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 # handles "lei q" command and provides internals for
5 # several other sub-commands (up, lcat, ...)
6 package PublicInbox::LeiQuery;
10 sub prep_ext { # externals_each callback
11 my ($lxs, $exclude, $loc) = @_;
12 $lxs->prepare_external($loc) unless $exclude->{$loc};
15 sub _start_query { # used by "lei q" and "lei up"
17 require PublicInbox::LeiOverview;
18 PublicInbox::LeiOverview->new($self) or return;
19 my $opt = $self->{opt};
20 my ($xj, $mj) = split(/,/, $opt->{jobs} // '');
21 (defined($xj) && $xj ne '' && $xj !~ /\A[1-9][0-9]*\z/) and
22 die "`$xj' search jobs must be >= 1\n";
23 my $lxs = $self->{lxs};
24 $xj ||= $lxs->concurrency($opt); # allow: "--jobs ,$WRITER_ONLY"
25 my $nproc = $lxs->detect_nproc || 1; # don't memoize, schedtool(1) exists
26 $xj = $nproc if $xj > $nproc;
27 $lxs->{-wq_nr_workers} = $xj;
28 (defined($mj) && $mj !~ /\A[1-9][0-9]*\z/) and
29 die "`$mj' writer jobs must be >= 1\n";
30 my $l2m = $self->{l2m};
31 # we use \1 (a ref) to distinguish between default vs. user-supplied
32 if ($l2m && grep { $opt->{$_} //= \1 } (qw(mail-sync import-remote
34 $self->_lei_store(1)->write_prepare($self);
36 $l2m and $l2m->{-wq_nr_workers} = $mj // do {
37 $mj = int($nproc * 0.75 + 0.5); # keep some CPU for git
40 # descending docid order is cheapest, MUA controls sorting order
41 $self->{mset_opt}->{relevance} //= -2 if $l2m || $opt->{threads};
43 my $tot = $self->{mset_opt}->{total} //= $self->{opt}->{limit} // 10000;
44 $self->{mset_opt}->{limit} = $tot > 10000 ? 10000 : $tot;
45 $self->{mset_opt}->{offset} //= 0;
46 $self->{mset_opt}->{threads} //= $opt->{threads};
49 require PublicInbox::LeiAuth;
50 $self->{auth} = PublicInbox::LeiAuth->new
52 $lxs->do_query($self);
55 sub qstr_add { # PublicInbox::InputPipe::consume callback for --stdin
56 my ($lei) = @_; # $_[1] = $rbuf
57 $_[1] // $lei->fail("error reading stdin: $!");
58 return $lei->{mset_opt}->{qstr} .= $_[1] if $_[1] ne '';
61 $lei->{mset_opt}->{q_raw} = $lei->{mset_opt}->{qstr};
62 $lei->{lse}->query_approxidate($lei->{lse}->git,
63 $lei->{mset_opt}->{qstr});
71 require PublicInbox::LeiXSearch;
72 # prepare any number of LeiXSearch || LeiSearch || Inbox || URL
73 my $lxs = $self->{lxs} = PublicInbox::LeiXSearch->new;
74 my $opt = $self->{opt};
75 my @only = @{$opt->{only} // []};
76 # --local is enabled by default unless --only is used
77 # we'll allow "--only $LOCATION --local"
78 my $sto = $self->_lei_store(1);
79 $self->{lse} = $sto->search;
80 if ($opt->{'local'} //= scalar(@only) ? 0 : 1) {
81 $lxs->prepare_external($self->{lse});
85 my @loc = $self->get_externals($loc) or return;
86 $lxs->prepare_external($_) for @loc;
89 my (@ilocals, @iremotes);
90 for my $loc (@{$opt->{include} // []}) {
91 my @loc = $self->get_externals($loc) or return;
92 $lxs->prepare_external($_) for @loc;
93 @ilocals = @{$lxs->{locals} // []};
94 @iremotes = @{$lxs->{remotes} // []};
96 # --external is enabled by default, but allow --no-external
97 if ($opt->{external} //= 1) {
98 my $ex = $self->canonicalize_excludes($opt->{exclude});
99 $self->externals_each(\&prep_ext, $lxs, $ex);
100 $opt->{remote} //= !($lxs->locals - $opt->{'local'});
101 $lxs->{locals} = \@ilocals if !$opt->{'local'};
102 $lxs->{remotes} = \@iremotes if !$opt->{remote};
105 ($lxs->locals || $lxs->remotes) ? ($self->{lxs} = $lxs) :
106 die("no local or remote inboxes to search\n");
109 # the main "lei q SEARCH_TERMS" method
111 my ($self, @argv) = @_;
112 PublicInbox::Config->json; # preload before forking
113 my $lxs = lxs_prepare($self) or return;
114 $self->ale->refresh_externals($lxs, $self);
115 my $opt = $self->{opt};
117 $mset_opt{asc} = $opt->{'reverse'} ? 1 : 0;
118 if (defined(my $sort = $opt->{'sort'})) {
119 if ($sort eq 'relevance') {
120 $mset_opt{relevance} = 1;
121 } elsif ($sort eq 'docid') {
122 $mset_opt{relevance} = $mset_opt{asc} ? -1 : -2;
123 } elsif ($sort =~ /\Areceived(?:-?[aA]t)?\z/) {
126 die "unrecognized --sort=$sort\n";
128 $opt->{save} and return
129 $self->fail('--save and --sort are incompatible');
131 $self->{mset_opt} = \%mset_opt;
134 return $self->fail(<<'') if @argv;
135 no query allowed on command-line with --stdin
137 require PublicInbox::InputPipe;
138 PublicInbox::InputPipe::consume($self->{0}, \&qstr_add, $self);
141 $mset_opt{q_raw} = [ @argv ]; # copy
143 $self->{lse}->query_argv_to_string($self->{lse}->git, \@argv);
147 # shell completion helper called by lei__complete
149 my ($self, @argv) = @_;
151 my $cb = $self->lazy_cb(qw(forget-external _complete_));
153 if ($argv[-1] =~ /\A(?:-I|(?:--(?:include|exclude|only)))\z/) {
154 my @c = $cb->($self, @cur);
155 # try basename match:
156 if (scalar(@cur) == 1 && index($cur[0], '/') < 0) {
157 my $all = $self->externals_each;
159 for my $loc (keys %$all) {
160 my $bn = (split(m!/!, $loc))[-1];
164 $bn{$_} == 1 && /\A\Q$cur[0]/
169 unshift(@cur, pop @argv);
174 # Stuff we may pass through to curl (as of 7.64.0), see curl manpage for
175 # details, so most options which make sense for HTTP/HTTPS (including proxy
176 # support for Tor and other methods of getting past weird networks).
177 # Most of these are untested by us, some may not make sense for our use case
178 # and typos below are likely.
179 # n.b. some short options (-$NUMBER) are not supported since they conflict
180 # with other "lei q" switches.
181 # FIXME: Getopt::Long doesn't easily let us support support options with
182 # '.' in them (e.g. --http1.1)
183 # TODO: should we depend on "-c http.*" options for things which have
184 # analogues in git(1)? that would reduce likelyhood of conflicts with
185 # our other CLI options
186 # Note: some names are renamed to avoid potential conflicts,
187 # see %lei2curl in lib/PublicInbox/LeiCurl.pm
190 abstract-unix-socket=s anyauth basic cacert=s capath=s
191 cert-status cert-type cert=s ciphers=s
192 connect-timeout=s connect-to=s cookie-jar=s cookie=s crlfile=s
193 digest disable dns-interface=s dns-ipv4-addr=s dns-ipv6-addr=s
194 dns-servers=s doh-url=s egd-file=s engine=s false-start
195 happy-eyeballs-timeout-ms=s haproxy-protocol header=s@
196 http2-prior-knowledge http2 insecure
197 interface=s ipv4 ipv6 junk-session-cookies
198 key-type=s key=s limit-rate=s local-port=s location-trusted location
199 max-redirs=i max-time=s negotiate netrc-file=s netrc-optional netrc
200 no-alpn no-buffer no-npn no-sessionid noproxy=s ntlm-wb ntlm
201 pass=s pinnedpubkey=s post301 post302 post303 preproxy=s
202 proxy-anyauth proxy-basic proxy-cacert=s proxy-capath=s
203 proxy-cert-type=s proxy-cert=s proxy-ciphers=s proxy-crlfile=s
204 proxy-digest proxy-header=s@ proxy-insecure
205 proxy-key-type=s proxy-key proxy-negotiate proxy-ntlm proxy-pass=s
206 proxy-pinnedpubkey=s proxy-service-name=s proxy-ssl-allow-beast
207 proxy-tls13-ciphers=s proxy-tlsauthtype=s proxy-tlspassword=s
208 proxy-tlsuser=s proxy-tlsv1 proxy-user=s proxy=s
209 proxytunnel=s pubkey=s random-file=s referer=s resolve=s
210 retry-connrefused retry-delay=s retry-max-time=s retry=i
211 sasl-ir service-name=s socks4=s socks4a=s socks5-basic
212 socks5-gssapi-service-name=s socks5-gssapi socks5-hostname=s socks5=s
213 speed-limit speed-type ssl-allow-beast sslv2 sslv3
214 suppress-connect-headers tcp-fastopen tls-max=s
215 tls13-ciphers=s tlsauthtype=s tlspassword=s tlsuser=s
216 tlsv1 trace-ascii=s trace-time trace=s
217 unix-socket=s user-agent=s user=s