From ff16ac125c01bcc6e1e4f766cb554f19199ee193 Mon Sep 17 00:00:00 2001 From: Christian Pointner Date: Thu, 29 Jun 2017 14:28:02 +0200 Subject: added importer for derive diff --git a/INFO b/INFO index 43d496b..566e1bf 100644 --- a/INFO +++ b/INFO @@ -8,3 +8,24 @@ Upload: immer ein paar Tage früher, meistens Mi oder Do davor Such-Kriterium: cba:broadcastDate RH: 4wöchentlich, Woche 2, Dienstag, 08:30-09:30 + +- Fokus Bildung: + https://cba.fro.at/series/fokus-bildung-bildung-im-fokus/feed + + Original: erster Montag im Monat + Upload: immer ein paar Tage früher, meistens Fr + Such-Kriterium: cba:broadcastDate + nicht immer gesetzt!! (~1 von 5 mal) + RH: 4wöchentlich, Woche 2, Mittwoch, 14:30-15:00 + +- derive: + https://cba.fro.at/series/derive-radio-fuer-stadtforschung/feed + + Original: erster Dienstag im Monat + Upload: meistens selber Tag oder einen davor + Such-Kriterium: cba:broadcastDate + wird erst seit wenigen Monaten gesetzt, + Sendung 03-2017 und 05-2017 fehlen in der CBA aber + sind laut O94 Homepage gelaufen... + RH: 4wöchentlich, Woche 3, Dienstag, 16:00-16:30 + 4wöchtenlich, Woche 2, Mittwoch, 14:00-14:30, Wiederholung diff --git a/Makefile b/Makefile index 95f55e1..9b9d1aa 100644 --- a/Makefile +++ b/Makefile @@ -26,7 +26,7 @@ endif .PHONY: clean distclean EXECUTABLE := rhautoimport -IMPORTER := btl dn ek er eu fb fl gd nw oi po rs ra tr wr +IMPORTER := btl dn dv ek er eu fb fl gd nw oi po rs ra tr wr EXECUTABLES := $(IMPORTER:%=$(EXECUTABLE)-%) all: $(EXECUTABLE) diff --git a/TODO b/TODO index b7f682e..e69de29 100644 --- a/TODO +++ b/TODO @@ -1,20 +0,0 @@ -- Fokus Bildung: - https://cba.fro.at/series/fokus-bildung-bildung-im-fokus/feed - - Original: erster Montag im Monat - Upload: immer ein paar Tage früher, meistens Fr - Such-Kriterium: cba:broadcastDate - nicht immer gesetzt!! (~1 von 5 mal) - RH: 4wöchentlich, Woche 2, Mittwoch, 14:30-15:00 - -- derive: - https://cba.fro.at/series/derive-radio-fuer-stadtforschung/feed - - Original: erster Dienstag im Monat - Upload: meistens selber Tag oder einen davor - Such-Kriterium: cba:broadcastDate - wird erst seit wenigen Monaten gesetzt, - Sendung 03-2017 und 05-2017 fehlen in der CBA aber - sind laut O94 Homepage gelaufen... - RH: 4wöchentlich, Woche 3, Dienstag, 16:00-16:30 - 4wöchtenlich, Woche 2, Mittwoch, 14:00-14:30, Wiederholung diff --git a/debian/changelog b/debian/changelog index 530d397..33f4236 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,8 +1,14 @@ -rhautoimport (31-1) UNRELEASED; urgency=medium +rhautoimport (32-1) unstable; urgency=medium + + * new importer for derive + + -- Christian Pointner Thu, 29 Jun 2017 14:27:28 +0200 + +rhautoimport (31-1) unstable; urgency=medium * new importer for Fokus Bildung - -- Christian Pointner Thu, 29 Jun 2017 13:51:30 +0200 + -- Christian Pointner Thu, 29 Jun 2017 13:51:30 +0200 rhautoimport (30-1) unstable; urgency=medium diff --git a/rhautoimport-dv b/rhautoimport-dv new file mode 100755 index 0000000..2fd75b4 --- /dev/null +++ b/rhautoimport-dv @@ -0,0 +1,249 @@ +#!/usr/bin/perl -w +# +# +# rhautoimport +# +# Copyright (C) 2009-2016 Christian Pointner +# +# 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 . +# + +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 RHRD::utils; + +use lib '/usr/local/share/rhautoimport/'; +use rhautoimport; + +my $STAT_FILE = $ENV{'HOME'} . "/rhautoimport-dv.stat"; +my $RSS_URL = "https://cba.fro.at/series/derive-radio-fuer-stadtforschung/feed"; +$XML::Feed::MULTIPLE_ENCLOSURES=1; +my $RD_GROUP = "derive"; +my $PV_ID = '293'; +my $TITLE = "dérive"; +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"); + +my $curweek = RHRD::utils::get_rd_week(); + +if($curweek == 3) { + 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 @today = Date::Calc::Today(); +print "today: " . Date::Calc::Date_to_Text(@today) . " (Week: " . $curweek . ")\n"; + +my @import_date = Date::Calc::Standard_to_Business(@today); +$import_date[2] = 2; +@import_date = Date::Calc::Business_to_Standard(@import_date); + +my $dow = Date::Calc::Day_of_Week(@today); +if($curweek == 1) { + @import_date = Date::Calc::Add_Delta_Days(@import_date, 14); +} elsif($curweek == 2) { + @import_date = Date::Calc::Add_Delta_Days(@import_date, 7); +} elsif($curweek == 3) { + if($dow > 2) { + @import_date = Date::Calc::Add_Delta_Days(@import_date, 28); + } +} else { + @import_date = Date::Calc::Add_Delta_Days(@import_date, 21); +} +print "day of next Radio Helsinki broadcast: " . Date::Calc::Date_to_Text(@import_date) . "\n"; +my @import_date_reb = Date::Calc::Add_Delta_Days(@import_date, 22); +print "day of this show's rebroadcast: " . Date::Calc::Date_to_Text(@import_date_reb) . "\n"; +my @reb_date_last = Date::Calc::Add_Delta_Days(@import_date_reb, -28); +print "day of last show's rebroadcast: " . Date::Calc::Date_to_Text(@reb_date_last) . "\n"; + +my @allowed_dbs = rhautoimport::get_dropboxes($RD_GROUP); +if(!defined $allowed_dbs[0] && defined $allowed_dbs[1]) { + print "$allowed_dbs[1]\n"; + exit 1; +} + +my $idx = 0; +my $idx_reb = 1; +if(scalar(@allowed_dbs) != 2) { + print "found more or less than 2 Dropboxes for this group?!\n"; + exit 1; +} +my $show_id = $allowed_dbs[$idx]->{'SHOWID'}; +my $show_title = $allowed_dbs[$idx]->{'SHOWTITLE'}; +my $show_id_reb = $allowed_dbs[$idx_reb]->{'SHOWID'}; +my $show_title_reb = $allowed_dbs[$idx_reb]->{'SHOWTITLE'}; + +my @broadcast_date = Date::Calc::Nth_Weekday_of_Month_Year($import_date[0], $import_date[1], 2, 1); +if(Date::Calc::Delta_Days(@broadcast_date, @import_date) <= 0) { + my @tmp_date = Date::Calc::Add_Delta_YM(@broadcast_date, 0, -1); + @broadcast_date = Date::Calc::Nth_Weekday_of_Month_Year($tmp_date[0], $tmp_date[1], 2, 1); +} +print "day of latest original broadcast before next Radio Helsinki broadcast: " . Date::Calc::Date_to_Text(@broadcast_date) . "\n"; + +if(Date::Calc::Delta_Days(@today, @reb_date_last) >= 0) { + print "the rebroadcast from last show has not been played yet!\n"; + exit 42; +} +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); + +print "looking for files for broadcast date $bd 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; + + my $entry_bdt = $entry->{'entry'}{'https://cba.fro.at/help#feeds'}{'broadcastDate'}; + next unless (defined $entry_bdt); + my $bdt = DateTime::Format::Strptime::strptime("%a, %d %b %Y %H:%M:%S", $entry_bdt); + + next unless($broadcast_date[0] == $bdt->year && $broadcast_date[1] == $bdt->month && $broadcast_date[2] == $bdt->day); + + 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_stat = `cat $STAT_FILE`; + my ($current_id, $current_file) = $current_stat =~ m/^(.*)\n(.*)/; + if($current_id eq $id && $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; +} + +my $exit_code = 0; +print "\n\nwill import '$uri' to show $show_id, $show_title\n"; +my ($ret, $log, $keptfile_uri) = rhautoimport::import_uri($show_id, $uri->as_string, "keep"); +if($ret == 0) { + print "\nImport Success:\n\n"; + print $log; + print "\n"; + ($ret, $log) = rhautoimport::pv_add_note($sum_title, $sum_text, $PV_ID, sprintf("%04d-%02d-%02d", @import_date), "1"); + print $log; + if($ret) { + print "\nIgnoring failed note import - manual intervention necessary!\n"; + $exit_code = 23; + } + + print "\n"; + + if(!defined($keptfile_uri)) { + print "rhimportd didn't provide a clue where to find the kept file... will import '$uri' to rebroadcast $show_id_reb, $show_title_reb\n"; + ($ret, $log) = rhautoimport::import_uri($show_id_reb, $uri->as_string); + } else { + print "re-using kept file '$keptfile_uri' to import rebroadcast $show_id_reb, $show_title_reb\n"; + ($ret, $log) = rhautoimport::import_uri($show_id_reb, $keptfile_uri, "delete"); + } + + if($ret == 0) { + print "\nImport Success:\n\n"; + print $log; + print "\n"; + ($ret, $log) = rhautoimport::pv_add_note($sum_title, $sum_text, $PV_ID, sprintf("%04d-%02d-%02d", @import_date_reb), "2"); + print $log; + if($ret) { + print "\nIgnoring failed note import - manual intervention necessary!\n"; + $exit_code = 23; + } + + } else { + print "\nImport Error:\n\n"; + print $log; + print "\n\nNot adding PV note!!"; + print "\n\nSince the import for the regular broadcast went through we will not retry this import!!!"; + print "\nYou need to manually import the rebroadcast."; + print "\n\n --> https://import.helsinki.at/shows/$show_id_reb\n"; + $exit_code = 23; + } +} else { + print "\nImport Error:\n\n"; + print $log; + print "\n\nNot adding PV note!!"; + exit 1; +} + +unlink($STAT_FILE); +open(my $fhs, '>', $STAT_FILE); +print $fhs "$id\n$file"; +close($fhs); + +exit $exit_code; -- cgit v0.10.2