<?php
/**
 * Minimal front matter parser (a safe subset of YAML).
 *
 * Supported: "key: value", quoted values, inline lists [a, "b c"], block lists
 * ("- item"), comment lines starting with "#". Not supported: nested maps,
 * multi-line scalars. No code is ever evaluated.
 *
 * @package Draftdrop
 */

if ( ! defined( 'ABSPATH' ) ) {
	exit;
}

/**
 * Splits a Markdown file into front matter and body.
 */
class Draftdrop_Frontmatter {

	/**
	 * Parse a raw file.
	 *
	 * @param string $raw File contents (UTF-8).
	 * @return array{meta: array, body: string, errors: string[]}
	 */
	public static function parse( $raw ) {
		$raw = (string) $raw;
		if ( 0 === strncmp( $raw, "\xEF\xBB\xBF", 3 ) ) {
			$raw = substr( $raw, 3 );
		}
		$raw = str_replace( array( "\r\n", "\r" ), "\n", $raw );

		$out = array(
			'meta'   => array(),
			'body'   => $raw,
			'errors' => array(),
		);

		if ( ! preg_match( '/\A---[ \t]*\n(.*?)\n---[ \t]*(?:\n|\z)(.*)\z/s', $raw, $m ) ) {
			$out['errors'][] = 'no_front_matter';
			return $out;
		}

		$out['body'] = $m[2];
		$list_key    = null;

		foreach ( explode( "\n", $m[1] ) as $line ) {
			$trim = trim( $line );
			if ( '' === $trim || '#' === $trim[0] ) {
				continue;
			}

			if ( preg_match( '/^([A-Za-z_][A-Za-z0-9_-]*)[ \t]*:(?:[ \t]+(.*))?$/', $line, $kv ) ) {
				$key = strtolower( str_replace( '-', '_', $kv[1] ) );
				$val = isset( $kv[2] ) ? trim( $kv[2] ) : '';
				if ( '' === $val ) {
					$out['meta'][ $key ] = '';
					$list_key            = $key;
				} else {
					$out['meta'][ $key ] = self::value( $val );
					$list_key            = null;
				}
			} elseif ( null !== $list_key && preg_match( '/^[ \t]*-[ \t]+(.*)$/', $line, $li ) ) {
				if ( ! is_array( $out['meta'][ $list_key ] ) ) {
					$out['meta'][ $list_key ] = array();
				}
				$out['meta'][ $list_key ][] = self::unquote( trim( $li[1] ) );
			} else {
				$out['errors'][] = 'bad_line';
			}
		}

		return $out;
	}

	/**
	 * Parse a scalar or an inline list.
	 *
	 * @param string $val Raw value.
	 * @return string|string[]
	 */
	private static function value( $val ) {
		if ( strlen( $val ) >= 2 && '[' === $val[0] && ']' === substr( $val, -1 ) ) {
			return self::split_list( substr( $val, 1, -1 ) );
		}
		return self::unquote( $val );
	}

	/**
	 * Split "a, 'b, c', d" into items, respecting quotes.
	 *
	 * @param string $inner List contents without brackets.
	 * @return string[]
	 */
	private static function split_list( $inner ) {
		$items = array();
		$cur   = '';
		$quote = '';
		$len   = strlen( $inner );

		for ( $i = 0; $i < $len; $i++ ) {
			$ch = $inner[ $i ];
			if ( '' !== $quote ) {
				if ( '\\' === $ch && '"' === $quote && $i + 1 < $len ) {
					$cur .= $ch . $inner[ ++$i ];
					continue;
				}
				if ( $ch === $quote ) {
					$quote = '';
				}
				$cur .= $ch;
				continue;
			}
			// A quote only opens a quoted item at the start of the item (so "driver's" is safe).
			if ( ( '"' === $ch || "'" === $ch ) && '' === trim( $cur ) ) {
				$quote = $ch;
				$cur  .= $ch;
				continue;
			}
			if ( ',' === $ch ) {
				$items[] = $cur;
				$cur     = '';
				continue;
			}
			$cur .= $ch;
		}
		$items[] = $cur;

		$out = array();
		foreach ( $items as $item ) {
			$item = self::unquote( trim( $item ) );
			if ( '' !== $item ) {
				$out[] = $item;
			}
		}
		return $out;
	}

	/**
	 * Remove matching surrounding quotes.
	 *
	 * @param string $s Value.
	 * @return string
	 */
	private static function unquote( $s ) {
		$n = strlen( $s );
		if ( $n >= 2 ) {
			$first = $s[0];
			$last  = $s[ $n - 1 ];
			if ( '"' === $first && '"' === $last ) {
				return str_replace( array( '\\"', '\\\\' ), array( '"', '\\' ), substr( $s, 1, -1 ) );
			}
			if ( "'" === $first && "'" === $last ) {
				return str_replace( "''", "'", substr( $s, 1, -1 ) );
			}
		}
		return $s;
	}
}