Add basic variable transformations

.githooks/pre-commit

@@ -0,0 +1,28 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+repo_root="$(git rev-parse --show-toplevel)"
+cd "$repo_root"
+
+if [[ ! -x vendor/bin/php-cs-fixer || ! -x vendor/bin/phpunit || ! -x vendor/bin/phpstan ]]; then
+    echo "Pre-commit checks require Composer dependencies. Run: composer install" >&2
+    exit 1
+fi
+
+if [[ -z "${OPENSTATSPEC_SPECIFICATION_DIR:-}" && -d "$repo_root/../specification" ]]; then
+    export OPENSTATSPEC_SPECIFICATION_DIR="$repo_root/../specification"
+fi
+
+check_dir="$(mktemp -d "${TMPDIR:-/tmp}/openstatspec-php-check.XXXXXX")"
+cleanup() {
+    rm -rf "$check_dir"
+}
+trap cleanup EXIT
+
+staged_tree="$(git write-tree)"
+git -c core.autocrlf=false archive "$staged_tree" | tar -x -C "$check_dir"
+mkdir "$check_dir/vendor"
+cp -a "$repo_root/vendor/." "$check_dir/vendor/"
+
+cd "$check_dir"
+composer check

README.md

@@ -187,6 +187,12 @@ composer install
 composer check
 ```
 
+Install the tracked pre-commit hook once per clone:
+
+```bash
+bash tools/install-git-hooks.sh
+```
+
 `composer check` validates Composer configuration, lints PHP, checks style, runs PHPStan and runs PHPUnit. Use `composer fix` for safe style fixes, then rerun `composer check`.
 
 GitHub Actions runs the regular suite on PHP 8.4 and 8.5. It also runs real

src/Frontend/Spss/Ast/DeleteVariablesStatement.php

@@ -0,0 +1,19 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OpenStatSpec\Frontend\Spss\Ast;
+
+final readonly class DeleteVariablesStatement implements Statement
+{
+    /** @param list<string> $variables */
+    public function __construct(
+        public int $lineNumber,
+        public array $variables,
+    ) {}
+
+    public function line(): int
+    {
+        return $this->lineNumber;
+    }
+}

src/Frontend/Spss/Ast/StringStatement.php

@@ -0,0 +1,20 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OpenStatSpec\Frontend\Spss\Ast;
+
+final readonly class StringStatement implements Statement
+{
+    /** @param list<string> $variables */
+    public function __construct(
+        public int $lineNumber,
+        public array $variables,
+        public int $width,
+    ) {}
+
+    public function line(): int
+    {
+        return $this->lineNumber;
+    }
+}

src/Frontend/Spss/Binder.php

@@ -5,8 +5,10 @@
 namespace OpenStatSpec\Frontend\Spss;
 
 use OpenStatSpec\Frontend\Spss\Ast\ElseInput;
+use OpenStatSpec\Frontend\Spss\Ast\DeleteVariablesStatement;
 use OpenStatSpec\Frontend\Spss\Ast\ExecuteStatement;
 use OpenStatSpec\Frontend\Spss\Ast\MissingInput;
+use OpenStatSpec\Frontend\Spss\Ast\StringStatement;
 use OpenStatSpec\Frontend\Spss\Ast\Program;
 use OpenStatSpec\Frontend\Spss\Ast\RangeInput;
 use OpenStatSpec\Frontend\Spss\Ast\RecodeStatement;
@@ -15,6 +17,8 @@
 use OpenStatSpec\Frontend\Spss\Binding\BoundProgram;
 use OpenStatSpec\Frontend\Spss\Binding\BoundRecode;
 use OpenStatSpec\Frontend\Spss\Binding\BoundValueLabels;
+use OpenStatSpec\Frontend\Spss\Binding\BoundCreateVariable;
+use OpenStatSpec\Frontend\Spss\Binding\BoundDeleteVariable;
 use OpenStatSpec\Frontend\Spss\Binding\BoundVariableLabel;
 
 final class Binder
@@ -30,6 +34,18 @@ public function bind(string $datasetId, Program $program): BoundProgram
             if ($statement instanceof ExecuteStatement) {
                 continue;
             }
+            if ($statement instanceof StringStatement) {
+                foreach ($statement->variables as $variable) {
+                    $bound[] = new BoundCreateVariable($variable, 'string', $statement->width);
+                }
+                continue;
+            }
+            if ($statement instanceof DeleteVariablesStatement) {
+                foreach ($statement->variables as $variable) {
+                    $bound[] = new BoundDeleteVariable($variable);
+                }
+                continue;
+            }
             if ($statement instanceof RecodeStatement) {
                 $targets = $statement->targets === [] ? $statement->sources : $statement->targets;
                 if (count($statement->sources) !== count($targets)) {

src/Frontend/Spss/Binding/BoundCreateVariable.php

@@ -0,0 +1,14 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OpenStatSpec\Frontend\Spss\Binding;
+
+final readonly class BoundCreateVariable implements BoundStatement
+{
+    public function __construct(
+        public string $variable,
+        public string $storageKind,
+        public int $declaredStringWidth,
+    ) {}
+}

src/Frontend/Spss/Binding/BoundDeleteVariable.php

@@ -0,0 +1,10 @@
+<?php
+
+declare(strict_types=1);
+
+namespace OpenStatSpec\Frontend\Spss\Binding;
+
+final readonly class BoundDeleteVariable implements BoundStatement
+{
+    public function __construct(public string $variable) {}
+}

src/Frontend/Spss/Compiler.php

@@ -18,7 +18,11 @@
 use OpenStatSpec\Frontend\Spss\Binding\BoundRecode;
 use OpenStatSpec\Frontend\Spss\Binding\BoundValueLabels;
 use OpenStatSpec\Frontend\Spss\Binding\BoundVariableLabel;
+use OpenStatSpec\Frontend\Spss\Binding\BoundCreateVariable;
+use OpenStatSpec\Frontend\Spss\Binding\BoundDeleteVariable;
 use OpenStatSpec\Transformation\Model\Action\AssignValueAction;
+use OpenStatSpec\Transformation\Model\CreateVariableOperation;
+use OpenStatSpec\Transformation\Model\DeleteVariableOperation;
 use OpenStatSpec\Transformation\Model\Action\CopySourceAction;
 use OpenStatSpec\Transformation\Model\Action\SetMissingAction;
 use OpenStatSpec\Transformation\Model\RecodeAction;
@@ -45,6 +49,18 @@ public function compile(BoundProgram $program): TransformationPlan
     {
         $operations = [];
         foreach ($program->statements as $statement) {
+            if ($statement instanceof BoundCreateVariable) {
+                $operations[] = new CreateVariableOperation(
+                    $statement->variable,
+                    $statement->storageKind,
+                    $statement->storageKind === 'string' ? $statement->declaredStringWidth : null,
+                );
+                continue;
+            }
+            if ($statement instanceof BoundDeleteVariable) {
+                $operations[] = new DeleteVariableOperation($statement->variable);
+                continue;
+            }
             if ($statement instanceof BoundRecode) {
                 $rules = [];
                 $hasElse = false;

src/Frontend/Spss/Parser.php

@@ -4,6 +4,7 @@
 
 namespace OpenStatSpec\Frontend\Spss;
 
+use OpenStatSpec\Frontend\Spss\Ast\DeleteVariablesStatement;
 use OpenStatSpec\Frontend\Spss\Ast\ElseInput;
 use OpenStatSpec\Frontend\Spss\Ast\ExecuteStatement;
 use OpenStatSpec\Frontend\Spss\Ast\MissingInput;
@@ -16,6 +17,7 @@
 use OpenStatSpec\Frontend\Spss\Ast\RecodeStatement;
 use OpenStatSpec\Frontend\Spss\Ast\ScalarValue;
 use OpenStatSpec\Frontend\Spss\Ast\SystemMissingInput;
+use OpenStatSpec\Frontend\Spss\Ast\StringStatement;
 use OpenStatSpec\Frontend\Spss\Ast\ValueInput;
 use OpenStatSpec\Frontend\Spss\Ast\ValueLabel;
 use OpenStatSpec\Frontend\Spss\Ast\ValueLabelGroup;
@@ -60,6 +62,11 @@ public function parseTokens(array $tokens): Program
                 $statements[] = $this->valueLabels($command);
             } elseif ($this->matchKeyword('EXECUTE')) {
                 $statements[] = new ExecuteStatement($command->line);
+            } elseif ($this->matchKeyword('STRING')) {
+                $statements[] = $this->string($command);
+            } elseif ($this->matchKeyword('DELETE')) {
+                $this->consumeKeyword('VARIABLES', 'Expected VARIABLES after DELETE.');
+                $statements[] = $this->deleteVariables($command);
             } else {
                 $this->fail($command, sprintf('Unsupported SPSS command %s.', $command->lexeme === '' ? '<end of input>' : $command->lexeme));
             }
@@ -180,6 +187,43 @@ private function valueLabels(Token $command): ValueLabelsStatement
         return new ValueLabelsStatement($command->line, $groups);
     }
 
+    private function string(Token $command): StringStatement
+    {
+        $variables = [];
+        do {
+            $variable = $this->consumeIdentifier('Expected a variable name in STRING.')->lexeme;
+            if (strcasecmp($variable, 'TO') === 0) {
+                $this->fail($this->previous(), 'STRING variable ranges using TO are not supported.');
+            }
+            $variables[] = $variable;
+        } while ($this->check(TokenType::Identifier));
+        $this->consume(TokenType::LeftParenthesis, 'Expected a width declaration in STRING.');
+        $width = $this->consume(TokenType::Identifier, 'Expected a string width such as A20.')->lexeme;
+        if (preg_match('/\AA([1-9][0-9]*)\z/i', $width, $matches) !== 1) {
+            $this->fail($this->previous(), 'STRING width must use the SPSS A<n> form.');
+        }
+        $widthValue = (int) $matches[1];
+        if ($widthValue > 32767) {
+            $this->fail($this->previous(), 'STRING width must be at most 32767.');
+        }
+        $this->consume(TokenType::RightParenthesis, 'Expected ) after STRING width.');
+
+        return new StringStatement($command->line, $variables, $widthValue);
+    }
+
+    private function deleteVariables(Token $command): DeleteVariablesStatement
+    {
+        $variables = [];
+        do {
+            $variable = $this->consumeIdentifier('Expected a variable name in DELETE VARIABLES.')->lexeme;
+            if (strcasecmp($variable, 'TO') === 0) {
+                $this->fail($this->previous(), 'DELETE VARIABLES ranges using TO are not supported.');
+            }
+            $variables[] = $variable;
+        } while ($this->check(TokenType::Identifier));
+
+        return new DeleteVariablesStatement($command->line, $variables);
+    }
     private function scalar(string $message): ScalarValue
     {
         if ($this->match(TokenType::String)) {