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') {
111 } # TODO HTML/Atom/...
114 # called once by parent (via PublicInbox::EOFpipe)
116 my ($self, $lei) = @_;
117 if ($self->{fmt} eq 'json') {
118 # JSON doesn't allow trailing commas, and preventing
119 # trailing commas is a PITA when parallelizing outputs
120 $lei->out("null]\n");
121 } elsif ($self->{fmt} eq 'concatjson') {
126 # prepares an smsg for JSON
128 my ($smsg, $mitem) = @_;
130 delete @$smsg{qw(lines bytes num tid)};
131 $smsg->{rt} = _iso8601(delete $smsg->{ts}); # JMAP receivedAt
132 $smsg->{dt} = _iso8601(delete $smsg->{ds}); # JMAP UTCDate
133 $smsg->{pct} = get_pct($mitem) if $mitem;
134 if (my $r = delete $smsg->{references}) {
135 $smsg->{refs} = [ map { "<$_>" } ($r =~ m/$MID_EXTRACT/go) ];
137 if (my $m = delete($smsg->{mid})) {
138 $smsg->{'m'} = "<$m>";
140 for my $f (qw(from to cc)) {
141 my $v = delete $smsg->{$f} or next;
142 $smsg->{substr($f, 0, 1)} = pairs($v);
144 $smsg->{'s'} = delete $smsg->{subject};
145 # can we be bothered to parse From/To/Cc into arrays?
146 scalar { %$smsg }; # unbless
149 sub ovv_atexit_child {
150 my ($self, $lei) = @_;
151 if (my $l2m = delete $lei->{l2m}) {
152 # gracefully stop lei2mail processes after all
153 # ->write_mail work is complete
154 delete $l2m->{-wq_s1};
155 if (my $rd = delete $l2m->{each_smsg_done}) {
156 read($rd, my $buf, 1); # wait for EOF
159 # order matters, git->{-tmp}->DESTROY must not fire until
160 # {each_smsg_done} hits EOF above
161 if (my $git = delete $self->{git}) {
162 $git->async_wait_all;
164 if (my $bref = delete $lei->{ovv_buf}) {
165 my $lk = $self->lock_for_scope;
170 # JSON module ->pretty output wastes too much vertical white space,
171 # this (IMHO) provides better use of screen real-estate while not
172 # being excessively compact:
174 my ($json, $k, $v) = @_;
175 if (ref $v eq 'ARRAY') {
177 my $sep = ",\n" . (' ' x (length($k) + 7));
178 if (ref($v->[0])) { # f/t/c
179 $v = '[' . join($sep, map {
180 my $pair = $json->encode($_);
181 $pair =~ s/(null|"),"/$1, "/g;
184 } else { # references
185 $v = '[' . join($sep, map {
186 substr($json->encode([$_]), 1, -1);
196 sub ovv_each_smsg_cb { # runs in wq worker usually
197 my ($self, $lei, $ibxish) = @_;
199 $lei->{1}->autoflush(1);
200 if (my $pkg = $self->{json}) {
202 $json->utf8->canonical;
203 $json->ascii(1) if $lei->{opt}->{ascii};
205 my $l2m = $lei->{l2m};
207 $dedupe = $lei->{dedupe} // die 'BUG: {dedupe} missing';
208 $dedupe->prepare_dedupe;
210 $lei->{ovv_buf} = \(my $buf = '') if !$l2m;
211 if ($l2m && !$ibxish) { # remote https?:// mboxrd
212 delete $l2m->{-wq_s1};
213 my $g2m = $l2m->can('git_to_mail');
214 my $wcb = $l2m->write_cb($lei);
216 my ($smsg, undef, $eml) = @_; # no mitem in $_[1]
217 $wcb->(undef, $smsg, $eml);
219 } elsif ($l2m && $l2m->{-wq_s1}) {
220 my ($lei_ipc, @io) = $lei->atfork_parent_wq($l2m);
221 # $io[-1] becomes a notification pipe that triggers EOF
222 # in this wq worker when all outstanding ->write_mail
224 pipe($l2m->{each_smsg_done}, $io[$#io + 1]) or die "pipe: $!";
225 fcntl($io[-1], 1031, 4096) if $^O eq 'linux'; # F_SETPIPE_SZ
226 delete @$lei_ipc{qw(l2m opt mset_opt cmd)};
227 $lei_ipc->{each_smsg_not_done} = $#io;
228 my $git = $ibxish->git; # (LeiXSearch|Inbox|ExtSearch)->git
230 my $git_dir = $git->{git_dir};
232 my ($smsg, $mitem) = @_;
233 $smsg->{pct} = get_pct($mitem) if $mitem;
234 $l2m->wq_do('write_mail', \@io, $git_dir, $smsg,
238 my $wcb = $l2m->write_cb($lei);
239 my $git = $ibxish->git; # (LeiXSearch|Inbox|ExtSearch)->git
240 $self->{git} = $git; # for ovv_atexit_child
241 my $g2m = $l2m->can('git_to_mail');
243 my ($smsg, $mitem) = @_;
244 $smsg->{pct} = get_pct($mitem) if $mitem;
245 $git->cat_async($smsg->{blob}, $g2m, [ $wcb, $smsg ]);
247 } elsif ($self->{fmt} =~ /\A(concat)?json\z/ && $lei->{opt}->{pretty}) {
248 my $EOR = ($1//'') eq 'concat' ? "\n}" : "\n},";
249 sub { # DIY prettiness :P
250 my ($smsg, $mitem) = @_;
251 return if $dedupe->is_smsg_dup($smsg);
252 $smsg = _unbless_smsg($smsg, $mitem);
254 $buf .= join(",\n", map {
257 _json_pretty($json, $_, $v);
259 $v = $json->encode([$v]);
260 qq{ "$_": }.substr($v, 1, -1);
264 return if length($buf) < 65536;
265 my $lk = $self->lock_for_scope;
270 my $ORS = $self->{fmt} eq 'json' ? ",\n" : "\n"; # JSONL
272 my ($smsg, $mitem) = @_;
273 return if $dedupe->is_smsg_dup($smsg);
274 $buf .= $json->encode(_unbless_smsg(@_)) . $ORS;
275 return if length($buf) < 65536;
276 my $lk = $self->lock_for_scope;
284 *DESTROY = \&ovv_out_lk_cancel;