jonc has asked for the wisdom of the Perl Monks concerning the following question:

I have been trying to get this to work, without success: http://stackoverflow.com/questions/6541701/help-returning-2-arrays-from-subroutine-depending-on-stop-list

THIS HAS BEEN CHANGED TO A TEST CASE:

For original, look at stackoverflow, or davido's excellent reproduction below.

This produces the same errors MAINLY an uninitialized value. The length is mainly from my trying to add a test file, and some comments. Sorry ,don't know how else to replicate opening a file.

#!usr/bin/perl
use strict;
use warnings;

my @parsed = (
'[sent. 1 len. 27]: [Others, ,, the, sea, butterflies, and, heteropods
+, ,, have, a, modified, foot, that, functions, as, a, swimming, organ
+, with, which, they, move, through, open, ocean, waters, .]
(ROOT(S(NP(NP (NNS Others))(, ,)(NP (DT the) (NN sea) (NNS butterflies
+)(CC and)(NNS heteropods))(, ,))(VP (VBP have)(NP(NP (DT a) (VBN modi
+fied) (NN foot))(SBAR(WHNP (WDT that))(S(VP (VBZ functions)(PP (IN as
+)(NP(NP (DT a) (VBG swimming) (NN organ))(SBAR(WHPP (IN with)(WHNP (W
+DT which)))(S(NP (PRP they))(VP (VBP move)(PP (IN through)(NP (JJ ope
+n) (NN ocean) (NNS waters)))))))))))))    (. .)))
nsubj(have-9, Others-1)
conj_and(butterflies-5, heteropods-7)
dobj(have-9, foot-12)
nsubj(functions-14, foot-12)
nsubj(move-22, they-21)
prep_through(move-22, waters-26)'  ,

'[sent. 2 len. 10]: [Radially, symmetrical, animals, move, slowly, or,
+ not, at, all, .]
(ROOT(S(NP(ADJP (RB Radially) (JJ symmetrical))(NNS animals))(VP (VBP 
+move)(ADVP (RB slowly)(CC or)(RB not))(ADVP (IN at) (DT all)))    (. 
+.)))
nsubj(move-4, animals-3)
advmod(move-4, at-8)
pobj(at-8, all-9)'  ,

'Parsing [sent. 155 len. 31]: [Flatworms, -LRB-, phylum, Platyhelminth
+es, -RRB-, have, no, body, cavity, ,, lack, organs, for, oxygen, tran
+sport, ,, have, only, one, entrance, to, the, gut, ,, and, move, by, 
+beating, their, cilia, .]
(ROOT(S(NP(NP (NNS Flatworms))(PRN (-LRB- -LRB-)(NP (NNP phylum) (NNP 
+Platyhelminthes))(-RRB- -RRB-)))(VP (VBP have)(S(NP(NP (DT no) (NN bo
+dy) (NN cavity))(, ,)(NP(NP (NN lack) (NNS organs))(PP (IN for)(NP (N
+N oxygen) (NN transport))))(, ,))(VP(VP (VB have)(NP (RB only) (CD on
+e) (NN entrance))(PP (TO to)(NP (DT the) (NN gut))))(, ,)(CC and)(VP 
+(VB move)(PP (IN by)(S(VP (VBG beating)(NP (PRP$ their) (NN cilia))))
+)))))    (. .)))
nsubj(have-6, Flatworms-1)
nsubj(move-26, cavity-9)'  ,

'Parsing [sent. 27 len. 20]: [Because, fluids, are, relatively, incomp
+ressible, ,, they, move, to, another, part, of, the, cavity, when, mu
+scles, surrounding, them, contract, .]
(ROOT(S(SBAR (IN Because)(S(NP (NNS fluids))(VP (VBP are)(ADJP (RB rel
+atively) (JJ incompressible)))))(, ,)(NP (PRP they))(VP (VBP move)(PP
+ (TO to)(NP(NP (DT another) (NN part))(PP (IN of)(NP (DT the) (NN cav
+ity)))(SBAR(WHADVP (WRB when))(S(NP (NNS muscles))(VP (JJ surrounding
+)(NP (PRP them))(NP (NN contract))))))))    (. .)))
advcl(move-8, incompressible-5)
nsubj(move-8, they-7)'  ,

'Parsing [sent. 18 len. 27]: [Cnidarians, also, have, epithelial, cell
+s, with, muscle, fibers, whose, contractions, enable, the, animals, t
+o, move, ,, as, well, as, nerve, nets, that, integrate, their, body, 
+activities, .]
(ROOT(S(NP (NNS Cnidarians))(ADVP (RB also))(VP (VBP have)(NP(NP(NP (J
+J epithelial) (NNS cells))(PP (IN with)(NP (NN muscle) (NNS fibers)))
+(SBAR(WHNP (WP$ whose) (NNS contractions))(S(VP (VBP enable)(S(NP (DT
+ the) (NNS animals))(VP (TO to)(VP (VB move))))))))(, ,)(CONJP (RB as
+) (RB well) (IN as))(NP(NP (NN nerve) (NNS nets))(SBAR(WHNP (WDT that
+))(S(VP (VB integrate)(NP (PRP$ their) (NN body) (NNS activities)))))
+)))    (. .)))
advmod(have-3, also-2)
nsubj(move-15, animals-13)'
);
# --- If qq is the same as reading in a file. :
# local $/ = 'Parsing';
# open(my $parse_corpus, '<', "/Users/jon/Desktop/stanford-postagger-f
+ull-2011-04-20/parsedLife2.txt") or die "Couldn't open directory $!";
#
# Note: split at Parsed, which is before each [sent. ...]

my @stopListNoun = ("theirs", "they");
# --- Unsure if same as real version: -- #
# open my $stop_list_noun, '<', $stopListNounFile or die "could not op
+en 'stoplist_noun.txt' $!"; ##Just open, no writing or reading?
# my @stopListNoun = <$stop_list_noun>;
# chomp @stopListNoun;
# close $stop_list_noun or die "could not close 'stoplist_noun.txt' $!
+";
#
# the file has a word each line.
</readmore>
my $search_key = "move";
my (@all_matches, @all_pronoun_matches);
my ($chapternumber, $sentencenumber, $sentence,
     $grammar_relation, $argument1, $argument2);

foreach my $sentblock (@parsed)
        {
    chomp $sentblock;
            
    next unless ($sentblock =~ /\[sent. (\d+) len. \d+\]: \[(.+)\]/);
    $sentencenumber = $1;
    $sentence = $2;
    $sentence =~ s/, / /g;
    $chapternumber = "1_1"; #From regex
    
    next unless ($sentblock =~ /\b$search_key\b/i); ##Ensure the sente
+nce contains the searchkey
    next unless ($sentblock =~ /\(VB\w*\s+\b$search_key\b[\)\s]+/i); #
+#Ensure searchkey is a verb
    my ($arg1, $arg2, $goodmatch);
    my @lines = split ("\n",$sentblock); ##Split by a newline
    
    for my $l (0..$#lines)
                {
                if (($lines[$l] =~ /subj\w*\(/) && ($lines[$l] =~ /\b$
+search_key\b/i))
                        {
                next unless ($lines[$l] =~ /\w+\(\w+\-\d+\,\s(\w+)\-\d
++\)/);
                my ($matches, $pronoun_matches) = &dependency_checks($
+lines[$l], $search_key, $chapternumber, $sentencenumber, $sentence);
                push @all_matches, $matches if ($matches);
                push @all_pronoun_matches, $pronoun_matches if ($prono
+un_matches);
                }
            }    
        }
    
my %counts;
foreach my $rowref (@all_matches)
{
    $counts{lc($rowref->[5])}++;  
}

my %pronouncounts;
foreach my $pronounrowref (@all_pronoun_matches)
{
    $pronouncounts{lc($pronounrowref->[5])}++;  
}

@all_matches = sort { $counts{lc($b->[5])} <=> $counts{lc($a->[5])} ||
                      lc($a->[5]) cmp lc($b->[5])
                    } @all_matches;
# for pronoun_matches, same sort, then concatenate to all_matches
@all_pronoun_matches = sort { $pronouncounts{lc($b->[5])} <=> $pronoun
+counts{lc($a->[5])} ||
                      lc($a->[5]) cmp lc($b->[5])
                    } @all_pronoun_matches;
@all_matches = (@all_matches, @all_pronoun_matches);

my %seen_header;
my %seen_subheader;
foreach my $match (@all_matches)
    {
    $match->[3] = "Subject";

    my $header = $counts{lc $match->[5]}." match(es) in which the ".$m
+atch->[3]." of ".$match->[4]." is ".$match->[5]." :\n\n";
    print $header unless $seen_subheader{lc $match->[5]}++;
    print "Section ".$match->[0].": ".$match->[2]."\n\n"; ##Section an
+d sentence (formatted)
    } #Foreach match

sub dependency_checks
    {
    my ($line, $verbform, $chapternumber, $sentencenumber, $sentence) 
+= @_;
    my @matches;
    my @pronoun_matches;
    return unless ($line =~ /(\w+)\((\w+)\-\d+\,\s(\w+)\-\d+\)/); #Cou
+ld pass this in
    $grammar_relation = $1;
    $argument1 = $2;
    $argument2 = $3;

    foreach my $pronoun (@stopListNoun)
        {
        if ((lc $pronoun eq lc $argument1) || (lc $pronoun eq lc $argu
+ment2))
            {
            push (@pronoun_matches, $chapternumber, $sentencenumber, $
+sentence, $grammar_relation, $argument2, $argument1) if ($argument2 =
+~ /$verbform/i);
            push (@pronoun_matches, $chapternumber, $sentencenumber, $
+sentence, $grammar_relation, $argument1, $argument2) if ($argument1 =
+~ /$verbform/i);    
            return;
            }
        }
    #Make sure searchkey is 1st arg:
    push (@matches, $chapternumber, $sentencenumber, $sentence, $gramm
+ar_relation, $argument2, $argument1) if ($argument2 =~ /$verbform/i);
    push (@matches, $chapternumber, $sentencenumber, $sentence, $gramm
+ar_relation, $argument1, $argument2) if ($argument1 =~ /$verbform/i);
    
    return (\@matches, \@pronoun_matches);
    }

[download]

Replies are listed 'Best First'.
Re: Return 2 arrays, sort the same, and concatenate them
by davido (Cardinal) on Jul 01, 2011 at 05:59 UTC

    Let me know if I should post more info.

    I've been following your other thread as well. I tidied up the formatting of the code you posted so that when I looked at it I would be able to read it (it's amazing how accustomed we get to careful formatting, and anything less makes code difficult to read). I'm going to paste here your posted code after being tidied up a bit, and will follow it with a question.

    # Part of the dependency_checks subroutine

foreach my $pronoun ( @stopListNoun ) {
    if ( 
        ( lc $pronoun eq lc $argument1 ) || 
        ( lc $pronoun eq lc $argument2 ) 
    ) {
        return;
    }
}
#Make sure searchkey is 1st arg:
push (
    @matches, 
    $chapternumber, $sentencenumber, $sentence, 
    $grammar_relation, $argument2, $argument1
) if ( $argument2 =~ /$verbform/i );    # USED TO BE 'eq', but that 
                                        # prevented protective from sh
+owing

push (
    @matches, 
    $chapternumber, $sentencenumber, $sentence, 
    $grammar_relation, $argument1, $argument2
) if ( $argument1 =~ /$verbform/i );

return \@matches;


# Called by:

$matches = &dependency_checks(
    $lines[$l], 
    $verbform, $chapternumber, 
    $sentencenumber, $sentence
);

push @all_matches, $matches if ( $matches );

    [download]

    The question: Would you be able to look at that code and the description of the problem that you posted, and come up with any kind of meaningful answer?

    There is so much missing that we cannot possibly know what this small snippet you posted is supposed to be doing. It's part of a bigger picture. If you look at a TV screen through a drinking straw and with the volume turned down, can you tell what's going on in the TV show?

    I would like to help because I enjoy the puzzle of working through problems. But I have no idea what's going on in your code.

    My recommendation is to post a snippet of code that is capable of compiling and running, but that shows the problem you're encountering. I don't mean 500 lines of code. I mean less than 30. If you are unable to reduce your problem to 30 lines of code, including a small sample of data, then the problem you think you're having isn't the problem. The real problem is disorganized and unmaintainable code which is leading you down a road to broken code.

    One thing that I've found is that as I work on boiling a problem down to a minimal amount of code that I can post to get help, I usually solve my problem anyway. Just a couple days ago I was working on a piece of code that seemed like it worked great until I ran a specific test on it. Long story short, I spent way too long on it, and finally started preparing a small snippet to post here for advice. Two minutes into my effort to boil it down to a snippet small enough to post, I discovered the problem. Two minutes later I fixed the original code and executed the test successfully.

    So I'll say it again. Please post an example of your input right before things break. Post an actual snippet of code that will compile and will take this input and break it. Tell us what output you should be getting, and how it differs from what's actually happening.

    Help us help you.


    Dave

[reply]
[d/l]
[reply]
        Thanks a lot for a detailed explanation of how to do the basics. I'm sorry for my inadequate posts. I will do what you say from here on. Looks like I've been a neglecting some reading here. You have inspired me. Thanks
[reply]

      +1 Dave, or + as many as possible. I have tried to do what you asked, and feel a little more confident now. Thanks!

      I made the code short, but couldn't figure how to make a test file input short. So I hope you can use your problem - solving abilities now! Thanks again for the valuable lesson.

[reply]
Re: Return 2 arrays, sort the same, and concatenate them
by jonc (Beadle) on Jul 01, 2011 at 13:20 UTC

    I've tried using the same %counts for the sort, because that struck me as a problem, but that didn't fix it.

    I have also tried using two separate subroutines. Note is fine if you use two separate print statements.

    Also, Sorry for the formatting, it didn't download from here the same as I have it...

[reply]
[d/l]

      Nice job, thinks are looking more answerable already

      Ok, here is the next step, start making functions :)

      Functions with meaningful names

      For an example see Re^6: Help with locating bp region in chromosome

      You would turn each loop into a separate function, so you end up with something like 

      ...
Main( @ARGV );
exit( 0 );
sub Main {
    ...
    
    my ( $allPronouns, $allMatches ) = MeaningfulMatches( \@parsed );
    my $PronounCountsHashref = CountPronouns( $allPronouns );
    my $CountsHashref        = CountMatches( $all_matches );
    SortThisMotherReference( $CountsHashref,        $allMatches );
    SortThisMotherReference( $PronounCountsHashref, $allPronouns );
    
    ...
}

      [download]

      See, isn't that easier to read? And see how the meaningful function names replace some of the comments?

      Now that you've got things separated into functions, its easy to use test and tweak each part individually without worrying about the whole thing working, until you've got it producing the correct output.

      So the next step is to make sure each function does its job by writing tests for each function. You give it input, it gives you output (or it modifies your input), and you compare the two to see if they match. You can use Test::Deep for this.

      Once they match, you move on to the next function

      Oh you say CountPronouns() is doing an inaccurate count? Or SortThisMotherReference() is not sorting correctly?

      Now its easy to write a small test program. Use Data::Dumper to create sample input, modify it by hand until you have sample output, and then modify the function until it produces the same output.

      For another example see Re^3: SEO Fixer Part II - Updated or watch this ~14min video String Calculator TDD Kata done in Perl of a programmer doing this live :)

[reply]
[d/l]

        Yes, it is helpful that way. More modular.

        I am almost positive that the wrong info (or none) is being pushed into @all_pronoun_matches.

        I came to this conclusion because after using your methods, I got an output that had @all_matches correct. BUT the info in @all_pronoun_matches was messed:

        The heading had the wrong count (didn't sort, just had # of @all_matches), and didn't have $match->[4] and $match[5].

        The matches under the heading didn't have anything (just printed "Section")

        I still can't say why the push isn't working though. THANKS

[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom