view release on metacpan or  search on metacpan

t/31_cqp_dickens.t.dont_run  view on Meta::CPAN

# Before `make install' is performed this script should be runnable with
# `make test'. After `make install' it should work as `perl test.pl'

BEGIN { $| = 1; print "Running tests for CWB/CQP.pm:\n"; }

# step 1: load the CQP/Perl library
END {
  unlink glob "tmp/DICKENS:*";  # clean up temporary corpus files
  if (not $loaded) {
    print "!! can't load CWB::CQP module (aborted)\n";
    exit 1;
  }
  if (not $deleted) {
    print "!! error deleting CWB::CQP object\n";
    exit 1;
  }
}

use CWB::CQP;
$loaded = 1;
print " - use CWB::CQP ok\n";

# step 2: start CQP process with our own test init file
$CQP = new CWB::CQP "-I data/files/init.cqp";
if (not defined $CQP) {
  print "!! can't start CQP process\n";
  exit 1;
}
print " - spawned CQP child process\n";

# step 3: we need an up-to-date version of CQP
$ok = $CQP->check_version(2, 2, 98);		# that´s 2.2.b98 or newer
if (not $ok) {
  print "!! CQP version too old, consider upgrading\n";
  exit 1;
}
$major = $CQP->{'major_version'};
$minor = $CQP->{'minor_version'};
$beta =  $CQP->{'beta_version'};
$date =  $CQP->{'compile_date'};
print " - CQP version $major.$minor";
print "b$beta"
  if $beta > 0;
print " (compile date: $date)\n";

# step 4: list available corpora (run CQP manually with -c to check output of interactive commands!)
@corpora = $CQP->exec("show corpora");
if ((not $CQP->ok) or (@corpora == 0) or (not grep {$_ eq "DICKENS"} @corpora)) {
  print "!! error in 'show corpora;' command\n";
  exit 1;
}
$N = @corpora;
print " - $N corpora found, including DICKENS\n";

# step 5: activate corpus
$CQP->exec("DICKENS");
if (not $CQP->ok) {
  print "!! can't activate corpus DICKENS\n";
  exit 1;
}
print " - corpus DICKENS activated\n";

# step 6: we need a little trick to get at the corpus size
@matches = $CQP->dump("DICKENS");		# don´t ask ...
if (@matches != 1) {
  print "!! error in 'dump DICKENS;' command\n";
  exit 1;
}

t/31_cqp_dickens.t.dont_run  view on Meta::CPAN

# step 7: try an error condition, but don´t let the default error handler mess up the screen
$CQP->set_error_handler('ignore');
$CQP->exec("[lemma == ']");
if ($CQP->ok) {
  print "!! query syntax error not detected\n";
  exit 1;
}
@msg = $CQP->error_message;
if (not ((@msg >= 2) and grep {/<--/} @msg)) {	# the <-- is used by CQP to indicate the location of a syntax error
  print "!! incomplete error message\n";
  exit 1;
}
$N = @msg;
print " - syntax error detected ($N-line error message)\n";

# step 8: try a different error condition and check whether our custom error handler is really called
$handler_ok = 0;
$CQP->set_error_handler(sub { if (@_ >= 2) { $handler_ok = @_ } } );
$CQP->exec("'('");
if (not $handler_ok) {
  print "!! custom error handler was not invoked\n";
  exit 1;
}
print " - syntax error handled ($handler_ok-line error message)\n";

# step 8b: try a query lock evaluation with an interactive command
$handler_ok = 0;
@nothing = $CQP->exec_query("show cd");
if (not ($handler_ok and @nothing == 0)) {
  print "!! query lock violation was not detected\n";
  exit 1;
}
print " - query lock violation caught\n";

# set stricter handler (any further CQP errors are _real_ errors ...)
$CQP->set_error_handler('die');

# step 9: check multi-line commands
@nothing = $CQP->exec(";;;;;");			# five blank lines as end-of-command markers are swallowed by CQP.pm
if (@nothing > 0) {
  print "!! unexpected output from empty comands\n";
  exit 1;
}
print " - command ';;;;;' ok, no output\n";

# step 10: repeatedly run a macro query defined in init.cqp and save the results as named queries
print " - executing macro ";
for ($i = 0; $i < 8; $i++) {
  print ".";
  $CQP->exec("A$i = /number_of_NPs[$i]");
}
print " ok\n";

# step 11: save results (named queries) to disk
print " - saving named queries ";
for ($i = 0; $i < 8; $i++) {
  print ".";
  $CQP->exec("save A$i");
}
print " ok\n";
@saved = glob "tmp/DICKENS:A*";
if (@saved != 8) {
  print "!! can't find saved query results on disk\n";
  exit 1;
}

# step 12: exit and re-start CQP
undef $CQP;
$deleted = 1;
print " - CQP process terminated\n";
$CQP = new CWB::CQP ("-I data/files/init.cqp");
if (not defined $CQP) {
  print "!! can't re-start CQP process\n";
  exit 1;
}
print " - new CQP child process spawned\n";

$CQP->set_error_handler('die'); # again, we're strict about possible errors

# step 13: named queries from step 10 should now be visible within CQP
@named = $CQP->exec("show named");
if ((grep {/-d-\s+DICKENS:A/} @named) != 8) {
  print "!! wrong output from 'show named;' command\n";
  exit 1;
}
print " - saved query results found\n";

# step 14: re-load query results and get number of matches
@size = ();
$total_size = 0;				# should add up to number of sentences in DICKENS
print " - loading query results ";
for ($i = 0; $i < 8; $i++) {
  print ".";
  ($s) = $CQP->exec("size A$i");		# "size" command returns just one number 
  if (not (defined $s and $s >= 0)) {
    print "\n!! load error\n";
    exit 1;
  }
  push @size, $s;
  $total_size += $s;
}
print " ok\n";
if (not ($size[5] > 0 and $size[7] == 0)) {
  print "!! wrong number of matches\n";
  exit 1;
}
print " - number of matches: ", join(", ", @size), "\n";

# step 15: check $total_size against simple query counting all sentences
print " - counting sentences ... ";
($size) = $CQP->exec("DICKENS; A = <s>[]; size A;"); # can execute multiple command with single call
print "$size found\n";
if ($total_size != $size) {
  print "!! inconsistency detected ($size sentences vs. $total_size matches to subcategories)\n";
  exit 1;
}

# step 16: cat the query results from A4 and A5 and check the number of lines
$CQP->exec("show -cpos");			# don't print cpos (check absence below)
@lines4 = $CQP->exec("cat A4");
@lines5 = $CQP->exec("cat DICKENS:A5");		# use both short and full form of named query