forked from rev138/pod_feeder
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpod_feeder.pl
executable file
·683 lines (564 loc) · 27.9 KB
/
pod_feeder.pl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
#!/usr/bin/perl
# vim:set sw=8 ts=8 sts=8 ft=perl expandtab:
##
## pod_feeder.pl
##
## A script to auto-post RSS/Atom feeds to a Diaspora account
##
## created 20150416 by Brian Ó <[email protected]>
## (on diaspora: [email protected])
##
## I owe a great debt to the code of diaspora-rss-bot (https://github.com/spkdev/diaspora-rss-bot)
## for helping me understand how play nice with CSRF tokens et al
##
use strict;
use warnings;
use utf8;
use LWP::UserAgent;
use URI::Escape;
use HTML::Entities;
use JSON;
use XML::Simple;
use DBI;
use Unicode::Normalize 'normalize';
use Getopt::Long;
use HTML::FormatMarkdown;
my $opts = {
'database' => './pod_feeder.db',
'limit' => 0,
'timeout' => 72, # hours
'via' => 'pod_feeder',
};
my @auto_tags = ();
my @ignored_tags = ();
my @aspect_ids = ();
GetOptions(
$opts,
'aspect-id|a=s' => \@aspect_ids,
'auto-tag|t=s' => \@auto_tags,
'body',
'category-tags|c',
'database|d=s',
'embed-image|b',
'feed-id|i=s',
'feed-url|f=s',
'fetch-only|o',
'help|h', => \&usage,
'ignore-tag|n=s', => \@ignored_tags,
'insecure|s=s',
'limit|x=i',
'no-branding',
'password|p=s',
'pod-url|l=s',
'post-raw-links|w',
'timeout|m=i',
'title-tags|e',
'url-tags|r',
'user-agent|g=s',
'username|u=s',
'via|v=s',
);
# defaults to 'public' if no aspect ids are specified
$aspect_ids[@aspect_ids] = 'public' unless @aspect_ids;
# some feeds block bots, so you can spoof the user agent string with something such as 'Mozilla/5.0' to get around this
my $user_agent = $opts->{'user-agent'} || undef;
# initialize the database if it does not exist
eval { init_database( $opts->{'database'} ) };
die "ERROR: Coult not initialize the database: $@" if $@;
# fetch the feed
my ( $fetched, $feed ) = fetch_feed( $opts->{'feed-url'}, $user_agent );
if( $fetched ){
eval {
# update the database
update_feed(
$feed,
db_file => $opts->{'database'},
feed_id => $opts->{'feed-id'},
auto_tags => hashtagify( \@auto_tags ),
ignored_tags => hashtagify( \@ignored_tags ),
extract_tags_from_url => $opts->{'url-tags'},
extract_tags_from_title => $opts->{'title-tags'},
tag_categories => $opts->{'category-tags'},
);
};
warn "$@" if $@;
eval {
# publish new feed items to the pod, unless the user specified --fetch-only
publish_feed_items(
db_file => $opts->{'database'},
embed_image => $opts->{'embed-image'},
feed_id => $opts->{'feed-id'},
timeout => $opts->{'timeout'},
pod_url => $opts->{'pod-url'},
username => $opts->{'username'},
password => $opts->{'password'},
aspect_ids => \@aspect_ids,
raw_link => $opts->{'post-raw-links'},
limit => $opts->{'limit'},
no_branding => $opts->{'no-branding'},
via => $opts->{'via'},
insecure => $opts->{'insecure'},
body => $opts->{'body'}
) unless $opts->{'fetch-only'};
};
warn "$@" if $@;
}
else {
die "Error fetching $opts->{'feed-url'} : " . $feed->code . ' ' . $feed->message;
}
# publishes un-posted items in the database
sub publish_feed_items {
my ( %params ) = @_;
my @updates = ();
my $query_string = "SELECT guid, title, link, image, image_title, hashtags, body FROM feeds WHERE feed_id == ? AND posted == 0 AND timestamp > ? ORDER BY timestamp";
my $dbh = connect_to_db( $params{'db_file'} );
# limit the number of items published if limit is specified
$query_string .= " LIMIT $params{'limit'}" if $params{'limit'} > 0;
my $sth = $dbh->prepare( $query_string ) or die "Can't prepare statement: $DBI::errstr";
$sth->execute( $params{'feed_id'}, time - ( $params{'timeout'} * 3600 ) ) or die "Can't execute statement: $DBI::errstr";
while( my $row = $sth->fetchrow_hashref() ){
push( @updates, $row );
}
foreach my $update ( @updates ){
my $content = $update->{'hashtags'};
if( $params{'embed_image'} and length $update->{'image'} ){
my $image_link = '[](' . $update->{'link'} . ')';
$content = "$image_link\n$content";
}
# to hyperlink the title or not to hyperlink the title...
if( $params{'raw_link'} ){
$content = '### ' . $update->{'title'} . "\n\n" . $update->{'link'} . "\n" . $content;
}
else {
$content = '### [' . $update->{'title'} . '](' . $update->{'link'} . ")\n\n" . $content;
}
$content .= "\n" . $update->{'body'} if $params{'body'};
print "Publishing $params{'feed_id'}\t$update->{'guid'}\n";
my $post = publish_post( $content, %params );
# mark the item as successfully posted
if( $post->is_success ){
$sth = $dbh->prepare( "UPDATE feeds SET posted = 1 WHERE guid = ?" ) or die "Can't prepare statement: $DBI::errstr";
$sth->execute( $update->{'guid'} ) or die "Can't execute statement: $DBI::errstr";
}
else {
warn $post->code . ' ' . $post->message;
}
# Now, don't be hasty, master Meriadoc
sleep 1;
}
$dbh->disconnect();
}
# adds new feed items to the database
sub update_feed {
my ( $feed, %params ) = @_;
$params{'auto_tags'} = [] unless defined $params{'auto_tags'};
$params{'extract_tags_from_url'} = 0 unless defined $params{'extract_tags_from_url'};
$params{'tag_categories'} = 0 unless defined $params{'tag_categories'};
$params{'ignored_tags'} = [] unless defined $params{'ignored_tags'};
my $items = get_feed_items( $feed, %params );
my $dbh = connect_to_db( $params{'db_file'} );
foreach my $item ( @$items ){
# strip junk
map { $item->{$_} =~ s/^\s+|\s+$//g } keys %$item;
map { $item->{$_} =~ s/^\n+|\n+$//g } keys %$item;
# decode uft8 strings before storing in the db
map { utf8::decode($item->{'title'}) } keys %$item;
map { utf8::decode($item->{'body'}) } keys %$item;
# check to see if it exists already
my $sth = $dbh->prepare("SELECT guid FROM feeds WHERE guid == ? LIMIT 1") or die "Can't prepare statement: $DBI::errstr";
$sth->execute( $item->{'guid'} ) or die "Can't execute statement: $DBI::errstr";
my $row = $sth->fetch();
# and if not, insert it
unless( defined $row ){
$sth = $dbh->prepare(
"INSERT INTO feeds( guid, feed_id, title, body, link, image, image_title, hashtags, posted, timestamp ) VALUES( ?, ?, ?, ?, ?, ?, ?, ?, ?, ? )"
) or die "Can't prepare statement: $DBI::errstr";
$sth->execute(
$item->{'guid'},
$params{'feed_id'},
$item->{'title'},
$item->{'body'},
$item->{'link'},
$item->{'image'},
$item->{'image_title'},
join( ' ', @{$item->{'hashtags'}} ),
0,
time,
) or die "Can't execute statement: $DBI::errstr";
}
}
$dbh->disconnect();
}
sub connect_to_db {
my ( $db_file ) = @_;
my $dbh = DBI->connect("dbi:SQLite:dbname=$db_file", '', '', { RaiseError => 1, sqlite_unicode => 1 } ) or die $DBI::errstr;
return $dbh;
}
# parse the individual items from the feed
sub get_feed_items {
my ( $feed, %params ) = @_;
my @items = ();
my $list = decode_feed( $feed );
$params{'auto_tags'} = [] unless defined $params{'auto_tags'};
$params{'extract_tags_from_url'} = 0 unless defined $params{'extract_tags_from_url'};
$params{'tag_categories'} = 0 unless defined $params{'tag_categories'};
$params{'ignored_tags'} = [] unless defined $params{'ignored_tags'};
foreach my $item ( @$list ){
my $link = $item->{'link'};
# no link, no go
next unless defined $link and ref $link ne 'HASH' and ref $link ne 'ARRAY';
my @hashtags = ();
my $guid = undef;
my $image = '';
my $image_title = '';
# strip trailing /
$link =~ s/\/+$// if defined $link;
# add user-specified tags
push( @hashtags, @{$params{'auto_tags'}} ) if defined $params{'auto_tags'};
# try to guess tags from the url
if( $params{'extract_tags_from_url'} ){
# grab last part of url
$link =~ m/\/([^\/]+)$/;
my $link_part = $1;
# strip off any url params
$link_part =~ s/(\?.*)$//;
# split up string
my @parts = split( /([^(\p{Letter}|\p{Number})]|\p{Punctuation})/, $link_part );
push( @hashtags, @parts );
}
# try to guess tags from the title
if( $params{'extract_tags_from_title'} ){
my $title = $item->{'title'};
# strip apostrophes
$title =~ s/('|`)//g;
# split up string on non-alphanumerics
my @parts = split( /([^(\p{Letter}|\p{Number})]|\p{Punctuation})/, $title );
my @tags = ();
foreach my $part ( @parts ){
push( @tags, $part ) unless $part =~ m/^(\s+)?$/;
}
push( @hashtags, @tags );
}
# try to extract tags from the feed categories (if they exist)
if( $params{'tag_categories'} and defined $item->{'category'} ){
my @categories = ();
if( ref $item->{'category'} ne 'ARRAY' ){
@categories = ( $item->{'category'} );
}
else {
@categories = @{$item->{'category'}};
}
push ( @hashtags, @categories );
}
# extract image link and hover text from content:encoded if it exists
if( defined $item->{'content:encoded'} ){
$item->{'content:encoded'} =~ /img .* ?src=\\?'(https?:\/\/[^']+)/ unless $item->{'content:encoded'} =~ /img .* ?src=\\?"(https?:\/\/[^"]+)/;
if( defined $1 ){
$image = $1;
$item->{'content:encoded'} =~ / title='([^']+)/ unless $item->{'content:encoded'} =~ / title="([^"]+)/;
$image_title = $1 if defined $1;
}
}
# extract image link and hover text from description if it exists
if( not length $image and defined $item->{'description'} ){
$item->{'description'} =~ /img .* ?src='(https?:\/\/[^']+)/ unless $item->{'description'} =~ /img .* ?src="(https?:\/\/[^"]+)/;
if( defined $1 ){
$image = $1;
$item->{'description'} =~ / title='([^']+)/ unless $item->{'description'} =~ / title="([^"]+)/;
$image_title = $1 if defined $1;
}
}
# extract the image link from the enclosure tag if it exists
if( not length $image and defined $item->{'enclosure'} and defined $item->{'enclosure'}->{'type'} and $item->{'enclosure'}->{'type'} =~ /^image\// ){
$image = $item->{'enclosure'}->{'url'} if defined $item->{'enclosure'}->{'url'};
}
# remove any query params from image link
$image =~ s/(\?.*)$//;
@hashtags = sort @hashtags;
if( defined $item->{'guid'} ){
if( ref $item->{'guid'} eq 'HASH' and defined $item->{'guid'}->{'content'} ){
$guid = $item->{'guid'}->{'content'};
}
elsif( ref $item->{'guid'} ne 'HASH' ) {
$guid = $item->{'guid'};
}
}
elsif( defined $item->{'id'} ){
$guid = $item->{'id'};
}
else { $guid = $link }
@hashtags = @{ hashtagify( \@hashtags ) };
# filter out ignored tags
for( my $t = 0; $t < @hashtags; $t++ ){
foreach my $ignored ( @{$params{'ignored_tags'}} ){
splice( @hashtags, $t, 1 ) if $hashtags[$t] eq $ignored;
}
}
my $body = HTML::FormatMarkdown->format_from_string($item->{'description'}, rm => 100000);
$body = HTML::FormatMarkdown->format_from_string($item->{'content:encoded'}, rm => 100000) if ($item->{'content:encoded'});
my $obj = {
guid => $guid,
title => $item->{'title'},
body => $body,
link => $link,
image => $image,
image_title => $image_title,
hashtags => \@hashtags,
};
$items[@items] = $obj;
}
# the last shall be first and the first shall be last
my @reversed = ();
for( my $i = $#items; $i >= 0; $i-- ){
$reversed[@reversed] = $items[$i];
}
return \@reversed;
}
# extract the data we need based on feed type (RSS v. Atom)
sub decode_feed{
my ( $feed ) = @_;
my @list = ();
# RSS
if( defined $feed->{'channel'} and defined $feed->{'channel'}->{'item'} ){
if( ref $feed->{'channel'}->{'item'} eq 'ARRAY' ){
@list = @{$feed->{'channel'}->{'item'}};
}
elsif( ref $feed->{'channel'}->{'item'} eq 'HASH' ){
if( length( keys %{$feed->{'channel'}->{'item'}} ) == 1 ){
$list[@list] = $feed->{'channel'}->{'item'}
}
else{
@list = values %{$feed->{'channel'}->{'item'}};
}
}
}
elsif( defined $feed->{'item'} ){
if( ref $feed->{'item'} eq 'ARRAY' ){
@list = @{$feed->{'item'}};
}
elsif( ref $feed->{'item'} eq 'HASH' ){
@list = values %{$feed->{'item'}};
}
}
# Atom
elsif( defined $feed->{'entry'} and ref $feed->{'entry'} eq 'HASH' ){
my $entries = $feed->{'entry'};
foreach my $guid ( keys %$entries ){
my $item = {
guid => $guid,
};
if( defined $entries->{$guid}->{'title'} ){
if( ref $entries->{$guid}->{'title'} eq 'HASH' and defined $entries->{$guid}->{'title'}->{'content'} ){
$item->{'title'} = $entries->{$guid}->{'title'}->{'content'};
}
elsif( ref $entries->{$guid}->{'title'} eq '' ){
$item->{'title'} = $entries->{$guid}->{'title'};
}
}
if( defined $entries->{$guid}->{'link'} ){
if( ref $entries->{$guid}->{'link'} eq 'HASH' and defined $entries->{$guid}->{'link'}->{'href'} ){
$item->{'link'} = $entries->{$guid}->{'link'}->{'href'};
}
elsif( ref $entries->{$guid}->{'link'} eq '' ){
$item->{'link'} = $entries->{$guid}->{'link'};
}
}
$item->{'category'} = $entries->{'category'} if defined $entries->{'category'};
if( defined $entries->{$guid}->{'summary'} ){
if( ref($entries->{$guid}->{'summary'}) eq 'HASH' and defined $entries->{$guid}->{'summary'}->{'content'} ){
$item->{'description'} = $entries->{$guid}->{'summary'}->{'content'};
}
else {
$item->{'description'} = $entries->{$guid}->{'summary'};
}
}
push( @list, $item ) if defined $item->{'link'} and defined $item->{'title'};
}
}
return \@list;
}
# fetch the feed and convert the XML to an data object
sub fetch_feed {
my ( $feed_url, $user_agent_string ) = @_;
my $ua = LWP::UserAgent->new();
$ua->agent( $user_agent_string ) if defined $user_agent_string;
my $response = $ua->get( $feed_url );
if( $response->is_success ){
return ( 1, XMLin normalize( 'D', $response->decoded_content ) );
}
else {
return ( 0, $response );
}
}
# sanitize and de-dupe tags
sub hashtagify {
my ( $list_ref ) = @_;
my %hashtags = ();
my @list = @$list_ref;
foreach my $item ( @list ){
# remove non-alphanumerics
$item =~ s/[^(\p{Letter}|\p{Number})]//g;
# drop stop words
# TODO : make these overridable
next if length( $item ) < 3;
next if lc( $item ) =~ m/^(a(lso|nd|ny|re)|been|but|can(not|t)?|e(ach|tc|very)|for|from|g(e|o)t|ha(d|ve)|has(nt)?|hers?|hi(m|s)|how|its|no(r|t)|ours?|she|some|th(an|at|em?|eirs?|(e|o)se|ey|eyre|is)|too|very|was|wh(at|en|o)|with|you(r|rs)?)$/;
# hashtagify it
$item = '#' . $item;
# use a hash here instead of an ordered list for auto-dedupe
$hashtags{ lc( $item ) } = undef;
}
my @deduped = keys %hashtags;
my @sorted = sort @deduped;
return \@sorted;
}
# publish a post to the pod
sub publish_post {
my ( $content, %params ) = @_;
my $posted = 0;
# create our user agent
my $ua = LWP::UserAgent->new( requests_redirectable => [ 'GET', 'HEAD', 'POST' ] );
# initialize an empty cookie jar
$ua->cookie_jar( {} );
# allow option for insecure certs
if( $params{'insecure'} ){
$ua->ssl_opts( verify_hostname => 0);
}
# log in
my $login_response = login( $ua, $params{'pod_url'}, $params{'username'}, $params{'password'} ) ;
# if we've logged in successfully, post the message
if( $login_response->is_success ){
# encode utf-8 characters
utf8::encode($content);
my $post = post_message( $ua, $params{'pod_url'}, $content, $params{'aspect_ids'}, %params );
#logout( $ua, $pod_url );
return $post;
}
else {
return $login_response;
}
}
# log in to the pod
sub login {
my ( $ua, $base_url, $username, $password ) = @_;
my $sign_in_url = "$base_url/users/sign_in";
$ua->cookie_jar->clear();
my ( $sign_in, $result ) = get_page( $ua, $sign_in_url );
if( $sign_in ){
my $csrf = extract_token( $result );
my $urlencoded_params = '';
my $params = {
$csrf->{'param'} => $csrf->{'token'},
'utf8' => '%E2%9C%93',
'user[username]' => $username,
'user[password]' => $password,
'user[remember_me]' => 1,
'commit' => 'Sign in'
};
foreach my $key ( keys %$params ){
$urlencoded_params .= uri_escape( $key ) . '=' . uri_escape( $params->{$key} ) . '&';
}
return $ua->post( $sign_in_url, 'Content' => $urlencoded_params, 'Content-Type' => 'application/x-www-form-urlencoded' );
}
else {
return $result;
}
}
# retreive a web page via GET
sub get_page {
my ( $ua, $url ) = @_;
my $response = $ua->get( $url );
if( $response->is_success ){
return ( 1, $response->decoded_content );
}
else {
return ( 0, $response );
}
}
# extract the CSRF token from the page's source code
sub extract_token {
my ( $html ) = @_;
my $csrf = {};
# parse out CSRF param and token
# $html =~ m/meta content="([^"]+)" name="csrf-param"/;
$html =~ m/meta name="csrf-param" content="([^"]+)"/;
$csrf->{'param'} = decode_entities( $1 ) if defined ( $1 );
# $html =~ m/meta content="([^"]+)" name="csrf-token"/;
$html =~ m/meta name="csrf-token" content="([^"]+)"/;
$csrf->{'token'} = decode_entities( $1 ) if defined ( $1 );
return $csrf if defined $csrf->{'param'} and defined $csrf->{'token'};
}
# make any necessary string manipulations to play nice with markdown
sub format_content {
my ( $content, %params ) = @_;
$content =~ s/\n/\n\n/g;
$content .= "\nposted by [pod_feeder](https://github.com/rev138/pod_feeder)" unless( $params{'no_branding'} );
return $content;
}
# post a message
sub post_message {
my ( $ua, $base_url, $content, $aspect_ids, %params ) = @_;
my ( $get_stream, $result ) = get_page( $ua, "$base_url/stream" );
if( $get_stream ){
my $csrf = extract_token( $result );
my $post_url = "$base_url/status_messages";
my $message = { status_message => { text => format_content( $content, %params ), provider_display_name => $params{'via'} }, aspect_ids => $aspect_ids };
my $json = JSON->new->allow_nonref;
$json = $json->utf8(0) unless utf8::is_utf8( $message );
my $json_message = $json->encode( $message );
return $ua->post( $post_url, 'Content' => $json_message, 'Content-Type' => 'application/json; charset=UTF-8', 'X-CSRF-Token' => $csrf->{'token'} );
}
}
# create a new sqlite db file with a 'feeds' table if it does not exist already
sub init_database {
my ( $db_file ) = @_;
my $dbh = connect_to_db( $db_file );
unless( -e $db_file ){
my $sth = $dbh->prepare(
'CREATE TABLE feeds(guid VARCHAR(255) PRIMARY KEY,feed_id VARCHAR(127),title VARCHAR(255),link VARCHAR(255),image VARCHAR(255),image_title VARCHAR(255),hashtags VARCHAR(255),timestamp INTEGER(10),posted INTEGER(1),body VARCHAR(10000))'
) or die "Can't prepare statement: $DBI::errstr";
$sth->execute() or die "Can't execute statement: $DBI::errstr";
$dbh->disconnect();
}
else {
my $sth = $dbh->column_info(undef, undef, 'feeds', undef);
my $body_exists = 0;
while( my( $tcat, $tscheme, $tname, $column_name ) = $sth->fetchrow_array() ) {
$body_exists = 1 if $column_name eq 'body';
}
unless( $body_exists ) {
$sth = $dbh->prepare('ALTER TABLE feeds ADD body VARCHAR(10000)');
$sth->execute() or die "Can't execute statement: $DBI::errstr";
}
$dbh->disconnect();
}
}
sub usage {
print "$0\n";
print "usage:\n";
print " -a --aspect-id <id> Aspects to share with. May specify multiple times (default: 'public')\n";
print " -b --embed-image Embed an image in the post if a link exists (default: off)\n";
print " -c --category-tags Attempt to automatically hashtagify RSS item 'categories' (default: off)\n";
print " -d --database <sqlite file> The SQLite file to store feed data (default: 'feed.db')\n";
print " -e --title-tags Automatically hashtagify RSS item title\n";
print " -f --feed-url <http://...> The feed URL\n";
print " -g --user-agent <string> Use this to spoof the user-agent if the feed blocks bots (ex: 'Mozilla/5.0')\n";
print " -i --feed-id <string> An arbitrary identifier to associate database entries with this feed\n";
print " -j --no-branding Do not include 'posted via pod_feeder' footer to posts\n";
print " -l --pod-url <https://...> The pod URL\n";
print " -m --timeout <hours> How long (in hours) to keep attempting failed posts (default 72)\n";
print " -n --ignore-tag <#hashtag> Hashtags to filter out. May be specified multiple times (default: none)\n";
print " -o --fetch-only Don't publish to Diaspora, just queue the new feed items for later\n";
print " -p --password <********> The D* user password\n";
print " -r --url-tags Attempt to automatically hashtagify the RSS link URL (default: off)\n";
print " -t --auto-tag <#hashtag> Hashtags to add to all posts. May be specified multiple times (default: none)\n";
print " -s --insecure Allows the option to bypass any errors caused from self-signed certificates(default: off)\n";
print " -u --username <user> The D* login username\n";
print " -v --via <string> Sets the 'posted via' text (default: 'pod_feeder')\n";
print " -w --post-raw-link Post the raw link instead of hyperlinking the article title (default: off)\n";
print " -x --limit <n> Only post n items per script run, to prevent post-spamming (default: no limit)\n";
print " --body Post the body of the feed (description or content:encoded item)\n";
print "\n";
exit;
}