# Table of Contents * [NAME](#name) * [SYNOPSIS](#synopsis) * [DESCRIPTION](#description) * [EXPORTED METHODS](#exported-methods) * [METHODS AND SUBROUTINES](#methods-and-subroutines) * [process\_csv](#process\csv) * [process\_file](#process\file) * [Custom Processors](#custom-processors) * [pre(file, options)](#prefile-options) * [next\_line(fh, lines, options)](#next\linefh-lines-options) * [filter(fh, lines, options, current\_line)](#filterfh-lines-options-current\line) * [process(fh, lines, options, current\_line)](#processfh-lines-options-current\line) * [post(fh, lines, options)](#postfh-lines-options) * [DEFAULT PROCESSORS](#default-processors) * [STATISTICS](#statistics) * [EXAMPLES](#examples) * [CAVEATS](#caveats) * [LICENSE](#license) * [SEE ALSO](#see-also) * [AUTHOR](#author) # NAME File::Process - process text files with custom handlers # SYNOPSIS use File::Process; my ($lines, $info) = process_file($file, process => sub { my ($fh, $lines, $args, $line) = @_; return uc $line; }); # DESCRIPTION Many scripts need to process one or more text files. The boiler-plate usually looks something like: open my $fh, '<', $file or croak "blah blah blah...\n"; while (<$fh> ) { # do something... } close $fh or croak "blah blah blah...\n"; The _do something..._ part often involves other common operations like removing new lines, skipping blank lines, etc. It gets tedious when you have to write the same template for processing different files in a script. This class provides a simple harness for processing files, taking the drudgery out of writing a simple text processor. It is most effect when used on relatively small files (see ["CAVEATS"](#caveats)). In it's most basic form the class will return all of the lines in a text file. The class exports one method (`process_file`) which invokes multiple subroutines that you can override or use in conjunction with your custom processors. _See `File::Process::Utils` for additional recipes._ # EXPORTED METHODS This module exports one method by default (`process_file`), since presumably you wanted to _process_ a file? You can export all of the default processor methods using the tag ':all'. use File::Process qw( pre post ); use File::Process qw( :all ); # METHODS AND SUBROUTINES ## process\_csv process_csv(file, options) See [File::Process::Utils](https://metacpan.org/pod/File%3A%3AProcess%3A%3AUtils) ## process\_file process_file(file, options) You start the processing of the file by calling `process_file` with the name of the file or a handle to an open file and a **list** of options. Note that the processors will pass and receive a **reference** to this list of options during the processing of the file. The method returns a list containing a reference to an array that contains each line of the file followed by the list of elements in the hash that was originally passed to it (along with any other data your custom method has inserted into it). my ($lines, %options) = process_file("foo.txt", chomp => 1); - file Path to the file to be processed or a handle to an open file. - options A **list** of options. You can send whatever options your custom processor supports. Before the default or your custom `process` subroutine is called, the `filter` subroutine is called. This is where you might massage the input in some way. The default `filter` subroutine supports various options to perform routine tasks. Options are described below. - skip\_blank\_lines Skip _blank lines_. A blank line is considered a line with only a new line character. - skip\_comments Set `skip_lines` to a true value to skip lines that beging with '#'. - merge\_lines Merges lines together rather that creating an array of lines. Typically used with the `chomp` option. When `merge_lines` is set to a true value, `IO::Scalar` is used to efficiently create a single scalar from all of the lines in the file. The first element of the return list then a scalar instead of an array reference. - chomp Set `chomp` to a true value to remove a trailing new line. - trim Set `trim` to one of _front_, _back_, _both_ to remove whitespace from the front, back or both front and back of a line. Note that this operation is performed _before_ your custom processor is called and may result in the line being skipped if the `skip_blank_lines` option is set to a true value. ## Custom Processors `process_file` will execute a set of subroutines for each line of the file. You can replace any of these subroutines to inject your own custom behaviors. They are executed in this order: - 1. `pre` - 2. `next_line` - 3. `filter` - 4. `process` - 5. `post` - You can terminate processing of a file by returning an undefined value for the `next_line()` hook - Returning undef for the `filter()`, and `process()` hooks will prevent a line from being accumulated in the line buffer - Any hook you define can terminate the process by throwing an exception. The default processors are described below. ### pre(file, options) The default `pre` processor opens the file and returns a file handle and a reference to an array that will be used to store the lines. If you provide your own `pre` process it should also return a tuple that contains the file handle and a reference to an array that will be used to store each processed line of the file. _Note that you don't have to adhere to this contract if your downstream processors don't require the same returns._ - file Path to a file that can be opened for reading or a handle to an open file. - options A reference to a hash that contains the options passed from `process_file`. The hash will be passed to the `process` method, so can be used to store data as you are processing the each line. The default `process` method will record counts of lines processed and other potentially useful statistics. ### next\_line(fh, lines, options) The `next_line` method is passed the file handle, the buffer of accumulated lines, and a reference to a hash of options passed to `process_file`. It is expected to return the _next line_ of the file, however your custom processor can return anything it likes. That object returned will be sent to the `process` subroutine for possible further processing. Returning `undef` will halt further processing. ### filter(fh, lines, options, current\_line) The default `filter` method will perform various tasks (chomp, trim, skip) controlled by the options described above. If the `chomp` option is set to true when you called `process_file`, the line will be chomped. You can also set the `skip_blank_lines` or `skip_comments` to skip blank lines or skip lines that begin with '# '. Filters should return the line or `undef` to skip the current line. If you really want to add `undef` to your buffer, do so in your filter: push @{$lines}, undef; If you want to halt processing here, `die` in your filter. Any exception will halt further processing. ### process(fh, lines, options, current\_line) The `process` method is passed the file handle, the buffer of accumulated lines, a reference to a hash of options passed to `process_file` and the next line of the the text file. The default processor simply returns the `current_line` value. - fh Handle to an open file or an object that supports an `IO::Handle` like interface. If `fh` is undefined the - lines A reference to an array that contains the lines read thus far. - options A reference to a hash of options passed to `process_file`. - current\_line The next line of data from the file. ### post(fh, lines, options) The `post` method is passed the same three arguments as passed to `process`. The default `post` method closes the file and records the end time of process. The default `post` method returns an array reference to the buffer of lines and list of options. Note that a reference to the list is passed in but a **list** is returned. This is also the return value of `process_file`. Your custon post can return anything it wants. # DEFAULT PROCESSORS Any of default processors (**pre**, **next\_line**, **filter**, **process**, **post**) can be called before or after your custom processors. Pass these methods the same list you receive. process_file( "foo.txt", post => sub { my @retval = post(@_); $retval[0] = join '', @{ $_[1] }; return @retval; } ); # STATISTICS - start\_time - end\_time - raw\_count - skipped # EXAMPLES - Return the all of the lines in a text file my ($lines) = process_file('foo.txt'); - Read JSON file print Dumper( process_file( $fh, chomp => 1, post => sub { post(@_); return decode_json( join '', @{ $_[1] } ); } ) ); ...or print Dumper( decode_json( process_file( $fh, chomp => 1, merge_lines => 1 ) ) ); - Read CSV file Presented here as example, however you can use the ["process\_csv"](#process_csv) method for processing CSV files. use File::Process qw(pre process_file); use Text::CSV_XS; use Data::Dumper; my $csv = Text::CSV_XS->new; my $file = shift; my ($csv_lines) = process_file( $file, csv => $csv, chomp => 1, has_headers => 1, pre => sub { my ( $fh, $args ) = @_; my ($fh, $all_lines) = pre($file, $args); if ( $args->{'has_headers'} ) { my @column_names = $args->{csv}->getline($fh); $args->{csv}->column_names(@column_names); } return ($fh, $all_lines); }, next_line => sub { my ( $fh, $all_lines, $args ) = @_; my $ref = $args->{csv}->getline_hr($fh); return $ref; } ) # CAVEATS Processing each line using hooks and callbacks can introduce inefficiencies in file processing. This class is meant to be used on moderately sized files. In it's basic forms, the methods will read all lines into memory as it iterates over the file. Your processing may not require that lines be accumulated at all. Your custom `process()` or `filter()` hook can choose to return an undefined value which prevents a line from being added to the buffer. Reading each line one-at-a-time may be inefficient as well, future version may introduce a slurp mode and/or the ability to send an array which represents a list of lines to process. Some example times: Timings were done a Linux system running on an _11th Gen Intel(R) Core(TM) i7-1160G7 @ 1.20GHz (8 threads, 4.40GHz)_ As a baseline: - Reading ~900K rows (pure Perl) .22s - Slurping ~900K rows (pure Perl) .03s Using `File::Process::process_file()` - Reading ~900K rows (no processing) ~1.6s - Reading ~900K rows from a CSV file with 5 columns - Returning an array of hashes: 7-8s - Returning an array of arrays: ~10s ...so there's room for improving the speed of these calls...caveat emptor. # LICENSE This module is free software. It may be used, redistributed and/or modified under the same terms as Perl itself. # SEE ALSO [File::Process::Utils](https://metacpan.org/pod/File%3A%3AProcess%3A%3AUtils) # AUTHOR Rob Lauer -