1 # Copyright (C) 2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # per-mitem/smsg iterators for search results
5 # "ovv" => "Overview viewer"
6 package PublicInbox::LeiOverview;
9 use parent qw(PublicInbox::Lock);
10 use POSIX qw(strftime);
11 use Fcntl qw(F_GETFL O_APPEND);
14 use PublicInbox::MID qw($MID_EXTRACT);
15 use PublicInbox::Address qw(pairs);
16 use PublicInbox::Config;
17 use PublicInbox::Search qw(get_pct);
18 use PublicInbox::LeiDedupe;
19 use PublicInbox::LeiToMail;
21 # cf. https://en.wikipedia.org/wiki/JSON_streaming
22 my $JSONL = 'ldjson|ndjson|jsonl'; # 3 names for the same thing
24 sub _iso8601 ($) { strftime('%Y-%m-%dT%H:%M:%SZ', gmtime($_[0])) }
26 # we open this in the parent process before ->wq_do handoff
27 sub ovv_out_lk_init ($) {
29 $self->{tmp_lk_id} = "$self.$$";
30 my $tmp = File::Temp->new("lei-ovv.dst.$$.lock-XXXXXX",
31 TMPDIR => 1, UNLINK => 0);
32 $self->{lock_path} = $tmp->filename;
35 sub ovv_out_lk_cancel ($) {
37 ($self->{tmp_lk_id}//'') eq "$self.$$" and
38 unlink(delete($self->{lock_path}));
43 if ($dst =~ m!\A([:/]+://)!) {
44 $lei->fail("$1 support not implemented, yet\n");
45 } elsif (!-e $dst || -d _) {
46 'maildir'; # the default TODO: MH?
47 } elsif (-f _ || -p _) {
48 $lei->fail("unable to determine mbox family of $dst\n");
50 $lei->fail("unable to determine format of $dst\n");
55 my ($class, $lei) = @_;
56 my $opt = $lei->{opt};
57 my $dst = $opt->{output} // '-';
58 $dst = '/dev/stdout' if $dst eq '-';
60 my $fmt = $opt->{'format'};
61 $fmt = lc($fmt) if defined $fmt;
62 if ($dst =~ s/\A([a-z0-9]+)://is) { # e.g. Maildir:/home/user/Mail/
65 return $lei->fail(<<"") if $fmt ne $ofmt;
66 --format=$fmt and --output=$ofmt conflict
69 $fmt //= 'json' if $dst eq '/dev/stdout';
70 $fmt //= detect_fmt($lei, $dst) or return;
72 if (index($dst, '://') < 0) { # not a URL, so assume path
73 $dst = File::Spec->canonpath($dst);
76 my $self = bless { fmt => $fmt, dst => $dst }, $class;
79 if ($fmt =~ /\A($JSONL|(?:concat)?json)\z/) {
80 $json = $self->{json} = ref(PublicInbox::Config->json);
82 my ($isatty, $seekable);
83 if ($dst eq '/dev/stdout') {
84 $isatty = -t $lei->{1};
85 $lei->start_pager if $isatty;
86 $opt->{pretty} //= $isatty;
87 if (!$isatty && -f _) {
88 my $fl = fcntl($lei->{1}, F_GETFL, 0) //
89 return $lei->fail("fcntl(stdout): $!");
90 ovv_out_lk_init($self) unless ($fl & O_APPEND);
92 ovv_out_lk_init($self);
96 $lei->{dedupe} //= PublicInbox::LeiDedupe->new($lei);
98 # default to the cheapest sort since MUA usually resorts
99 $lei->{opt}->{'sort'} //= 'docid' if $dst ne '/dev/stdout';
100 $lei->{l2m} = eval { PublicInbox::LeiToMail->new($lei) };
101 return $lei->fail($@) if $@;
106 # called once by parent
108 my ($self, $lei) = @_;
109 if ($self->{fmt} eq 'json') {
110 print { $lei->{1} } '[';
111 } # TODO HTML/Atom/...
114 # called once by parent (via PublicInbox::EOFpipe)
116 my ($self, $lei) = @_;
117 my $out = $lei->{1} or return;
118 if ($self->{fmt} eq 'json') {
119 # JSON doesn't allow trailing commas, and preventing
120 # trailing commas is a PITA when parallelizing outputs
121 print $out "null]\n";
122 } elsif ($self->{fmt} eq 'concatjson') {
127 sub ovv_atfork_child {
132 # prepares an smsg for JSON
134 my ($smsg, $mitem) = @_;
136 delete @$smsg{qw(lines bytes num tid)};
137 $smsg->{rt} = _iso8601(delete $smsg->{ts}); # JMAP receivedAt
138 $smsg->{dt} = _iso8601(delete $smsg->{ds}); # JMAP UTCDate
139 $smsg->{pct} = get_pct($mitem) if $mitem;
140 if (my $r = delete $smsg->{references}) {
141 $smsg->{refs} = [ map { "<$_>" } ($r =~ m/$MID_EXTRACT/go) ];
143 if (my $m = delete($smsg->{mid})) {
144 $smsg->{'m'} = "<$m>";
146 for my $f (qw(from to cc)) {
147 my $v = delete $smsg->{$f} or next;
148 $smsg->{substr($f, 0, 1)} = pairs($v);
150 $smsg->{'s'} = delete $smsg->{subject};
151 # can we be bothered to parse From/To/Cc into arrays?
152 scalar { %$smsg }; # unbless
155 sub ovv_atexit_child {
156 my ($self, $lei) = @_;
157 if (my $l2m = delete $lei->{l2m}) {
158 # gracefully stop lei2mail processes after all
159 # ->write_mail work is complete
160 delete $l2m->{-wq_s1};
161 if (my $rd = delete $l2m->{each_smsg_done}) {
162 read($rd, my $buf, 1); # wait for EOF
165 # order matters, git->{-tmp}->DESTROY must not fire until
166 # {each_smsg_done} hits EOF above
167 if (my $git = delete $self->{git}) {
168 $git->async_wait_all;
170 if (my $bref = delete $lei->{ovv_buf}) {
171 my $out = $lei->{1} or return;
172 my $lk = $self->lock_for_scope;
177 # JSON module ->pretty output wastes too much vertical white space,
178 # this (IMHO) provides better use of screen real-estate while not
179 # being excessively compact:
181 my ($json, $k, $v) = @_;
182 if (ref $v eq 'ARRAY') {
184 my $sep = ",\n" . (' ' x (length($k) + 7));
185 if (ref($v->[0])) { # f/t/c
186 $v = '[' . join($sep, map {
187 my $pair = $json->encode($_);
188 $pair =~ s/(null|"),"/$1, "/g;
191 } else { # references
192 $v = '[' . join($sep, map {
193 substr($json->encode([$_]), 1, -1);
203 sub ovv_each_smsg_cb { # runs in wq worker usually
204 my ($self, $lei, $ibxish) = @_;
206 $lei->{1}->autoflush(1);
207 if (my $pkg = $self->{json}) {
209 $json->utf8->canonical;
210 $json->ascii(1) if $lei->{opt}->{ascii};
212 my $l2m = $lei->{l2m};
214 $dedupe = $lei->{dedupe} // die 'BUG: {dedupe} missing';
215 $dedupe->prepare_dedupe;
217 $lei->{ovv_buf} = \(my $buf = '') if !$l2m;
218 if ($l2m && !$ibxish) { # remote https?:// mboxrd
219 delete $l2m->{-wq_s1};
220 my $g2m = $l2m->can('git_to_mail');
221 my $wcb = $l2m->write_cb($lei);
223 my ($smsg, undef, $eml) = @_; # no mitem in $_[1]
224 $wcb->(undef, $smsg, $eml);
226 } elsif ($l2m && $l2m->{-wq_s1}) {
227 my ($lei_ipc, @io) = $lei->atfork_parent_wq($l2m);
228 # $io[-1] becomes a notification pipe that triggers EOF
229 # in this wq worker when all outstanding ->write_mail
231 pipe($l2m->{each_smsg_done}, $io[$#io + 1]) or die "pipe: $!";
232 fcntl($io[-1], 1031, 4096) if $^O eq 'linux'; # F_SETPIPE_SZ
233 delete @$lei_ipc{qw(l2m opt mset_opt cmd)};
234 $lei_ipc->{each_smsg_not_done} = $#io;
235 my $git = $ibxish->git; # (LeiXSearch|Inbox|ExtSearch)->git
237 my $git_dir = $git->{git_dir};
239 my ($smsg, $mitem) = @_;
240 $smsg->{pct} = get_pct($mitem) if $mitem;
241 $l2m->wq_do('write_mail', \@io, $git_dir, $smsg,
245 my $wcb = $l2m->write_cb($lei);
246 my $git = $ibxish->git; # (LeiXSearch|Inbox|ExtSearch)->git
247 $self->{git} = $git; # for ovv_atexit_child
248 my $g2m = $l2m->can('git_to_mail');
250 my ($smsg, $mitem) = @_;
251 $smsg->{pct} = get_pct($mitem) if $mitem;
252 $git->cat_async($smsg->{blob}, $g2m, [ $wcb, $smsg ]);
254 } elsif ($self->{fmt} =~ /\A(concat)?json\z/ && $lei->{opt}->{pretty}) {
255 my $EOR = ($1//'') eq 'concat' ? "\n}" : "\n},";
256 sub { # DIY prettiness :P
257 my ($smsg, $mitem) = @_;
258 return if $dedupe->is_smsg_dup($smsg);
259 $smsg = _unbless_smsg($smsg, $mitem);
261 $buf .= join(",\n", map {
264 _json_pretty($json, $_, $v);
266 $v = $json->encode([$v]);
267 qq{ "$_": }.substr($v, 1, -1);
271 if (length($buf) > 65536) {
272 my $lk = $self->lock_for_scope;
273 print { $lei->{1} } $buf;
278 my $ORS = $self->{fmt} eq 'json' ? ",\n" : "\n"; # JSONL
280 my ($smsg, $mitem) = @_;
281 return if $dedupe->is_smsg_dup($smsg);
282 $buf .= $json->encode(_unbless_smsg(@_)) . $ORS;
283 if (length($buf) > 65536) {
284 my $lk = $self->lock_for_scope;
285 print { $lei->{1} } $buf;
293 *DESTROY = \&ovv_out_lk_cancel;