<?php
declare(strict_types=1);
/**
* Add or Edit Joomla! Articles Via API Using Streamed CSV
* - When id = 0 in csv it's doing a POST. If alias exists it add a random slug at the end of your alias and do POST again
* - When id > 0 in csv it's doing a PATCH. If alias exists it add a random slug at the end of your alias and do PATCH again
*
* @author Alexandre ELISÉ <This email address is being protected from spambots. You need JavaScript enabled to view it. >
* @copyright (c) 2009 - present. Alexandre ELISÉ. All rights reserved.
* @license GPL-2.0-and-later GNU General Public License v2.0 or later
* @link https://apiadept.com
*/
// Public url of the sample csv used in this example (CHANGE WITH YOUR OWN CSV URL IF YOU WISH)
$csvUrl = 'https://docs.google.com/spreadsheets/d/e/2PACX-1vTO8DC8xzEEyP754B0kBu1sa2P9Rn3I8OLmq_RJYHwOwTlY8OGvpjp1yvaE84Imj0HYQeJcNKT2TOFR/pub?output=csv';
// Your Joomla! 4.x website base url
$baseUrl = '';
// Your Joomla! 4.x Api Token (DO NOT STORE IT IN YOUR REPO USE A VAULT OR A PASSWORD MANAGER)
$token = '';
$basePath = 'api/index.php/v1';
// Request timeout
$timeout = 10;
// Add custom fields support (shout-out to Marc DECHÈVRE : CUSTOM KING)
// The keys are the columns in the csv with the custom fields names (that's how Joomla! Web Services Api work as of today)
// For the custom fields to work they need to be added in the csv and to exists in the Joomla! site.
$customFieldKeys = []; //['with-coffee','with-dessert','extra-water-bottle'];
// This time we need endpoint to be a function to make it more dynamic
$endpoint = function (string $givenBaseUrl, string $givenBasePath, int $givenResourceId = 0): string {
return $givenResourceId ? sprintf('%s/%s/%s/%d', $givenBaseUrl, $givenBasePath, 'content/articles', $givenResourceId)
: sprintf('%s/%s/%s', $givenBaseUrl, $givenBasePath, 'content/articles');
};
// PHP Generator to efficiently read the csv file
$generator = function (string $url, array $keys = []): Generator {
if (empty($url))
{
yield new RuntimeException('Url MUST NOT be empty', 422);
}
$defaultKeys = [
'id',
'title',
'alias',
'catid',
'articletext',
'introtext',
'fulltext',
'language',
'metadesc',
'metakey',
'state',
'featured',
'publish_up',
'publish_down',
'featured_up',
'featured_down',
];
$mergedKeys = array_unique(array_merge($defaultKeys, $keys));
// Assess robustness of the code by trying random key order
//shuffle($mergedKeys);
$resource = fopen($url, 'r');
if ($resource === false)
{
yield new RuntimeException('Could not read csv file', 500);
}
try
{
//NON-BLOCKING I/O (Does not wait before processing next line.)
stream_set_blocking($resource, false);
$firstLine = stream_get_line(
$resource,
0,
"\r\n"
);
if (empty($firstLine))
{
yield new RuntimeException('First line MUST NOT be empty. It is the header', 422);
}
$csvHeaderKeys = str_getcsv($firstLine);
$commonKeys = array_intersect($csvHeaderKeys, $mergedKeys);
do
{
$currentLine = stream_get_line(
$resource,
0,
"\r\n"
);
if (empty($currentLine))
{
yield new RuntimeException('Current line MUST NOT be empty', 422);
}
$extractedContent = str_getcsv($currentLine);
// Allow using csv keys in any order
$commonValues = array_intersect_key($extractedContent, $commonKeys);
$encodedContent = json_encode(array_combine($commonKeys, $commonValues));
if ($encodedContent !== false)
{
yield $encodedContent;
}
yield new RuntimeException('Current line seem to be invalid', 422);
} while (!feof($resource));
} finally
{
fclose($resource);
}
};
// Process data returned by the PHP Generator
$process = function (string $givenHttpVerb, string $endpoint, string $dataString, array $headers, int $timeout, $transport) {
curl_setopt_array($transport, [
CURLOPT_URL => $endpoint,
CURLOPT_RETURNTRANSFER => true,
CURLOPT_ENCODING => 'utf-8',
CURLOPT_MAXREDIRS => 10,
CURLOPT_TIMEOUT => $timeout,
CURLOPT_FOLLOWLOCATION => true,
CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_2TLS,
CURLOPT_CUSTOMREQUEST => $givenHttpVerb,
CURLOPT_POSTFIELDS => $dataString,
CURLOPT_HTTPHEADER => $headers,
]
);
$response = curl_exec($transport);
// Continue even on partial failure
if (empty($response))
{
throw new RuntimeException('Empty output', 422);
}
return $response;
};
// Read CSV in a PHP Generator using streams in non-blocking I/O mode
$streamCsv = $generator($csvUrl, $customFieldKeys);
$storage = [];
foreach ($streamCsv as $dataKey => $dataString)
{
if (!is_string($dataString))
{
continue;
}
$curl = curl_init();
try
{
// HTTP request headers
$headers = [
'Accept: application/vnd.api+json',
'Content-Type: application/json',
'Content-Length: ' . mb_strlen($dataString),
sprintf('X-Joomla-Token: %s', trim($token)),
];
$decodedDataString = json_decode($dataString, true);
// Article primary key. Usually 'id'
$pk = (int) $decodedDataString['id'];
$output = $process($pk ? 'PATCH' : 'POST', $endpoint($baseUrl, $basePath, $pk), $dataString, $headers, $timeout, $curl);
$decodedJsonOutput = json_decode($output, true);
// don't show errors, handle them gracefully
if (isset($decodedJsonOutput['errors']))
{
// If article is potentially a duplicate (already exists with same alias)
$storage[$dataKey] = ['mightExists' => $decodedJsonOutput['errors'][0]['code'] === 400, 'decodedDataString' => $decodedDataString];
continue;
}
echo $output . PHP_EOL;
}
catch (Throwable $e)
{
echo $e->getMessage() . PHP_EOL;
continue;
} finally
{
curl_close($curl);
}
}
// Handle errors and retries
foreach ($storage as $item)
{
$curl = curl_init();
try
{
if ($item['mightExists'])
{
$pk = (int) $item['decodedDataString']['id'];
$item['decodedDataString']['alias'] = sprintf('%s-%s', $item['decodedDataString']['alias'], bin2hex(random_bytes(4)));
// back to json string after changing alias
$dataString = json_encode($item['decodedDataString']);
// HTTP request headers
$headers = [
'Accept: application/vnd.api+json',
'Content-Type: application/json',
'Content-Length: ' . mb_strlen($dataString),
sprintf('X-Joomla-Token: %s', trim($token)),
];
$output = $process($pk ? 'PATCH' : 'POST', $endpoint($baseUrl, $basePath, $pk), $dataString, $headers, $timeout, $curl);
echo $output . PHP_EOL;
}
}
catch (Throwable $e)
{
echo $e->getMessage() . PHP_EOL;
continue;
} finally
{
curl_close($curl);
}
}
For more convenience, you might want to go on the repo here
