Man pages sections > man3 > Bio::DB::HIV::HIVQueryHelper

Bio::DB::HIV::HIVQueryHelper - Routines and packages used by Bio::DB::HIV and

Bio::DB::HIV::HIVQueryHelper(3pm) User Contributed Perl Documentation Bio::DB::HIV::HIVQueryHelper(3pm)

NAME

Bio::DB::HIV::HIVQueryHelper - Routines and packages used by Bio::DB::HIV and Bio::DB::Query::HIVQuery

SYNOPSIS

  Used in Bio::DB::Query::HIVQuery. No need to use directly.

DESCRIPTION

"Bio::DB::HIV::HIVQueryHelper" contains a number of packages for use by Bio::DB::Query::HIVQuery. Package "HIVSchema" parses the "lanl-schema.xml" file, and allows access to it in the context of the relational database it represents (see APPENDIX for excruciating detail). Packages "QRY", "R", and "Q" together create the query string parser that enables NCBI-like queries to be understood by "Bio::DB::Query::HIVQuery". They provide objects and operators to perform and simplify logical expressions involving "AND", "OR", and "()" and return hash structures that can be handled by "Bio::DB::Query::HIVQuery" routines.

FEEDBACK

Mailing Lists

User feedback is an integral part of the evolution of this and other Bioperl modules. Send your comments and suggestions preferably to the Bioperl mailing list. Your participation is much appreciated.
  bioperl-l@bioperl.org                  - General discussion
  http://bioperl.org/wiki/Mailing_lists  - About the mailing lists

Support

Please direct usage questions or support issues to the mailing list:
bioperl-l@bioperl.org
rather than to the module maintainer directly. Many experienced and reponsive experts will be able look at the problem and quickly address it. Please include a thorough description of the problem with code and data examples if at all possible.

Reporting Bugs

Report bugs to the Bioperl bug tracking system to help us keep track of the bugs and their resolution. Bug reports can be submitted via the web:
  https://github.com/bioperl/bioperl-live/issues

AUTHOR - Mark A. Jensen

Email maj@fortinbras.us

CONTRIBUTORS

Mark A. Jensen

APPENDIX

The rest of the documentation details each of the contained packages. Internal methods are usually preceded with a _

HIVSchema - objects/methods to manipulate a version of the LANL HIV DB schema

HIVSchema SYNOPSIS
    $schema = new HIVSchema( 'lanl-schema.xml' );
    @tables = $schema->tables;
    @validFields = $schema->fields;
    @validAliases = $schema->aliases;
    @query_aliases_for_coreceptor = $schema->aliases( 'SEQ_SAMple.SSAM_second_receptor' );
    $pk_for_SequenceEntry = $schema->primarykey('SequenceEntry');    # returns 'SequenceEntry.SE_id'
    $fk_for_SEQ_SAMple_to_SequenceEntry =
              $schema->foreignkey('SEQ_SAMple', 'SequenceEntry');    # returns 'SEQ_SAMple.SSAM_SE_id'
    $table = $schema->tablepart('SEQ_SAMple.SSAM_badseq');           # returns 'SEQ_SAMple'
    $column = $schema->columnpart('SEQ_SAMple.SSAM_badseq');         # returns 'SSAM_badseq'
HIVSchema DESCRIPTION
HIVSchema methods are used in Bio::DB::Query::HIVQuery for table, column, primary/foreign key manipulations based on the observed Los Alamos HIV Sequence Database (LANL DB) naming conventions for their CGI parameters. The schema is contained in an XML file ("lanl-schema.xml") which is read into an HIVSchema object, in turn a property of the HIVQuery object. HIVSchema methods are used to build correct cgi queries in a way that attempts to preserve the context of the relational database the query parameters represent.
HIVSchema CONSTRUCTOR
HIVSchema::new
 Title   : new
 Usage   : $schema = new HIVSchema( "lanl-schema.xml ");
 Function:
 Example :
 Returns : an HIVSchema object
 Args    : XML filename
HIVSchema INSTANCE METHODS
HIVSchema tables
 Title   : tables
 Usage   : $schema->tables()
 Function: get all table names in schema
 Example :
 Returns : array of table names
 Args    : none
HIVSchema columns
 Title   : columns
 Usage   : $schema->columns( [$tablename] );
 Function: return array of columns for specified table, or all columns in
           schema, if called w/o args
 Example :
 Returns :
 Args    : tablename or fieldname string
HIVSchema fields
 Title   : fields
 Usage   : $schema->fields();
 Function: return array of all fields in schema, in format "table.column"
 Example :
 Returns : array of all fields
 Args    : none
HIVSchema options
 Title   : options
 Usage   : $schema->options(@fieldnames)
 Function: get array of options (i.e., valid match data strings) available
           to specified field
 Example :
 Returns : array of match data strings
 Args    : [array of] fieldname string[s] in "table.column" format
HIVSchema aliases
 Title   : aliases
 Usage   : $schema->aliases(@fieldnames)
 Function: get array of aliases to specified field[s]
 Example :
 Returns : array of valid query aliases for fields as spec'd in XML file
 Args    : [an array of] fieldname[s] in "table.column" format
HIVSchema ankh
 Title   : ankh (annotation key hash)
 Usage   : $schema->ankh(@fieldnames)
 Function: return a hash translating fields to annotation keys for the
           spec'd fields.
           (Annotation keys are used for parsing the tab-delimited response
           to Bio::DB::Query::HIVQuery::_do_lanl_request.)
 Example :
 Returns : hash ref
 Args    : [an array of] fieldname[s] in "table.column" format
HIVSchema tablepart
 Title   : tablepart (alias: tbl)
 Usage   : $schema->tbl(@fieldnames)
 Function: return the portion of the fieldname[s] that refer to the
           db table
 Example : $schema->tbl('SequenceEntry.SE_id'); # returns 'SequenceEntry'
 Returns : table name as string
 Args    : [an array of] fieldname[s] in "table.column" format
HIVSchema columnpart
 Title   : columnpart (alias: col)
 Usage   : $schema->col(@fieldnames)
 Function: return the portion of the fieldname[s] that refer to the
           db column
 Example : $schema->col('SequenceEntry.SE_id'); # returns 'SE_id'
 Returns : column name as string
 Args    : [an array of] fieldname[s] in "table.column" format
HIVSchema primarykey
 Title   : primarykey [alias: pk]
 Usage   : $schema->pk(@tablenames);
 Function: return the primary key of the specified table[s], as judged by
           the syntax of the table's[s'] fieldnames
 Example : $schema->pk('SequenceEntry') # returns 'SequenceEntry.SE_id'
 Returns : primary key fieldname[s] in "table.column" format, or null if
           no pk exists
 Args    : [an array of] table name[s] (fieldnames are ok, table part used)
HIVSchema foreignkey
 Title   : foreignkey [alias: fk]
 Usage   : $schema->fk($intable [, $totable])
 Function: return foreign key fieldname in table $intable referring to
           table $totable, or all foreign keys in $intable if $totable
           unspec'd
 Example : $schema->fk('AUthor', 'SequenceEntry'); # returns 'AUthor_AU_SE_id'
 Returns : foreign key fieldname[s] in "table.column" format
 Args    : tablename [, optional foreign table name] (fieldnames are ok,
           table part used)
HIVSchema foreigntable
 Title   : foreigntable [alias ftbl]
 Usage   : $schema->ftbl( @foreign_key_fieldnames );
 Function: return tablename of table that foreign keys points to
 Example : $schema->ftbl( 'AUthor.AU_SE_id' ); # returns 'SequenceEntry'
 Returns : tablename
 Args    : [an array of] fieldname[s] in "table.column" format
HIVSchema find_join
 Title   : find_join
 Usage   : $sch->find_join('Table1', 'Table2')
 Function: Retrieves a set of foreign and primary keys (in table.column
           format) that represents a join path from Table1 to Table2
 Example :
 Returns : an array of keys (as table.column strings) -or- an empty
           array if Table1 == Table2 -or- undef if no path exists
 Args    : two table names as strings
HIVSchema _find_join_guts
 Title   : _find_join_guts
 Usage   : $sch->_find_join_guts($table1, $table2, $stackref, \$found, $reverse)
           (call with $stackref = [], $found=0)
 Function: recursive guts of find_join
 Example :
 Returns : if a path is found, $found==1 and @$stackref contains the keys
           in table.column format representing the path; if a path is not
           found, $found == 0 and @$stackref contains garbage
 Args    : $table1, $table2 : table names as strings
           $stackref : an arrayref to an empty array
           \$found   : a scalar ref to the value 0
           $rev : if $rev==1, the arrays of table names will be reversed;
                  this can give a shorter path if cycles exist in the
                  schema graph
HIVSchema loadSchema
 Title   : loadHIVSchema [alias: loadSchema]
 Usage   : $schema->loadSchema( $XMLfilename )
 Function: read (LANL DB) schema spec from XML
 Example : $schema->loadSchema('lanl-schema.xml');
 Returns : hashref to schema data
           Keys are fieldnames in "table.column" format.
           Each value is a hashref with the following properties:
           {name}    : HIVWEB 'table.column' format fieldname,
                       can be used directly in the cgi query
           {aliases} : ref to array containing valid aliases/shortcuts for
                       {name}; can be used in routines creating the HTML query
           {options} : ref to array containing valid matchdata for this field
                       can be used directly in the HTML query
           {ankey}   : contains the annotation key for this field used with
                       Bioperl annotation objects
           {..attr..}: ..value_of_attr.. for this field (app-specific metadata)
 Args    :
HIVSchema _sfieldh
 Title   : _sfieldh
 Usage   : $schema->_sfieldh($fieldname)
 Function: get hashref to the specified field hash
 Example :
 Returns : hashref
 Args    : fieldname in "table.column" format

Class QRY - a query algebra for HIVQuery

QRY SYNOPSIS
  $Q = new QRY(
               new R(
                      new Q('coreceptor', 'CXCR4'),
                      new Q('country', 'ZA')
                     )
               );
  QRY::Eq(QRY::And($Q, $Q), $Q);                     # returns 1
  QRY::Eq(QRY::Or($Q, $Q), $Q);                      # returns 1
  $Q2 = $Q1->clone;
  $Q2 = new QRY(
                new R(
                       new Q( 'coreceptor', 'CCR5' ),
                       new Q( 'country', 'ZA')
                      )
               );
  (QRY::And($Q, $Q2))->isnull;                       # returns 1
  $Q3 = QRY::Or($Q, $Q2);
  print $Q3->A;                                      # prints '(CCR5 CXCR4)[coreceptor] (ZA)[country]'
QRY DESCRIPTION
The QRY package provides a query parser for Bio::DB::Query::HIVQuery. Currently, the parser supports AND, OR, and () operations. The structure of the LANL cgi makes it tricky to perform NOTs, though this could be implemented if the desire were great.
Two class methods do the work. "QRY::_parse_q" does a first-pass parse of the query string. "QRY::_make_q" interprets the parse tree as returned by "QRY::_parse_q" and produces an array of hash structures that can be used directly by "Bio::DB::Query::HIVQuery" query execution methods. Validation of query fields and options is performed at the "Bio::DB::Query::HIVQuery" level, not here.
"QRY" objects are collections of "R" (or request) objects, which are in turn collections of "Q" (or atomic query) objects. "Q" objects represent a query on a single field, with match data options "OR"ed together, e.g. "(A B)[subtype]". "R" objects collect "Q" objects that could be processed in a single HTTP request; i.e., a set of atomic queries each having different fields "AND"ed together, such as
  (A B)[subtype] AND ('CCR5')[coreceptor] AND (US CA)[country]
The "QRY" object collects "R"s that cannot be reduced (through logical operations) to a single HTTP request, e.g.
  ((C)[subtype] AND (SI)[phenotype]) OR ( (D)[subtype] AND (NSI)[phenotype] ),
which cannot be got in one go through the current LANL cgi implementation (as far as I can tell). The parser will simplify something like
  ((C)[subtype] AND (SI)[phenotype]) OR ((C)[subtype] AND (NSI)[phenotype])
to the single request
  (C)[subtype] AND (NSI SI)[phenotype]
however.
The operators "&" and "|" are overloaded to "QRY::And" and "QRY::Or", to get Perl precedence and grouping for free. "bool" is overloaded to get symbolic tests such as "if ($QRY) {stuff}". "==" is overloaded with "QRY::Eq" for convenience. No overloading is done for "R" or "Q".
QRY _make_q
 Title   : _make_q
 Usage   : QRY::_make_q($parsetree)
 Function: creates hash structures suitable for HIVQuery from parse tree
           returned by QRY::_parse_q
 Example :
 Returns : array of hashrefs of query specs
 Args    : a hashref
QRY _make_q_guts
 Title   : _make_q_guts (Internal class method)
 Usage   : _make_q_guts($ptree, $q_expr, $qarry, $anarry)
 Function: traverses the parse tree returned from QRY::_parse_q, checking
           syntax and creating HIVQuery-compliant query structures
 Example :
 Returns :
 Args    : $parse_tree (hashref), $query_expression (scalar string ref),
           $query_array (array ref : stack for returning query structures),
           $annotation_array (array ref : stack for returning annotation
           fields)
QRY _parse_q
 Title   : _parse_q
 Usage   : QRY::_parse_q($query_string)
 Function: perform first pass parse of a query string with some syntax
           checking, return a parse tree suitable for QRY::_make_q
 Example : QRY::_parse_q(" to[be] OR (not to)[be] ");
 Returns : hashref
 Args    : query string
QRY CONSTRUCTOR
QRY Constructor
 Title   : QRY constructor
 Usage   : $QRY = new QRY()
 Function:
 Example :
 Returns :
 Args    : array of R objects, optional
QRY INSTANCE METHODS
QRY requests
 Title   : requests
 Usage   : $QRY->requests
 Function: get/set array of requests comprising this QRY object
 Example :
 Returns :
 Args    : array of class R objects
QRY put_requests
 Title   : put_requests
 Usage   : $QRY->put_request(@R)
 Function: add object of class R to $QRY
 Example :
 Returns :
 Args    : [an array of] of class R object[s]
QRY isnull
 Title   : isnull
 Usage   : $QRY->isnull
 Function: test if QRY object is null
 Example :
 Returns : 1 if null, 0 otherwise
 Args    :
QRY A
 Title   : A
 Usage   : print $QRY->A
 Function: get a string representation of QRY object
 Example :
 Returns : string scalar
 Args    :
QRY len
 Title   : len
 Usage   : $QRY->len
 Function: get number of class R objects contained by QRY object
 Example :
 Returns : scalar
 Args    :
QRY clone
 Title   : clone
 Usage   : $QRY2 = $QRY1->clone;
 Function: create and return a clone of the object
 Example :
 Returns : object of class QRY
 Args    :
QRY CLASS METHODS
QRY Or
 Title   : Or
 Usage   : $QRY3 = QRY::Or($QRY1, $QRY2)
 Function: logical OR for QRY objects
 Example :
 Returns : a QRY object
 Args    : two class QRY objects
QRY And
 Title   : And
 Usage   : $QRY3 = QRY::And($QRY1, $QRY2)
 Function: logical AND for QRY objects
 Example :
 Returns : a QRY object
 Args    : two class QRY objects
QRY Bool
 Title   : Bool
 Usage   : QRY::Bool($QRY1)
 Function: allows symbolic testing of QRY object when bool overloaded
 Example : do {stuff} if $QRY1 *same as* do {stuff} if !$QRY1->isnull
 Returns :
 Args    : a class QRY object
QRY Eq
 Title   : Eq
 Usage   : QRY::Eq($QRY1, $QRY2)
 Function: test if R objects in two QRY objects are the same
           (irrespective of order)
 Example :
 Returns : 1 if equal, 0 otherwise
 Args    : two class QRY objects

Class R - request objects for QRY algebra

R SYNOPSIS
  $R = new R( $q1, $q2 );
  $R->put_atoms($q3);
  $R->del_atoms('coreceptor', 'phenotype');
  return $R->clone;
  $R1 = new R( new Q('subtype', 'B') );
  $R2 = new R( new Q('subtype', 'B C'),
               new Q('country', 'US') );
  R::Eq( (R::And($R1, $R2))[0],
         new R( new Q('subtype', 'B' ),
                new Q('country', 'US') ));                 # returns 1
  QRY::Eq( new QRY(R::Or($R1, $R2)), new QRY($R1, $R2) );  # returns 1
  R::In( (R::And($R1, $R2))[0], $R1 );                     # returns 1
R DESCRIPTION
Class R objects contain a list of atomic queries (class Q objects). Each class R object represents a single HTTP request to the LANL DB. When converted to a DB query, the class Q objects contained by an R object are effectively "AND"ed.
R CONSTRUCTOR
R constructor
 Title   : R constructor
 Usage   : $R = new R()
 Function: create a new R (request) object
 Example :
 Returns : class R (request) object
 Args    : optional, array of class Q objects
R INSTANCE METHODS
R len
 Title   : len
 Usage   : $R->len
 Function: get number of class Q objects contained in R object
 Example :
 Returns : scalar
 Args    :
R atoms
 Title   : atoms
 Usage   : $R->atoms( [optional $field])
 Function: get array of class Q (atomic query) objects in class R object
 Example : $R->atoms(); $R->atoms('coreceptor')
 Returns : array of class Q objects (all Qs or those corresponding to $field
           if present)
 Args    : optional, scalar string
R fields
 Title   : fields
 Usage   : $R->fields
 Function: get array of fields of all Q objects contained in $R
 Example :
 Returns : array of scalars
 Args    :
R put_atoms
 Title   : put_atoms
 Usage   : $R->put_atoms( @q )
 Function: AND an atomic query (class Q object) to the class R object's list
 Example :
 Returns : void
 Args    : an [array of] class Q object[s]
R del_atoms
 Title   : del_atoms
 Usage   : $R->del_atoms( @qfields )
 Function: removes class Q objects from R object's list according to the
           field names given in arguments
 Example :
 Returns : the class Q objects deleted
 Args    : scalar array of field names
R isnull
 Title   : isnull
 Usage   : $R->isnull
 Function: test if class R object is null
 Example :
 Returns : 1 if null, 0 otherwise
 Args    :
R A
 Title   : A
 Usage   : print $R->A
 Function: get a string representation of class R object
 Example :
 Returns : string scalar
 Args    :
R clone
 Title   : clone
 Usage   : $R2 = $R1->clone;
 Function: create and return a clone of the object
 Example :
 Returns : object of class R
 Args    :
R CLASS METHODS
R In
 Title   : In
 Usage   : R::In($R1, $R2)
 Function: tests whether the query represented by $R1 would return a subset
           of items returned by the query represented by $R2
 Example : print "R2 gets those and more" if R::In($R1, $R2);
 Returns : 1 if R1 is subset of R2, 0 otherwise
 Args    : two class R objects
R And
 Title   : And
 Usage   : @Rresult = R::And($R1, $R2)
 Function: logical AND for R objects
 Example :
 Returns : an array containing class R objects
 Args    : two class R objects
R Or
 Title   : Or
 Usage   : @Rresult = R::Or($R1, $R2)
 Function: logical OR for R objects
 Example :
 Returns : an array containing class R objects
 Args    : two class R objects
R Eq
 Title   : Eq
 Usage   : R::Eq($R1, $R2)
 Function: test if class Q objects in two R objects are the same
           (irrespective of order)
 Example :
 Returns : 1 if equal, 0 otherwise
 Args    : two class R objects

Class Q - atomic query objects for QRY algebra

Q SYNOPSIS
    $q = new Q('coreceptor', 'CXCR4 CCR5');
    $u = new Q('coreceptor', 'CXCR4');
    $q->fld;                                 # returns 'coreceptor'
    $q->dta;                                 # returns 'CXCR4 CCR5'
    print $q->A;                             # prints '(CXCR4 CCR5)[coreceptor]
    Q::qeq($q, $u);                          # returns 0
    Q::qeq( Q::qor($q, $q), $q );            # returns 1
    Q::qin($u, $q)                           # returns 1
    Q::qeq(Q::qand($u, $q), $u );            # returns 1
Q DESCRIPTION
Class Q objects represent atomic queries, that can be described by a single LANL cgi parameter=value pair. Class R objects (requests) are built from class Qs. The logical operations at the higher levels ("QRY, R") ultimately depend on the lower level operations on Qs: "qeq, qin, qand, qor".
Q CONSTRUCTOR
Q constructor
 Title   : Q constructor
 Usage   : $q = new Q($field, $data)
 Function: create a new Q (atomic query) object
 Example :
 Returns : class Q object
 Args    : optional $field, $data strings
Q INSTANCE METHODS
Q isnull
 Title   : isnull
 Usage   : $q->isnull
 Function: test if class Q object is null
 Example :
 Returns : 1 if null, 0 otherwise
 Args    :
Q fld
 Title   : fld
 Usage   : $q->fld($field)
 Function: get/set fld (field name) property
 Example :
 Returns : scalar
 Args    : scalar
Q dta
 Title   : dta
 Usage   : $q->dta($data)
 Function: get/set dta (whsp-separated data string) property
 Example :
 Returns : scalar
 Args    : scalar
Q A
 Title   : A
 Usage   : print $q->A
 Function: get a string representation of class Q object
 Example :
 Returns : string scalar
 Args    :
Q clone
 Title   : clone
 Usage   : $q2 = $q1->clone;
 Function: create and return a clone of the object
 Example :
 Returns : object of class Q
 Args    :
Q CLASS METHODS
Q qin
 Title   : qin
 Usage   : Q::qin($q1, $q2)
 Function: tests whether the query represented by $q1 would return a subset
           of items returned by the query represented by $q2
 Example : print "q2 gets those and more" if Q::qin($q1, $q2);
 Returns : 1 if q1 is subset of q2, 0 otherwise
 Args    : two class Q objects
Q qeq
 Title   : qeq
 Usage   : Q::qeq($q1, $q2)
 Function: test if fld and dta properties in two class Q objects are the same
           (irrespective of order)
 Example :
 Returns : 1 if equal, 0 otherwise
 Args    : two class Q objects
Q qor
 Title   : qor
 Usage   : @qresult = Q::qor($q1, $q2)
 Function: logical OR for Q objects
 Example :
 Returns : an array of class Q objects
 Args    : two class Q objects
Q qand
 Title   : qand
 Usage   : @qresult = Q::And($q1, $q2)
 Function: logical AND for R objects
 Example :
 Returns : an array of class Q objects
 Args    : two class Q objects
Q INTERNALS
Q unique
 Title   : unique
 Usage   : @ua = unique(@a)
 Function: return contents of @a with duplicates removed
 Example :
 Returns :
 Args    : an array

Additional tools for Bio::AnnotationCollectionI

Bio::AnnotationCollectionI SYNOPSIS (additional methods)
    $seq->annotation->put_value('patient_id', 1401)
    $seq->annotation->get_value('patient_ids')                   # returns 1401
    $seq->annotation->put_value('patient_group', 'MassGenH')
    $seq->annotation->put_value(['clinical', 'cd4count'], 503);
    $seq->annotation->put_value(['clinical', 'virus_load'], 150805);
    foreach ( qw( cd4count virus_load ) ) {
        $blood_readings{$_} = $seq->annonation->get_value(['clinical', $_]);
    }
Bio::AnnotationCollectionI DESCRIPTION (additional methods)
"get_value()" and "put_value" allow easy creation of and access to an annotation collection tree with nodes of Bio::Annotation::SimpleValue. These methods obiviate direct accession of the SimpleValue objects.

get_value

 Title   : get_value
 Usage   : $ac->get_value($tagname) -or-
           $ac->get_value( $tag_level1, $tag_level2,... )
 Function: access the annotation value assocated with the given tags
 Example :
 Returns : a scalar
 Args    : an array of tagnames that descend into the annotation tree

put_value

 Title   : put_value
 Usage   : $ac->put_value($tagname, $value) -or-
           $ac->put_value([$tag_level1, $tag_level2, ...], $value) -or-
           $ac->put_value( [$tag_level1, $tag_level2, ...] )
 Function: create a node in an annotation tree, and assign a scalar value to it
           if a value is specified
 Example :
 Returns : scalar or a Bio::AnnotationCollection object
 Args    : $tagname, $value scalars (can be specified as -KEYS=>$tagname,
           -VALUE=>$value) -or-
           \@tagnames, $value (or as -KEYS=>\@tagnames, -VALUE=>$value )
 Note    : If intervening nodes do not exist, put_value creates them, replacing
           existing nodes. So if $ac->put_value('x', 10) was done, then later,
           $ac->put_value(['x', 'y'], 20), the original value of 'x' is trashed,
           and $ac->get_value('x') will now return the annotation collection
           with tagname 'y'.

get_keys

 Title   : get_keys
 Usage   : $ac->get_keys($tagname_level_1, $tagname_level_2,...)
 Function: Get an array of tagnames underneath the named tag nodes
 Example : # prints the values of the members of Category 1...
           print map { $ac->get_value($_) } $ac->get_keys('Category 1') ;
 Returns : array of tagnames or empty list if the arguments represent a leaf
 Args    : [array of] tagname[s]
2017-08-30 perl v5.26.0