Changeset 37184 for main/trunk/greenstone2/perllib/plugins/SplitJSONFile.pm

Timestamp:

2023-01-25T23:15:40+13:00 (15 months ago)

Author:

davidb

Message:

Further refinement of idea, with emphasis on using plugins arguments rather than having things hardwired

File:

• main/trunk/greenstone2/perllib/plugins/SplitJSONFile.pm (modified) (3 diffs)

main/trunk/greenstone2/perllib/plugins/SplitJSONFile.pm

@@ -57 +57 @@
       'deft' => "",
       'reqd' => "no" },
-    { 'name' => "metadata_exp",
+    { 'name' => "file_exp",
       'desc' => "{SplitJSONFile.file_exp}",
       'type' => "string",
@@ -105 +105 @@
 
     my $json = JSON->new();
+    my $json_text_content = JSON::from_json($$textref);
+
+    my $json_array = [ ];
     
     # Split the text into several smaller segments
     my $split_exp = $self->{'split_exp'};
 
+    my @split_exps = split(/\s*,\s*/,$split_exp);
+    
+    if (scalar(@split_exps) == 0) {
+    @split_exps = ("");
+    }
+    
+    foreach my $full_json_field (@split_exps) {
 
-    my $json_text_content = JSON::from_json($$textref);
-    
-    my $json_array1 = $json_text_content->{'contentGroups'};
-    my $json_array2 = $json_text_content->{'contentItems'};
+    $full_json_field =~ s/^\.?//;
+    
+    my @json_field_parts = ($full_json_field eq "") ? () : split(/\./,$full_json_field);
 
-    # my $json_array = [ @$json_array1, @$json_array2 ];
+    ## print STDERR "**** json_field_parts = ", JSON::to_json(\@json_field_parts), "\n";
+    my $json_cursor = $json_text_content;
 
-    my $json_array = [  @$json_array2 ];
-    
-    if ($verbosity>1) {
+    my $had_error = 0;
+    
+    foreach my $json_field_part (@json_field_parts) {
+        $json_cursor = $json_cursor->{$json_field_part};
+        if (!defined $json_cursor) {
+        print STDERR "Warning: failed to find '$full_json_field' in JSON file.  '$json_field_part' did not exist\n";
+        print STDERR "-> Skipping\n";
+        $had_error = 1;
+        last;
+        }
+    }
+
+    if (!$had_error) {
+        
+        if (ref($json_cursor) eq "ARRAY") {
+        push(@$json_array,@$json_cursor);
+        }
+        else {
+        if ($full_json_field eq "") {
+            print STDERR "Warning: The top-level of the JSON file is not an array field\n";
+        }
+        else {
+            print STDERR "Warning: The specified field '$full_json_field' was detected in the JSON file, however it is not an array field\n";
+        }
+        print STDERR "-> Skipping\n";
+        }
+    }
+    }
+
+    if ($verbosity>=3) {
     print $outhandle "----------\n";
     print $outhandle "SplitJSONFile -- Segments\n";
@@ -133 +170 @@
     my $seg_json_unicode_str = $json->encode($seg_json_rec); # expects unicode string
 
-    if ($verbosity>1) {
+    if ($verbosity>=3) {
         my $seg_json_utf8_printable_str = Encode::encode("utf8",$seg_json_unicode_str);