axlint

#!/usr/bin/env php
<?php

function msg_log($msg){
	global $log_handle;
	if (!isset($log_handle)){
		$log_handle = fopen('axlint.log', 'a');
	}
	fwrite($log_handle, $msg . PHP_EOL);
}

/**
 * Walk a tree, applying an action to every object.
 * Returns a modified tree.
 * From https://github.com/Vinai/pandocfilters-php/blob/master/pandocfilters.php
 */
function ast_walk($value, $action, $format, $meta)
{
	if (is_array($value)) {
		$array = array();
		foreach ($value as $item) {
			if (is_object($item) && isset($item->t)) {
				$result = $action($item->t, $item->c, $format, $meta);
				if (is_null($result)) {
					$array[] = ast_walk($item, $action, $format, $meta);
				} elseif (is_array($result)) {
					foreach ($result as $res_value) {
						$array[] = ast_walk($res_value, $action, $format, $meta);
					}
				} elseif (is_object($result)) {
					$array[] = ast_walk($result, $action, $format, $meta);
				} else {
					$obj = clone $item;
					$obj->c = (string)$result;
					$array[] = $obj;
				}
				continue;
			}
			
			$array[] = ast_walk($item, $action, $format, $meta);
		}
		return $array;
	}
		
	if (is_object($value)) {
		$obj = clone $value;
		foreach (get_object_vars($value) as $prop => $prop_value) {
			$obj->{$prop} = ast_walk($prop_value, $action, $format, $meta);
		}
		return $obj;
	}
	
	return $value;
}

function html_get_title($filename) {
	$html = file_get_contents($filename);
	if (!$html) 
		return null;

	$res = preg_match("/<title>(.*)<\/title>/siU", $html, $matches);
	if (!$res) 
		return null; 

	$title = html_entity_decode($matches[1], ENT_COMPAT | ENT_QUOTES, 'UTF-8');
	// Clean up title: remove EOL's and excessive whitespace.
	$title = preg_replace('/\s+/', ' ', $title);
	$title = trim($title);
	return $title;
}

/*
The argument is a list of functions action(key, value, format, meta),
where key is the type of the pandoc object (e.g. 'Str', 'Para'),
value is the contents of the object (e.g. a string for 'Str',
a list of inline elements for 'Para'), format is the target
output format (which will be taken for the first command line
argument if present), and meta is the document's metadata.
If the function returns null, the object to which it applies
will remain unchanged.  If it returns an object, the object will
be replaced.	If it returns a list, the list will be spliced in to
the list to which the target object belongs. So, returning an
empty list deletes the object.
*/

function link_filter($type, $value, $format, $meta){
	if($type !== 'Link'){
		return null;
	}
	
	$url = $value[1][0];
	// If title already defined
	if(count($value[0]) != 1 || $value[0][0]->c != $url){
		return null;
	}
	
	$title = html_get_title($url);
	if(is_null($title)){
		$title = $url;
	}
	$value[0] = array((object)array('t' => 'Str', 'c' => $title));
	
	return (object)array('t' => 'Link', 'c' => $value);
}

// echo -e "Text with link to http://google.com homepage.\n- test 1\n- test 2\n- test 3\n\n## test code\n\tfunction hello(){\n\t\techo 'hello';\n\t}" | ./axlint
// http://pandoc.org/README.html
// http://pandoc.org/scripting.html
function main(){
	/*
	if (count($GLOBALS['argv']) > 1) {
		$file = $GLOBALS['argv'][1];
		if(!is_readable($file) || !is_file($file)){
			fwrite(STDERR, sprintf('%d is not a readable file' . PHP_EOL, $file));
			exit(1);
		}
		if(!is_writable($file)){
			fwrite(STDERR, sprintf('%d is not a readable file' . PHP_EOL, $file));
			exit(1);
		}
	} else {
		$file = 'php://stdin';
	}
	file_get_contents($file);
	*/
	
	$format = 'markdown_github';
	$input_args = '--normalize --preserve-tabs';
	$output_args = '--atx-headers';
	/*
	markdown_github (GitHub-Flavored Markdown):
	- pipe_tables
	- raw_html
	- fenced_code_blocks
	- auto_identifiers
	- ascii_identifiers
	- backtick_code_blocks
	- autolink_bare_uris
	- intraword_underscores
	- strikeout
	- hard_line_breaks
	- emoji
	- shortcut_reference_links
	*/
	
	// FORMAT to JSON AST
	$cmd = sprintf('pandoc -f %s -t json %s', $format, $input_args);
	$descriptorspec = array(
		0 => STDIN/* = 'php://stdin' *//* array('pipe', 'r') */, // stdin
		1 => array('pipe', 'w'), // stdout
		2 => array('pipe', 'a') // stderr
	);
	$process = proc_open($cmd, $descriptorspec, $pipes);
	
	// proc_open fail
	if (!is_resource($process)) {
		$last_error = error_get_last();
		fwrite(STDERR, 'Unable exec pandoc: ' . $last_error['message'] . PHP_EOL);
		exit(1);
	}
	
	$process_stdout = stream_get_contents($pipes[1]);
	fclose($pipes[1]);
	
	$process_stderr = stream_get_contents($pipes[2]);
	fclose($pipes[2]);
	
	$process_exit_code = proc_close($process);
	
	// proc_close fail
	if($process_exit_code < 0){
		$last_error = error_get_last();
		fwrite(STDERR, 'Unable exec pandoc: ' . $last_error['message'] . PHP_EOL);
		exit(1);
	}
	
	// pandoc error
	if($process_exit_code > 0){
		fwrite(STDERR, 'pandoc error:' . PHP_EOL . $process_stderr);
		exit($process_exit_code);
	}
	
	// In case of something in STDERR to log
	fwrite(STDERR, $process_stderr);
	
	$ast = json_decode($process_stdout);

	$json_error = json_last_error();
	if($json_error > JSON_ERROR_NONE){
		fwrite(STDERR, 'Pandoc JSON AST error ' . $json_error . PHP_EOL);
		exit(1);
	}
	
	// Update
	$ast = ast_walk($ast, 'link_filter', "json", $ast[0]->unMeta);

	// TODO
	// check if h1 title exist
	// resolve link title (+include anchors name)
	// resolve entries #AX***
	// resolve browser bugtracking ID
	// update entries relations
	// remove duplicated links in list
	// add tag based on **keyword**
	
	// JSON AST to FORMAT
	$cmd = sprintf('pandoc -f json -t %s %s', $format, $output_args);
	$descriptorspec = array(
		0 => array('pipe', 'r'), // stdin
		1 => STDOUT, // stdout
		2 => array('pipe', 'a') // stderr
	);
	$process = proc_open($cmd, $descriptorspec, $pipes);
	
	// proc_open fail
	if (!is_resource($process)) {
		$last_error = error_get_last();
		fwrite(STDERR, 'Unable exec pandoc: ' . $last_error['message'] . PHP_EOL);
		exit(1);
	}

	$json = json_encode($ast, JSON_HEX_TAG|JSON_HEX_AMP|JSON_UNESCAPED_SLASHES|JSON_UNESCAPED_UNICODE);
	fwrite($pipes[0], $json);
	fclose($pipes[0]);
	
	$process_stderr = stream_get_contents($pipes[2]);
	fclose($pipes[2]);
	
	$process_exit_code = proc_close($process);
	
	// proc_close fail
	if($process_exit_code < 0){
		$last_error = error_get_last();
		fwrite(STDERR, 'Unable exec pandoc: ' . $last_error['message'] . PHP_EOL);
		exit(1);
	}
	
	// pandoc error
	if($process_exit_code > 0){
		fwrite(STDERR, 'pandoc error:' . PHP_EOL . $process_stderr);
		exit($process_exit_code);
	}
	
	// In case of something in STDERR to log
	fwrite(STDERR, $process_stderr);
}

main();