Merge pull request #1 from elecena/initial-version

Initial version

.gitignore

@@ -1,2 +1,5 @@
-.idea/
+/.idea/
 /vendor/
+/.php-cs-fixer.cache
+/.phpunit.cache/
+/.phpunit.result.cache

phpunit.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<phpunit xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+         xsi:noNamespaceSchemaLocation="https://schema.phpunit.de/10.3/phpunit.xsd"
+         bootstrap="vendor/autoload.php"
+         cacheDirectory=".phpunit.cache"
+         executionOrder="depends,defects"
+         beStrictAboutCoverageMetadata="true"
+         beStrictAboutOutputDuringTests="true"
+         failOnRisky="true"
+         failOnWarning="true"
+         displayDetailsOnTestsThatTriggerDeprecations="true"
+         displayDetailsOnTestsThatTriggerErrors="true"
+         displayDetailsOnTestsThatTriggerNotices="true"
+         displayDetailsOnTestsThatTriggerWarnings="true"
+    >
+    <testsuites>
+        <testsuite name="default">
+            <directory>tests</directory>
+        </testsuite>
+    </testsuites>
+
+    <source>
+        <include>
+            <directory>src</directory>
+        </include>
+    </source>
+</phpunit>

src/JsonlParser.php

@@ -0,0 +1,106 @@
+<?php
+
+namespace Elecena\JsonlParser;
+
+class JsonlParser implements \Countable
+{
+    const LINES_SEPARATOR = "\n";
+
+    /**
+     * @param resource $stream
+     */
+    public function __construct(protected $stream)
+    {
+    }
+
+    public function push(array|string $item): void
+    {
+        $encoded = json_encode($item);
+        fwrite($this->stream, $encoded . self::LINES_SEPARATOR);
+    }
+
+    public function pushItems(\Iterator $items): void
+    {
+        foreach($items as $item) {
+            $this->push($item);
+        }
+    }
+
+    /**
+     * This method returns the last item from the file and removes it.
+     */
+    public function pop(): null|array|string
+    {
+        /***
+         * Rewind to the end of the file and try to find the last newline
+         *
+         * @see https://www.php.net/manual/en/function.fseek.php
+         */
+        fseek($this->stream, 0, SEEK_END);
+
+        // this stream is now empty
+        if (ftell($this->stream) === 0) {
+            return null;
+        }
+
+        // start reading from the end of the stream in reverse order, byte by byte
+        fseek($this->stream, -1, SEEK_END);
+        $buffer = fread($this->stream, 1);
+
+        while(ftell($this->stream) > 1) {
+            // move two bytes back (one already read and the one before it)
+            fseek($this->stream, -2, SEEK_CUR);
+
+            $char = fread($this->stream, 1);
+            $buffer .= $char;
+
+            if ($char === self::LINES_SEPARATOR) {
+                break;
+            }
+
+            if (ftell($this->stream) === 0) {
+                break;
+            }
+        }
+
+        $buffer = strrev($buffer);
+
+        // truncate the stream and remove the trailing newline
+        $pos = ftell($this->stream);
+        ftruncate($this->stream, $pos < 1 ? 0 : $pos-1);
+
+        return json_decode($buffer, associative: true);
+    }
+
+    /**
+     * @return \Generator<array,string>
+     */
+    public function iterate(): \Generator
+    {
+        while(!is_null($item=$this->pop())) {
+            yield $item;
+        }
+    }
+
+    /**
+     * This method returns how many JSON-encoded lines are in the stream.
+     *
+     * This can be heavy on large files this method rewinds and then reads the entire stream content.
+     *
+     * @return int
+     */
+    public function count(): int
+    {
+        $count = 0;
+        rewind($this->stream);
+
+        /**
+         * https://www.php.net/manual/en/function.stream-get-line.php
+         */
+        while(stream_get_line($this->stream, 1024 * 1024, self::LINES_SEPARATOR) !== false) {
+            $count++;
+        }
+
+        return $count;
+    }
+}

tests/BaseTestCase.php

@@ -0,0 +1,20 @@
+<?php
+
+use PHPUnit\Framework\TestCase;
+
+abstract class BaseTestCase extends TestCase
+{
+    /**
+     * Returns a readable, in-memory stream with the provided string as a content
+     *
+     * @return resource
+     */
+    protected static function streamFromString(string $string)
+    {
+        $stream = fopen('php://memory', 'r+');
+        fwrite($stream, $string);
+        rewind($stream);
+
+        return $stream;
+    }
+}

tests/JsonParserTest.php

@@ -0,0 +1,109 @@
+<?php
+
+use \Elecena\JsonlParser\JsonlParser;
+
+class JsonParserTest extends BaseTestCase
+{
+    const ITEM = ['foo' => 'bar', 'ok' => true];
+    const ITEM_ONE = ['foo' => 'bar', 'ok' => true];
+    const ITEM_TWO = ['foo' => 'test', 'ok' => true];
+
+    public function testOpensAnEmptyString(): void
+    {
+        $stream = self::streamFromString('');
+        $parser = new JsonlParser($stream);
+        $this->assertNull($parser->pop());
+        $this->assertCount(0, $parser);
+    }
+
+    public function testOpensASingleLine(): void
+    {
+        $stream = self::streamFromString(json_encode(self::ITEM));
+        $parser = new JsonlParser($stream);
+        $this->assertCount(1, $parser);
+        $this->assertSame(self::ITEM, $parser->pop());
+        $this->assertCount(0, $parser);
+    }
+    public function testOpensASingleLineWithTrailingNewLine(): void
+    {
+        $stream = self::streamFromString(json_encode(self::ITEM) . JsonlParser::LINES_SEPARATOR);
+        $parser = new JsonlParser($stream);
+        $this->assertCount(1, $parser);
+        $this->assertSame(self::ITEM, $parser->pop());
+        $this->assertCount(0, $parser);
+    }
+    public function testOpensTwoLines(): void
+    {
+        $stream = self::streamFromString(json_encode(self::ITEM_ONE) . JsonlParser::LINES_SEPARATOR . json_encode(self::ITEM_TWO));
+        $parser = new JsonlParser($stream);
+        $this->assertCount(2, $parser);
+        $this->assertSame(self::ITEM_TWO, $parser->pop());
+        $this->assertCount(1, $parser);
+        $this->assertSame(self::ITEM_ONE, $parser->pop());
+        $this->assertCount(0, $parser);
+    }
+
+    public function testOpensAnEmptyStringAndAddsAnItem(): void
+    {
+        $stream = self::streamFromString('');
+        $parser = new JsonlParser($stream);
+        $this->assertCount(0, $parser);
+        $parser->push(self::ITEM);
+        $this->assertCount(1, $parser);
+        $this->assertSame(self::ITEM, $parser->pop());
+        $this->assertCount(0, $parser);
+    }
+
+    public function testHandlesStrings(): void
+    {
+        $item = 'https://foo.bar.net';
+
+        $stream = self::streamFromString('');
+        $parser = new JsonlParser($stream);
+        $this->assertCount(0, $parser);
+        $parser->push($item);
+        $this->assertCount(1, $parser);
+        $this->assertSame($item, $parser->pop());
+        $this->assertCount(0, $parser);
+        $this->assertNull($parser->pop());
+    }
+
+    public function testIterator(): void
+    {
+        $stream = self::streamFromString('');
+        $parser = new JsonlParser($stream);
+
+        $parser->push('one');
+        $parser->push('two');
+        $parser->push('three');
+        $this->assertCount(3, $parser);
+
+        $list = iterator_to_array($parser->iterate());
+
+        $this->assertCount(0, $parser);
+        $this->assertCount(3, $list);
+        $this->assertSame(['three', 'two', 'one'], $list);
+    }
+
+    public function testPushItems(): void
+    {
+        $stream = self::streamFromString('');
+        $parser = new JsonlParser($stream);
+
+        function iterator(): Generator
+        {
+            yield 'one';
+            yield 'two';
+            yield 'three';
+        }
+
+        $parser->pushItems(items:iterator());
+        $this->assertCount(3, $parser);
+
+        $list = iterator_to_array($parser->iterate());
+
+        $this->assertCount(0, $parser);
+        $this->assertCount(3, $list);
+        $this->assertSame(['three', 'two', 'one'], $list);
+    }
+}