我如何通过光标的结果遍历每个页面

Question

So ive been trying for a few hours now to basically get the recentAverageRap of the user selected with this api. So i can get part of it with the code. But i can not figure out how to go to the next page if there is a next page. All I know is in the url i got to change: cursor= to the json value of nextPageCursor to get to next page. But ive been trying and cant get the correct info from all the pages.

Heres my code so far which only gets first page:

<?php
    $userid = htmlentities($_GET['userid'], ENT_QUOTES);
    $assettypes = array("Hat", "Face", "Gear", "HairAccessory", "FaceAccessory", "NeckAccessory", "ShoulderAccessory", "FrontAccessory", "BackAccessory", "WaistAccessory");

    $rap2 = 0;
    foreach($assettypes as $assettype){
        $url = "https://inventory.roblox.com/v1/users/" . $userid . "/assets/collectibles?assetType=" . $assettype . "&sortOrder=Asc&limit=100&cursor=";
        $get = file_get_contents($url);
        $json = json_decode($get);
        $rap = 0;

        foreach($json->data as $val){
            $rap += $val->recentAveragePrice;
        }

        $rap2 += $rap;
    }

    echo $rap2;
?>

Answer 1

Sorry to resurrect an old question, but I stumbled on this issue myself, and wrote a small PHP class to help me accomplish fetching all pages of Roblox's paginated WebAPI endpoints in a single request.

The basic solution (sequential & slow) would be to activate some kind of loop that continually calls the api, while incrementing ?page and ?cursor query param values until the response for a given AssetType no longer returns a nextPageCursor value, then breaking the loop and moving on.

<?php
    function get_page($userid, $assettype, $page, $cursor){
        $url = "https://inventory.roblox.com/v1/users/" . $userid . "/assets/collectibles?assetType=" . $assettype . "&sortOrder=Asc&limit=100&cursor=" . $cursor . "&page=" . $page;
        $get = file_get_contents($url);
        $json = json_decode($get);
        $rap = 0;
        $items = 0;
        foreach($json->data as $val){
            $rap += $val->recentAveragePrice;
            $items += 1;
        }
        return [$rap, $items, $json->nextPageCursor];
    }
    function get_all_pages_for_asset_type($userid, $assettype){
        $keep_fetching = true;
        $cursor = '';
        $page = 1;
        $rap = 0;
        $items = 0;
        while($keep_fetching){
            $results = get_page($userid, $assettype, $page, $cursor);
            $rap += $results[0];
            $items += $results[1];
            if(!$results[2] || empty(trim($results[2]))){
                // nextPageCursor empty, break
                $keep_fetching = false;
                break;
            }else{
                // nextPageCursor present, continue & fetch next page...
                $page+=1;
                $cursor = $results[2];
            }
        }
        return [$page, $items, $rap];
    }
    $userid = htmlentities($_GET['userid'], ENT_QUOTES);
    $assettypes = array("Hat", "Face", "Gear", "HairAccessory", "FaceAccessory", "NeckAccessory", "ShoulderAccessory", "FrontAccessory", "BackAccessory", "WaistAccessory");
    $rap2 = 0;
    $exec_time_start = microtime(true);
    $items_total = 0;
    $pages_total = 0;
    foreach($assettypes as $assettype){
        $results = get_all_pages_for_asset_type($userid, $assettype);
        $pages_total += $results[0];
        $items_total += $results[1];
        $rap2 += $results[2];
    }
    echo json_encode([
        'total_rap' => $rap2,
        'items' => $items_total,
        'pages' => $pages_total,
        'rendertime' => (microtime(true) - $exec_time_start)
    ]);
?>

A more advanced solution (concurrent & faster) would be to fire off multiple dynamic promise chains, 1 for each AssetType, which would dynamically run in their own, individual loops until nextPageCursor is empty, upon which that particular AssetType's promise chain could be resolved. Once all 10 promise chains resolve completely, (using GuzzleHttp\\Promise\\Settle , which is like JS's Promise.all(promises[]) the final calculation is done & returned.

I've posted the full code for both solutions to Github here: https://github.com/jakedowns/roblox-api-fetch-all-pages

The index.php file includes:

1. your original question example code
2. basic while-loop based solution
3. the complicated Guzzle,Promise,CurlMulti approach

Some basic timing comparisons between the approaches:

// ~4 seconds for 10 pages of 607 items (incomplete)
original_example();

// ~15 seconds for 31 pages of 2483
basic_solution();

// ~4 seconds for 31 pages of 2483 items
advanced_solution();