NAME
Data::Roundtrip - convert between Perl data structures, YAML and JSON with unicode support (I believe ...)
VERSION
Version 0.29
SYNOPSIS
This module contains a collection of utilities for converting between JSON, YAML, Perl variable and a Perl variable's string representation (aka dump). Hopefully, all unicode content will be handled correctly between the conversions and optionally escaped or un-escaped. Also JSON can be presented in a pretty format or in a condensed, machine-readable format (not spaces, indendation or line breaks).
use Data::Roundtrip qw/:all/;
#use Data::Roundtrip qw/json2yaml/;
#use Data::Roundtrip qw/:json/; # see EXPORT
$jsonstr = '{"Songname": "Απόκληρος της κοινωνίας",'
.'"Artist": "Καζαντζίδης Στέλιος/Βίρβος Κώστας"}'
;
$yamlstr = json2yaml($jsonstr);
print $yamlstr;
# NOTE: long strings have been broken into multilines
# and/or truncated (replaced with ...)
#---
#Artist: Καζαντζίδης Στέλιος/Βίρβος Κώστας
#Songname: Απόκληρος της κοινωνίας
$yamlstr = json2yaml($jsonstr, {'escape-unicode'=>1});
print $yamlstr;
#---
#Artist: \u039a\u03b1\u03b6\u03b1 ...
#Songname: \u0391\u03c0\u03cc\u03ba ...
$backtojson = yaml2json($yamlstr);
# $backtojson is a string representation
# of following JSON structure:
# {"Artist":"Καζαντζίδης Στέλιος/Βίρβος Κώστας",
# "Songname":"Απόκληρος της κοινωνίας"}
# This is useful when sending JSON via
# a POST request and it needs unicode escaped:
$backtojson = yaml2json($yamlstr, {'escape-unicode'=>1});
# $backtojson is a string representation
# of following JSON structure:
# but this time with unicode escaped
# (pod content truncated for readbility)
# {"Artist":"\u039a\u03b1\u03b6 ...",
# "Songname":"\u0391\u03c0\u03cc ..."}
# this is the usual Data::Dumper dump:
print json2dump($jsonstr);
#$VAR1 = {
# 'Songname' => "\x{391}\x{3c0}\x{3cc} ...",
# 'Artist' => "\x{39a}\x{3b1}\x{3b6} ...",
#};
# and this is a more human-readable version:
print json2dump($jsonstr, {'dont-bloody-escape-unicode'=>1});
# $VAR1 = {
# "Artist" => "Καζαντζίδης Στέλιος/Βίρβος Κώστας",
# "Songname" => "Απόκληρος της κοινωνίας"
# };
# pass some parameters to Data::Dumper
# like: be terse (no $VAR1):
print json2dump($jsonstr,
{'dont-bloody-escape-unicode'=>0, 'terse'=>1}
#{'dont-bloody-escape-unicode'=>0, 'terse'=>1, 'indent'=>0}
);
# {
# "Artist" => "Καζαντζίδης Στέλιος/Βίρβος Κώστας",
# "Songname" => "Απόκληρος της κοινωνίας"
# }
# this is how to reformat a JSON string to
# have its unicode content escaped:
my $json_with_unicode_escaped =
json2json($jsonstr, {'escape-unicode'=>1});
# With version 0.18 and up two more exported-on-demand
# subs were added to read JSON or YAML directly from a file:
# jsonfile2perl() and yamlfile2perl()
my $perldata = jsonfile2perl("file.json");
my $perldata = yamlfile2perl("file.yaml");
die "failed" unless defined $perldata;
# For some of the above functions there exist command-line scripts:
perl2json.pl -i "perl-data-structure.pl" -o "output.json" --pretty
json2json.pl -i "with-unicode.json" -o "unicode-escaped.json" --escape-unicode
# etc.
# only for *2dump: perl2dump, json2dump, yaml2dump
# and if no escape-unicode is required (i.e.
# setting 'dont-bloody-escape-unicode' => 1 permanently)
# and if efficiency is important,
# meaning that perl2dump is run in a loop thousand of times,
# then import the module like this:
use Data::Roundtrip qw/:all no-unicode-escape-permanently/;
# or like this
use Data::Roundtrip qw/:all unicode-escape-permanently/;
# then perl2dump() is more efficient but unicode characters
# will be permanently not-escaped (1st case) or escaped (2nd case).
EXPORT
By default no symbols are exported. However, the following export tags are available (:all will export all of them):
:json
:perl2json()
,json2perl()
,json2dump()
,json2yaml()
,json2json()
,jsonfile2perl()
:yaml
:perl2yaml()
,yaml2perl()
,yaml2dump()
,yaml2yaml()
,yaml2json()
,yamlfile2perl()
:dump
:perl2dump()
,perl2dump_filtered()
,perl2dump_homebrew()
:io
:read_from_file()
,write_to_file()
,read_from_filehandle()
,write_to_filehandle()
,:all
: everything above.Additionally, these four subs:
dump2perl()
,dump2json()
,dump2yaml()
,dump2dump()
do not belong to any export tag. However they can be imported explicitly by the caller in the usual way (e.g.use Data::Roundtrip qw/dump2perl perl2json .../
). Section CAVEATS, under "dump2perl", describes how these subseval()
a string possibly coming from user, possibly being unchecked.no-unicode-escape-permanently
: this is not an export keyword/parameter but a parameter which affects all the*2dump*
subs by setting unicode escaping permanently to false. See "EFFICIENCY".unicode-escape-permanently
: this is not an export keyword/parameter but a parameter which affects all the*2dump*
subs by setting unicode escaping permanently to true. See "EFFICIENCY".
EFFICIENCY
The export keyword/parameter no-unicode-escape-permanently
affects all the *2dump*
subs by setting unicode escaping permanently to false. This improves efficiency, although one will ever need to use this in extreme situations where a *2dump*
sub is called repeatedly in a loop of a few hundreds or thousands of iterations or more.
Each time a *2dump*
is called, the dont-bloody-escape-unicode
flag is checked and if it is set, then Data::Dumper's qquote()
is overriden with _qquote_redefinition_by_Corion()
just for that instance and will be restored as soon as the dump is finished. Similarly, a filter for not escaping unicode is added to Data::Dump just for that particular call and is removed immediately after. This has some computational cost and can be avoided completely by overriding the sub and adding the filter once, at loading (in import()
).
The price to pay for this added efficiency is that unicode in any dump will never be escaped (e.g. \x{3b1})
, but will be rendered (e.g. α
, a greek alpha). Always. The option dont-bloody-escape-unicode
will permanently be set to true.
Similarly, the export keyword/parameter unicode-escape-permanently
affects all the *2dump*
subs by setting unicode escaping permanently to true. This improves efficiency as well.
See "BENCHMARKS" on how to find the fastest *2dump*
sub.
BENCHMARKS
The special Makefile target benchmarks
will time calls to each of the *2dump*
subs under
use Data::Roundtrip;
use Data::Roundtrip qw/no-unicode-escape-permanently/;
use Data::Roundtrip qw/unicode-escape-permanently/;
and for 'dont-bloody-escape-unicode' => 0
and 'dont-bloody-escape-unicode' => 1
.
In general, "perl2dump" is faster by 25% when one of the permanent import parameters is used (either of the last two cases above).
SUBROUTINES
perl2json
my $ret = perl2json($perlvar, $optional_paramshashref)
Arguments:
$perlvar
$optional_paramshashref
Return value:
$ret
Given an input $perlvar
(which can be a simple scalar or a nested data structure, but not an object), it will return the equivalent JSON string. In $optional_paramshashref
one can specify whether to escape unicode with 'escape-unicode' => 1
and/or prettify the returned result with 'pretty' => 1
. The output can be fed back to "json2perl" for getting the Perl variable back.
It returns the JSON string on success or undef
on failure.
json2perl
Arguments:
$jsonstring
Return value:
$ret
Given an input $jsonstring
as a string, it will return the equivalent Perl data structure using JSON::decode_json(Encode::encode_utf8($jsonstring))
.
It returns the Perl data structure on success or undef
on failure.
perl2yaml
my $ret = perl2yaml($perlvar, $optional_paramshashref)
Arguments:
$perlvar
$optional_paramshashref
Return value:
$ret
Given an input $perlvar
(which can be a simple scalar or a nested data structure, but not an object), it will return the equivalent YAML string. In $optional_paramshashref
one can specify whether to escape unicode with 'escape-unicode' => 1
. Prettify is not supported yet. The output can be fed to "yaml2perl" for getting the Perl variable back.
It returns the YAML string on success or undef
on failure.
yaml2perl
my $ret = yaml2perl($yamlstring);
Arguments:
$yamlstring
Return value:
$ret
Given an input $yamlstring
as a string, it will return the equivalent Perl data structure using YAML::PP::Load($yamlstring)
It returns the Perl data structure on success or undef
on failure.
yamlfile2perl
my $ret = yamlfile2perl($filename)
Arguments:
$filename
Return value:
$ret
Given an input $filename
which points to a file containing YAML content, it will return the equivalent Perl data structure.
It returns the Perl data structure on success or undef
on failure.
perl2dump
my $ret = perl2dump($perlvar, $optional_paramshashref)
Arguments:
$perlvar
$optional_paramshashref
Return value:
$ret
Given an input $perlvar
(which can be a simple scalar or a nested data structure, but not an object), it will return the equivalent string (via Data::Dumper). In $optional_paramshashref
one can specify whether to escape unicode with 'dont-bloody-escape-unicode' => 0
, (or 'escape-unicode' => 1
). The DEFAULT behaviour is to NOT ESCAPE unicode.
Additionally, use terse output with 'terse' => 1
and remove all the incessant indentation with 'indent' => 1
which unfortunately goes to the other extreme of producing a space-less output, not fit for human consumption. The output can be fed to "dump2perl" for getting the Perl variable back.
It returns the string representation of the input perl variable on success or undef
on failure.
The output can be fed back to "dump2perl".
CAVEAT: when not escaping unicode (which is the default behaviour), each call to this sub will override Data::Dumper's qquote()
sub then call Data::Dumper's Dumper()
and save its output to a temporary variable, restore qquote()
sub to its original code ref and return the contents. This exercise is done every time this perl2dump()
is called. It may be expensive. The alternative is to redefine qquote()
once, when the module is loaded, with all the side-effects this may cause.
Note that there are two other alternative subs which offer more-or-less the same functionality and their output can be fed back to all the dump2*()
subs. These are "perl2dump_filtered" which uses Data::Dump::Filtered to add a filter to control unicode escaping but lacks in aesthetics and functionality and handling all the cases Dump and Dumper do quite well.
There is also perl2dump_homebrew()
which uses the same dump-recursively engine as "perl2dump_filtered" but does not involve Data::Dump at all.
perl2dump_filtered
my $ret = perl2dump_filtered($perlvar, $optional_paramshashref)
Arguments:
$perlvar
$optional_paramshashref
Return value:
$ret
It does the same job as "perl2dump" which is to stringify a perl variable. And takes the same options.
It returns the string representation of the input perl variable on success or undef
on failure.
It uses Data::Dump::Filtered to add a filter to Data::Dump.
perl2dump_homebrew
my $ret = perl2dump_homebrew($perlvar, $optional_paramshashref)
Arguments:
$perlvar
$optional_paramshashref
Return value:
$ret
It does the same job as "perl2dump" which is to stringify a perl variable. And takes the same options.
It returns the string representation of the input perl variable on success or undef
on failure.
The output can be fed back to "dump2perl".
It uses its own basic dumper. Which is recursive. So, beware of extremely deep nested data structures. Deep not long! But it probably is as efficient as it can be but definetely lacks in aesthetics and functionality compared to Dump and Dumper.
dump_perl_var_recursively
my $ret = dump_perl_var_recursively($perl_var)
Arguments:
$perl_var
, a Perl variable like a scalar or an arbitrarily nested data structure. For the latter, it requires references, e.g. hash-ref or arrayref.
Return value:
$ret
, the stringified version of the input Perl variable.
This sub will take a Perl var (as a scalar or an arbitrarily nested data structure) and emulate a very very basic Dump/Dumper but with enforced rendering unicode (for keys or values or array items), and not escaping unicode - this is not an option, it returns a string representation of the input perl var
There are 2 obvious limitations:
- 1. indentation is very basic,
- 2. it supports only scalars, hashes and arrays, (which will dive into them no problem) This sub can be used in conjuction with DataDumpFilterino() to create a Data::Dump filter like,
-
Data::Dump::Filtered::add_dump_filter( \& DataDumpFilterino ); or dumpf($perl_var, \& DataDumpFilterino);
the input is a Perl variable as a reference, so no
%inp
but$inp={}
and$inp=[]
.This function is recursive. Beware of extremely deep nested data structures. Deep not long! But it probably is as efficient as it can be but definetely lacks in aesthetics and functionality compared to Dump and Dumper.
The output is a, possibly multiline, string. Which it can then be fed back to "dump2perl".
dump2perl
# CAVEAT: it will eval($dumpstring) internally, so
# check $dumpstring for malicious code beforehand
# it is a security risk if you don't.
# Don't use it if $dumpstring comes from
# untrusted sources (user input for example).
my $ret = dump2perl($dumpstring)
Arguments:
$dumpstring
, this comes from the output of Data::Dump, Data::Dumper or our own "perl2dump", "perl2dump_filtered", "perl2dump_homebrew". Escaped, or unescaped.
Return value:
$ret
, the Perl data structure on success orundef
on failure.
CAVEAT: it eval()'s the input $dumpstring
in order to create the Perl data structure. eval()'ing unknown or unchecked input is a security risk. Always check input to eval() which comes from untrusted sources, like user input, scraped documents, email content. Anything really.
json2perl
my $ret = json2perl($jsonstring)
Arguments:
$jsonstring
Return value:
$ret
Given an input $jsonstring
as a string, it will return the equivalent Perl data structure using JSON::decode_json(Encode::encode_utf8($jsonstring))
.
It returns the Perl data structure on success or undef
on failure.
jsonfile2perl
my $ret = jsonfile2perl($filename)
Arguments:
$filename
Return value:
$ret
Given an input $filename
which points to a file containing JSON content, it will return the equivalent Perl data structure.
It returns the Perl data structure on success or undef
on failure.
json2yaml
my $ret = json2yaml($jsonstring, $optional_paramshashref)
Arguments:
$jsonstring
$optional_paramshashref
Return value:
$ret
Given an input JSON string $jsonstring
, it will return the equivalent YAML string YAML by first converting JSON to a Perl variable and then converting that variable to YAML using "perl2yaml". All the parameters supported by "perl2yaml" are accepted.
It returns the YAML string on success or undef
on failure.
yaml2json
my $ret = yaml2json($yamlstring, $optional_paramshashref)
Arguments:
$yamlstring
$optional_paramshashref
Return value:
$ret
Given an input YAML string $yamlstring
, it will return the equivalent YAML string YAML by first converting YAML to a Perl variable and then converting that variable to JSON using "perl2json". All the parameters supported by "perl2json" are accepted.
It returns the JSON string on success or undef
on failure.
json2json
yaml2yaml
Transform a json or yaml string via pretty printing or via escaping unicode or via un-escaping unicode. Parameters like above will be accepted.
json2dump
dump2json
yaml2dump
dump2yaml
These subs offer similar functionality as their counterparts described above.
Section CAVEATS, under "dump2perl", describes how dump2*()
subs eval()
a string possibly coming from user, possibly being unchecked.
dump2dump
my $ret = dump2dump($dumpstring, $optional_paramshashref)
Arguments:
$dumpstring
$optional_paramshashref
Return value:
$ret
For example:
my $dumpstr = '...';
my $newdumpstr = dump2dump(
$dumpstr,
{
'dont-bloody-escape-unicode' => 1,
'terse' => 0,
}
);
It returns the a dump string similar to
read_from_file
my $contents = read_from_file($filename)
Arguments:
$filename
: the input filename.
Return value:
$contents
Given a filename, it opens it using :encoding(UTF-8)
, slurps its contents and closes it. It's a convenience sub which could have also been private. If you want to retain the filehandle, use "read_from_filehandle".
It returns the file contents on success or undef
on failure.
read_from_filehandle
my $contents = read_from_filehandle($filehandle)
Arguments:
$filehandle
: the handle to an already opened file.
Return value:
$contents
: the file contents slurped.
It slurps all content from the specified input file handle. Upon return the file handle is still open. It returns the file contents on success or undef
on failure.
write_to_file
write_to_file($filename, $contents) or die
Arguments:
$filename
: the output filename.$contents
: any string to write it to file.
Return value:
1 on success, 0 on failure
Given a filename, it opens it using :encoding(UTF-8)
, writes all specified content and closes the file. It's a convenience sub which could have also been private. If you want to retain the filehandle, use "write_to_filehandle".
It returns 1 on success or 0 on failure.
write_to_filehandle
write_to_filehandle($filehandle, $contents) or die
Arguments:
$filehandle
: the handle to an already opened file (for writing).
Return value:
1 on success or 0 on failure.
It writes content to the specified file handle. Upon return the file handle is still open.
It returns 1 on success or 0 on failure.
SCRIPTS
A few scripts have been put together and offer the functionality of this module to the command line. They are part of this distribution and can be found in the script
directory.
These are: json2json.pl
, json2yaml.pl
, yaml2json.pl
, json2perl.pl
, perl2json.pl
, yaml2perl.pl
CAVEATS
I have to apologise here to the authors of YAML::PP for defaming them because I clumsily wrote YAML::PP when I wanted to write YAML.
So, the reality is that YAML::PP does not have any problem in handling the edge-case below.
A valid Perl variable may kill YAML's Load()
because of escapes and quotes. For example this:
my $yamlstr = <<'EOS';
---
- 682224
- "\"w": 1
EOS
my $pv = eval { YAML::Load($yamlstr) };
if( $@ ){ die "failed(1): ". $@ }
# it's dead
Strangely, there is no problem for this:
my $yamlstr = <<'EOS';
---
- 682224
- "\"w"
EOS
# this is OK also:
# - \"w: 1
my $pv = eval { YAML::Load($yamlstr) };
if( $@ ){ die "failed(1): ". $@ }
# it's OK! still alive.
I have provided an author-only test (make deficiencies
) which tests all three of them on the edge cases. Both YAML::PP and YAML::XS pass the tests.
This YAML issue is relevant. Many thanks to CPAN authors TINITA and INGY for their work on this, and on YAML*
too.
For now, the plan is to still use YAML::PP and avoid explicitly requiring YAML::XS until YAML::Any is ready.
Be warned that sub dump2perl()
eval()
's its input. If this comes from the user and it is not checked then it is considered a security problem. Subs dump2json()
, dump2yaml()
, dump2dump()
use dump2perl()
. The four subs will issue a warning whenever you call them. Additionally, as from version 0.28, they need to be explicitly imported like:
use Data::Roundtrip qw/... dump2perl .../
They are no longer part of export tag :dump
nor :all
. If their input comes from the user please check the input not to contain malicious code which when eval()
'ed can create security concerns.
AUTHOR
Andreas Hadjiprocopis, <bliako at cpan.org> / <andreashad2 at gmail.com>
BUGS
Please report any bugs or feature requests to bug-data-roundtrip at rt.cpan.org
, or through the web interface at https://rt.cpan.org/NoAuth/ReportBug.html?Queue=Data-Roundtrip. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.
SEE ALSO
- Convert JSON to Perl and back with unicode
- RFC: Perl<->JSON<->YAML<->Dumper : roundtripping and possibly with unicode
SUPPORT
You can find documentation for this module with the perldoc command.
perldoc Data::Roundtrip
You can also look for information at:
RT: CPAN's request tracker (report bugs here)
AnnoCPAN: Annotated CPAN documentation
Review this module at PerlMonks
Search CPAN
ACKNOWLEDGEMENTS
Several Monks at PerlMonks.org (in no particular order):
- haukex
- Corion (the
_qquote_redefinition_by_Corion()
which harnesses Data::Dumper's incessant unicode escaping) - kcott (The EXPORT section among other suggestions)
- jwkrahn
- leszekdubiel
- marto
- Haarg
- and an anonymous monk
DEDICATIONS
Almaz!
LICENSE AND COPYRIGHT
This software, EXCEPT the portions created by [Corion] @ Perlmonks and [kcott] @ Perlmonks, is Copyright (c) 2020 by Andreas Hadjiprocopis.
This is free software, licensed under:
The Artistic License 2.0 (GPL Compatible)
1 POD Error
The following errors were encountered while parsing the POD:
- Around line 1386:
=back doesn't take any parameters, but you said =back Given an input string C<$dumpstring>, which can have been produced by e.g. C<perl2dump()> and is identical to L<Data::Dumper>'s C<Dumper()> output, it will roundtrip back to the same string, possibly with altered format via the parameters in C<$optional_paramshashref>.