AVRO-4190: [php] Add logical types (#3526)

* wip

* update logical types

* wip - added logical types

* update PHP workflow

* add composer.json to workflow path

* bump composer version

* add license to AvroLogicalType.php

* add test on logical types

* update test and fix default decimal value

* improve exception messages

* fix lint

* improve types

* add test for decimal schema validation

* add test for decimal schema validation

* remove is_numeric

* add some test for decimal logical type (bytes)

* Add AvroDuration type and extend fixed schema

* additional tests for decimal

Author: mattiabasone

.github/workflows/test-lang-php.yml

@@ -22,6 +22,7 @@ on:
     branches: [ main ]
     paths:
     - .github/workflows/test-lang-php.yml
+    - composer.json
     - lang/php/**
 
 defaults:
@@ -44,6 +45,10 @@ jobs:
           - '7.3'
           - '7.4'
           - '8.0'
+          - '8.1'
+          - '8.2'
+          - '8.3'
+          - '8.4'
 
     steps:
       - uses: actions/checkout@v5
@@ -52,7 +57,7 @@ jobs:
         uses: shivammathur/setup-php@v2
         with:
           php-version: ${{ matrix.php }}
-          tools: composer:2.2.5
+          tools: composer:2.8.12
 
       - name: Get Composer Cache Directory
         id: composer-cache
@@ -82,6 +87,10 @@ jobs:
           - '7.3'
           - '7.4'
           - '8.0'
+          - '8.1'
+          - '8.2'
+          - '8.3'
+          - '8.4'
 
     steps:
       - uses: actions/checkout@v5
@@ -90,7 +99,7 @@ jobs:
         uses: shivammathur/setup-php@v2
         with:
           php-version: ${{ matrix.php }}
-          tools: composer:2.2.5
+          tools: composer:2.8.12
 
       - name: Cache Local Maven Repository
         uses: actions/cache@v4

composer.json

@@ -23,7 +23,7 @@
     "issues": "https://issues.apache.org/jira/browse/AVRO"
   },
   "require": {
-    "php": "^7.1 || ^8.0"
+    "php": "^7.3 || ^8.0"
   },
   "deps": [
     "vendor/phpunit/phpunit",

lang/php/lib/DataFile/AvroDataIOWriter.php

@@ -136,7 +136,7 @@ private static function generateSyncMarker()
     /**
      * Writes the header of the AvroIO object container
      */
-    private function writeHeader()
+    private function writeHeader(): void
     {
         $this->write(AvroDataIO::magic());
         $this->datum_writer->writeData(

lang/php/lib/Datum/AvroIOBinaryEncoder.php

@@ -21,6 +21,7 @@
 namespace Apache\Avro\Datum;
 
 use Apache\Avro\Avro;
+use Apache\Avro\AvroException;
 use Apache\Avro\AvroGMP;
 use Apache\Avro\AvroIO;
 
@@ -41,7 +42,7 @@ class AvroIOBinaryEncoder
      * @param AvroIO $io object to which data is to be written.
      *
      */
-    public function __construct($io)
+    public function __construct(AvroIO $io)
     {
         Avro::checkPlatform();
         $this->io = $io;
@@ -50,9 +51,9 @@ public function __construct($io)
     /**
      * @param null $datum actual value is ignored
      */
-    public function writeNull($datum)
+    public function writeNull($datum): void
     {
-        return null;
+        return;
     }
 
     /**
@@ -67,23 +68,23 @@ public function writeBoolean($datum)
     /**
      * @param string $datum
      */
-    public function write($datum)
+    public function write($datum): void
     {
         $this->io->write($datum);
     }
 
     /**
      * @param int $datum
      */
-    public function writeInt($datum)
+    public function writeInt($datum): void
     {
         $this->writeLong($datum);
     }
 
     /**
      * @param int $n
      */
-    public function writeLong($n)
+    public function writeLong($n): void
     {
         if (Avro::usesGmp()) {
             $this->write(AvroGMP::encodeLong($n));
@@ -125,7 +126,7 @@ public static function encodeLong($n)
      * @param float $datum
      * @uses self::floatToIntBits()
      */
-    public function writeFloat($datum)
+    public function writeFloat($datum): void
     {
         $this->write(self::floatToIntBits($datum));
     }
@@ -142,7 +143,7 @@ public function writeFloat($datum)
      * @returns string bytes
      * @see Avro::checkPlatform()
      */
-    public static function floatToIntBits($float)
+    public static function floatToIntBits($float): string
     {
         return pack('g', (float) $float);
     }
@@ -151,7 +152,7 @@ public static function floatToIntBits($float)
      * @param float $datum
      * @uses self::doubleToLongBits()
      */
-    public function writeDouble($datum)
+    public function writeDouble($datum): void
     {
         $this->write(self::doubleToLongBits($datum));
     }
@@ -165,7 +166,7 @@ public function writeDouble($datum)
      * @param double $double
      * @returns string bytes
      */
-    public static function doubleToLongBits($double)
+    public static function doubleToLongBits($double): string
     {
         return pack('e', (double) $double);
     }
@@ -174,17 +175,64 @@ public static function doubleToLongBits($double)
      * @param string $str
      * @uses self::writeBytes()
      */
-    public function writeString($str)
+    public function writeString($str): void
     {
         $this->writeBytes($str);
     }
 
     /**
      * @param string $bytes
      */
-    public function writeBytes($bytes)
+    public function writeBytes($bytes): void
     {
         $this->writeLong(strlen($bytes));
         $this->write($bytes);
     }
+
+    public function writeDecimal($decimal, int $scale, int $precision): void
+    {
+        if (!is_numeric($decimal)) {
+            throw new AvroException("Decimal value '{$decimal}' must be numeric");
+        }
+
+        $value = $decimal * (10 ** $scale);
+        if (!is_int($value)) {
+            $value = (int) round($value);
+        }
+
+        $maxValue = 10 ** $precision;
+        if (abs($value) >= $maxValue) {
+            throw new AvroException(
+                "Decimal value '{$decimal}' is out of range for precision={$precision}, scale={$scale}"
+            );
+        }
+
+        $packed = pack('J', $value);
+
+        $significantBit = self::getMostSignificantBitAt($packed, 0);
+        $trimByte = $significantBit ? 0xff : 0x00;
+
+        $offset = 0;
+        $packedLength = strlen($packed);
+        while ($offset < $packedLength - 1) {
+            if (ord($packed[$offset]) !== $trimByte) {
+                break;
+            }
+
+            if (self::getMostSignificantBitAt($packed, $offset + 1) !== $significantBit) {
+                break;
+            }
+
+            $offset++;
+        }
+
+        $value = substr($packed, $offset);
+
+        $this->writeBytes($value);
+    }
+
+    private static function getMostSignificantBitAt($bytes, $offset): int
+    {
+        return ord($bytes[$offset]) & 0x80;
+    }
 }