0100
+
 rhautoimport (70-1) unstable; urgency=medium
 
   * remove importer for derive
diff --git a/rhautoimport-ar b/rhautoimport-ar
new file mode 100755
index 0000000..bb1d1b2
--- /dev/null
+++ b/rhautoimport-ar
@@ -0,0 +1,187 @@
+#!/usr/bin/perl -w
+#
+#
+#  rhautoimport
+#
+#  Copyright (C) 2009-2017 Christian Pointner <equinox@helsinki.at>
+#
+#  This file is part of rhautoimport.
+#
+#  rhautoimport is free software: you can redistribute it and/or modify
+#  it under the terms of the GNU General Public License as published by
+#  the Free Software Foundation, either version 3 of the License, or
+#  any later version.
+#
+#  rhautoimport is distributed in the hope that it will be useful,
+#  but WITHOUT ANY WARRANTY; without even the implied warranty of
+#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#  GNU General Public License for more details.
+#
+#  You should have received a copy of the GNU General Public License
+#  along with rhautoimport. If not, see <http://www.gnu.org/licenses/>.
+#
+
+use strict;
+use Date::Calc;
+use DateTime::Format::Strptime;
+use XML::Feed;
+use XML::Feed::Entry;
+use XML::Feed::Content;
+use XML::Feed::Enclosure;
+use URI::URL;
+use HTML::Entities;
+
+use lib '/usr/local/share/rhautoimport/';
+use rhautoimport;
+
+my $STAT_FILE = $ENV{'HOME'} . "/rhautoimport-ar.stat";
+my $RSS_URL = "https://cba.fro.at/podcast/anarchistisches-radio/feed";
+$XML::Feed::MULTIPLE_ENCLOSURES=1;
+my $RD_GROUP = "anarcradio";
+my $PV_ID = '294';
+my $TITLE="Anarchistisches Radio";
+my $LAST_RUN = 0;
+my $upload_delay_days = 0;   # files are usally there on the same day of the original broadcast
+
+binmode(STDIN, ":utf8");
+binmode(STDOUT, ":utf8");
+binmode(STDERR, ":utf8");
+
+if($#ARGV >= 0 && $ARGV[0] eq 'last') {
+  print "!!!This is the last attempt, there won't be a retry on error!!!\n";
+  $LAST_RUN = 1;
+}
+rhautoimport::report_title_and_last($TITLE, $LAST_RUN);
+
+
+my @allowed_dbs = rhautoimport::get_dropboxes($RD_GROUP);
+if(!defined $allowed_dbs[0] && defined $allowed_dbs[1]) {
+  print "$allowed_dbs[1]\n";
+  exit 1;
+}
+
+if(scalar(@allowed_dbs) != 1) {
+  print "found more or less than one Dropbox for this group?!\n";
+  exit 1;
+}
+my $show_id = $allowed_dbs[0]->{'SHOWID'};
+my $show_title = $allowed_dbs[0]->{'SHOWTITLE'};
+
+
+my @today = Date::Calc::Today();
+my $dow = Date::Calc::Day_of_Week(@today);
+if ($dow > 3) {
+  print "File won't be available by now!\n";
+  exit 42;
+}
+my @tmp = Date::Calc::Standard_to_Business(@today);
+$tmp[2] = 7;
+my @broadcast_date = Date::Calc::Business_to_Standard(@tmp);
+@broadcast_date = Date::Calc::Add_Delta_Days(@broadcast_date, -7);
+$tmp[2] = 3;
+my @import_date = Date::Calc::Business_to_Standard(@tmp);
+
+print "today: " . Date::Calc::Date_to_Text(@today) . "\n";
+print "day of original broadcast: " . Date::Calc::Date_to_Text(@broadcast_date) . "\n";
+print "day of Radio Helsinki broadcast: " . Date::Calc::Date_to_Text(@import_date) . "\n";
+
+if(Date::Calc::Delta_Days(@broadcast_date, @today) <= $upload_delay_days) {
+  print "File won't be available by now!\n";
+  exit 42;
+}
+
+my $id = sprintf("%04d-%02d-%02d", @import_date);
+my $bd = sprintf("%04d-%02d-%02d", @broadcast_date);
+my $bdtitle = sprintf("%02d/%02d/%04d", $broadcast_date[2], $broadcast_date[1], $broadcast_date[0]);
+
+print "looking for entries with titles containing '$bdtitle' in RSS Feed\n";
+print " -> $RSS_URL\n";
+
+my ($result, $feed) = rhautoimport::fetch_parse_rss_cba($RSS_URL);
+unless ($result) {
+  print "Error fetching feed: $feed\n";
+  exit 1;
+}
+
+my $uri = "";
+my $file = "";
+my $sum_title = "";
+my $sum_text = "";
+
+my $i = 0;
+for my $entry ($feed->entries) {
+  $i++;
+  next unless $entry->enclosure;
+
+  next if(index($entry->title, $bdtitle) == -1);
+
+  my $j = 0;
+  for my $enclosure($entry->enclosure) {
+    $j++;
+    if($enclosure->type eq "audio/mpeg" || $enclosure->type eq "audio/ogg") {
+      print "$i/$j: (" . $enclosure->type . ", " . $enclosure->length . ") "  . $enclosure->url . "\n";
+      my $orig_uri = $enclosure->url;
+      if($orig_uri =~ /^(.*)_cut(\.[^\.]+)$/) {
+        $orig_uri = $1 . $2;
+      }
+
+      $uri = new URI::URL($orig_uri);
+      my @path = $uri->path_components;
+      $file = $path[-1];
+
+      my $current_file = `cat $STAT_FILE`;
+      if($current_file eq $file) {
+        print "Already downloaded file of today\n";
+        exit 42;
+      }
+      if(!rhautoimport::check_file_extension($file)) {
+        print "\n\nThe extension of the matching file '". $file . "' seems to be wrong - manual import necessary!!!\n";
+        print "\n\n --> https://import.helsinki.at/shows/$show_id\n";
+        exit 1;
+      }
+
+      $sum_title = decode_entities($entry->title);
+      $sum_text = decode_entities($entry->content->body);
+      print "summary:\n" . $sum_title . "\n\n" . $sum_text . "\n";
+      last;
+    }
+  }
+  last if $uri ne "";
+}
+if($uri eq "") {
+  print "No Entry found from $bd - ";
+  if($LAST_RUN) {
+    print "giving up, manual import necessary!!!\n";
+    print "\n\n --> https://import.helsinki.at/shows/$show_id\n";
+  } else {
+    print "will retry later\n";
+  }
+  exit 1;
+}
+
+print "\n\nwill import '$uri' to show $show_id, $show_title\n\n";
+
+my ($ret, $log) = rhautoimport::import_uri($show_id, $uri->as_string);
+if($ret) {
+  print "\nImport Error:\n\n";
+  print $log;
+  print "\n\nNot adding PV note!!";
+  exit 1;
+}
+print "\nImport Success:\n\n";
+print $log;
+print "\n";
+my $exit_code = 0;
+($ret, $log) = rhautoimport::pv_add_note($sum_title, $sum_text, $PV_ID, $id, "1");
+print $log;
+if($ret) {
+  print "\nIgnoring failed note import - manual intervention necessary!\n";
+  $exit_code = 23;
+}
+
+unlink($STAT_FILE);
+open(my $fhs, '>', $STAT_FILE);
+print $fhs "$file";
+close($fhs);
+
+exit $exit_code;