README

Industrial-grade, memory-efficient streaming reader for PHP.

Reading gigabytes of text data (logs, SQL dumps, CSVs) in PHP can be tricky. Standard fread breaks multi-byte characters (UTF-8, UTF-16) if a chunk boundary splits a character in half. Sorelvi\StreamReader solves this problem by strictly analyzing byte sequences and ensuring every yielded chunk contains only valid, complete characters.

Key Features

• Memory Efficient: Uses Generators (yield) to process files of any size (even terabytes) with constant O(1) memory usage.
• Multibyte Safe: Guaranteed character integrity. Never splits a UTF-8 emoji or a UTF-16 surrogate pair across chunks.
• Zero Dependencies: Works on pure PHP (no mbstring required).
• Polyglot: Built-in support via Preset enums:
  • UTF-8 (optimized bitwise scanning)
  • UTF-16 (LE/BE auto-detection via BOM or manual config)
  • UTF-32
  • Fixed-byte encodings (ASCII, Windows-1251, ISO-8859-1 via Preset::BYTE1)
  • Fixed-width protocols (Preset::BYTE2 through Preset::BYTE10)
• Network/Pipe Ready: Built-in retry logic handles slow or non-blocking network streams.
• Resumable: Architecture supports pausing and resuming reading via HandleContext.
• Extensible: Full support for custom StreamInterface and EstimatorInterface implementations.

Requirements

• PHP 8.2 or higher

Installation

composer require sorelvi/stream-reader

Quick Start

Reading a Large File

The easiest way to read a file is using the static factory method with a Preset.

use Sorelvi\StreamReader\Reader;
use Sorelvi\StreamReader\Enum\Preset;
use Sorelvi\StreamReader\Exception\StreamReaderException;

try {
    // 1. Create a reader for a UTF-8 file
    $reader = Reader::createForFile('/path/to/huge_dump.sql');

    // Optional: Adjust chunk size (default is 64KB)
    $reader->setChunkLength(8192); // 8KB

    // 2. Iterate cleanly
    foreach ($reader->readChunk() as $chunk) {
        // $chunk is guaranteed to end on a valid character boundary
        // e.g. send to database or parse CSV line
        process_data($chunk);
    }
} catch (StreamReaderException $e) {
    // Handles IO errors, corrupted streams, and encoding violations
    error_log('[' . $e->getCode() . '] ' . $e->getMessage());
}

Reading from String

use Sorelvi\StreamReader\Reader;

$sql = "INSERT INTO ... 🚀"; // String with emojis
$reader = Reader::createForString($sql);

foreach ($reader->readChunk() as $chunk) {
    echo $chunk;
}

Per-call Chunk Size

The readChunk() method accepts an optional chunk size that overrides the instance default for that single call:

$reader = Reader::createForString('Hello World');
$reader->setChunkLength(2); // default: 2 bytes

$first  = $reader->readChunk(1)->current(); // 'H'  — one-off override
$second = $reader->readChunk()->current();  // 'el' — uses default 2

Integrity vs. Validation

Important: This library focuses on Stream Integrity, not Content Validation.

• What it DOES: It ensures that chunks are not split in the middle of a byte sequence defined by the encoding (e.g., it won't cut a 4-byte Emoji in half).
• What it does NOT: It does not validate if the bytes actually represent valid characters in that encoding. If you read a binary file as UTF-8, the reader will still process it without error, ensuring "structural" completeness based on UTF-8 bit-mask rules, even if the resulting output is garbage.

Garbage In → Complete Garbage Out (not Half-Garbage).

Advanced Usage

Dependency Injection (Manual Construction)

For strict control over the estimator or context, you can instantiate the class manually.

use Sorelvi\StreamReader\Reader;
use Sorelvi\StreamReader\StreamCompletenessEstimatorFactory;
use Sorelvi\StreamReader\Enum\Preset;
use Sorelvi\StreamReader\HandleContext;

// 1. Prepare resources
$inputFile = fopen('data.csv', 'rb');
$context = new HandleContext();

// 2. Create the specific estimator
$estimator = StreamCompletenessEstimatorFactory::create(Preset::UTF16LE);

// 3. Instantiate Reader with explicit dependencies
$reader = new Reader($inputFile, $estimator, $context);

foreach ($reader->readChunk() as $chunk) {
    // ...
}

Resume Reading (Context Management)

The HandleContext tracks the number of bytes read. You can persist this state to resume reading later (e.g., between HTTP requests or background jobs).

use Sorelvi\StreamReader\Reader;
use Sorelvi\StreamReader\HandleContext;
use Sorelvi\StreamReader\Enum\Preset;

// Step 1: Read some data
$context = new HandleContext();
$reader = Reader::createForFile('large.log', $context, Preset::UTF16);

foreach ($reader->readChunk() as $chunk) {
    process($chunk);
    if (should_pause()) {
        break;
    }
}

// Step 2: Save context
$state = $context->toArray();
save_to_db($state);

// --- Later ---

// Step 3: Resume
$newContext = HandleContext::fromArray(load_from_db());

// The reader will automatically seek to the correct position
$reader = Reader::createForFile('large.log', $newContext, Preset::UTF16);

HandleContext API

Supporting UTF-16

The reader intelligently handles UTF-16 endianness:

1. Uses the explicitly configured endianness if Preset::UTF16LE or Preset::UTF16BE is specified.
2. For Preset::UTF16 (auto mode), detects endianness from the Byte Order Mark (BOM): 0xFF 0xFE → LE, anything else → BE (per RFC 2781).
3. Correctly handles Surrogate Pairs (4-byte characters in UTF-16) to ensure they are not split across chunks.
4. The detected endianness is stored in HandleContext and reused on subsequent calls, ensuring consistency across resumed sessions.

// Auto-detect BOM
$reader = Reader::createForFile('export_utf16.csv', null, Preset::UTF16);

// Or specify explicitly
$reader = Reader::createForFile('export_utf16le.csv', null, Preset::UTF16LE);

Network and Pipe Streams

When working with non-blocking or slow network streams, fread may return an empty string without reaching EOF. The Stream class handles this transparently with configurable retry logic.

use Sorelvi\StreamReader\Stream;
use Sorelvi\StreamReader\Reader;
use Sorelvi\StreamReader\StreamCompletenessEstimatorFactory;
use Sorelvi\StreamReader\Enum\Preset;

$socket = fsockopen('tcp://example.com', 9000);

$stream = new Stream($socket);
$stream->setMaxEmptyAttempts(50);          // max retries on empty reads (default: 20)
$stream->setAttemptsDelayMicroseconds(5000); // delay between retries in µs (default: 10000)

$estimator = StreamCompletenessEstimatorFactory::create(Preset::UTF8);
$reader = new Reader($stream, $estimator);

foreach ($reader->readChunk() as $chunk) {
    // ...
}

If the stream returns empty data more times than maxEmptyAttempts, a TooManyEmptyAttempts exception is thrown.

Custom Estimator

You can implement EstimatorInterface to support any custom encoding or fixed-width binary protocol.

use Sorelvi\StreamReader\EstimatorInterface;
use Sorelvi\StreamReader\HandleContext;
use Sorelvi\StreamReader\Reader;

class MyProtocolEstimator implements EstimatorInterface
{
    public function handle(string $buffer, HandleContext $context): int
    {
        // Return 0 if the buffer ends on a complete unit,
        // or N > 0 to request N more bytes before yielding.
        $remainder = strlen($buffer) % 7; // 7-byte fixed records
        return $remainder ? 7 - $remainder : 0;
    }

    public function getMaxAddReadBytes(): int
    {
        return 6; // max bytes ever requested in one handle() call
    }
}

$reader = Reader::createForFile('records.bin', null, new MyProtocolEstimator());

Custom Stream

You can implement StreamInterface to wrap any data source (e.g., in-memory buffers, S3 streams, database BLOBs).

use Sorelvi\StreamReader\StreamInterface;

class MyCustomStream implements StreamInterface
{
    public function isEndOfStream(): bool { /* ... */ }
    public function read(int $length): string { /* ... */ }
    public function skip(int $offset): void { /* ... */ }
    public function getCurrentPosition(): int { /* ... */ }
}

$reader = new Reader(new MyCustomStream(), $estimator);

Supported Presets

Use the Sorelvi\StreamReader\Enum\Preset enum to select your encoding:

Exception Reference

All library exceptions extend Sorelvi\StreamReader\Exception\StreamReaderException.

ErrorCode Enum

Sorelvi\StreamReader\Exception\ErrorCode provides integer codes for programmatic error handling:

use Sorelvi\StreamReader\Exception\StreamDamaged;
use Sorelvi\StreamReader\Exception\ErrorCode;

try {
    foreach ($reader->readChunk() as $chunk) { /* ... */ }
} catch (StreamDamaged $e) {
    match ($e->getCode()) {
        ErrorCode::INCOMPLETE_SEQUENCE_AT_EOF->value  => handleTruncated(),
        ErrorCode::ORPHAN_CONTINUATION_BYTE->value    => handleCorrupted(),
        ErrorCode::INVALID_START_BYTE_DETECTED->value => handleInvalid(),
        default => throw $e,
    };
}

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please ensure to run static analysis tools (composer check) before submitting.

See CONTRIBUTING.md for full development workflow.

License

Licensed under the Apache License, Version 2.0.

Copyright 2026 Sorelvi.