Buffered Stream to Stream copy with transforms

readme.md

Stream-to-Stream Transformer

This class is a simple utility to perform string transformations while copying from stream-to-stream. It has been tested on a 1.48GiB malformed XML document to correct some syntax errors, where find and replace in Sublime Text would fail.

Usage is simple:

$transformer = new TransformingStream(STDIN, STDOUT);
$transformer->transform('/foo(bar)?/', 'baz$1');
$transformer->process();

This class is more suited to CLI usage than web, due to the possible runtime of text transformation over very large files.

TransformingStream.php

<?php

class TransformingStream {

    protected $input;
    protected $output;

    // Chunk size...
    protected $chunkSize = 4096;

    // Find/replace operations
    protected $search = [];
    protected $replace = [];

    /** @var callable|null */
    protected $progress;

    /**
     * TransformingStream constructor.
     *
     * @param resource $input
     * @param resource $output
     */
    public function __construct($input, $output)
    {
        $this->input  = $input;
        $this->output = $output;
    }

    /**
     * @return int
     */
    public function getChunkSize(): int
    {
        return $this->chunkSize;
    }

    /**
     * @param int $chunkSize
     * @return \TransformingStream
     */
    public function setChunkSize(int $chunkSize): self
    {
        $this->chunkSize = $chunkSize;

        return $this;
    }

    /**
     * @param callable $progress The onProgress callback
     */
    public function setProgress(callable $progress)
    {
        $this->progress = $progress;
    }

    public function transform(string $pattern, string $replace): self
    {
        $this->search[]  = $pattern;
        $this->replace[] = $replace;

        return $this;
    }

    public function process()
    {
        // Early read to fill to twice the buffer size, if possible
        $buffer = fread($this->input, $this->chunkSize);
        $chunk  = 1;
        $cb     = $this->progress;

        do {
            $buffer .= fread($this->input, $this->chunkSize);

            // Apply transforms to the whole buffer to fix any operations skipped by
            // boundaries of the buffer chunks.
            $buffer = preg_replace($this->search, $this->replace, $buffer);

            // pull off the output chunk...
            fwrite($this->output, substr($buffer, 0, $this->chunkSize));
            $buffer = substr($buffer, $this->chunkSize);

            if ($cb) {
                $cb($chunk);
            }

            $chunk++;
        } while (!feof($this->input));

        if ($buffer) {
            fwrite($this->output, $buffer);
        }
    }

}