$urlDerived = "http://www." . $domain . "/";
// This MUST be the URL we tried the very first time.
// Get the most recent result from HTTP Archive for this URL.
$row = doRowQuery("select pageid, urlHtml from {$gPagesTable} where url='{$urlDerived}' order by pageid desc limit 1;");
if (!$row) {
// This is the first time for this URL.
// We'll have to clean it up next time.
continue;
}
$pageid = $row['pageid'];
$urlHtml = $row['urlHtml'];
if ($urlDerived != $urlHtml) {
$xdiff++;
// CVSNO
// Get more info about the initial URL.
$row = doRowQuery("select requestid, status, resp_location, resp_cache_control, resp_expires from {$gRequestsTable} where pageid={$pageid} and url='{$urlDerived}' order by requestid asc limit 1;");
$requestid = $row['requestid'];
$status = $row['status'];
$resp_cache_control = $row['resp_cache_control'];
$resp_expires = $row['resp_expires'];
$resp_location = $row['resp_location'];
if (0 === strpos($resp_location, "/")) {
// relative location
$resp_location = $urlDerived + strstr($resp_location, 1);
}
if (301 == $status || 302 == $status) {
$x300++;
// CVSNO
if (false != strpos($urlHtml, "?")) {
// Don't store a derived URL that contains a querystring.
$snotfixed .= "NOTFIXED (querystring): {$urlDerived} != {$urlHtml}\n";
if ($minid === $maxid) {
tprint("The rows have already been copied.");
} else {
$cmd = "replace into {$pagesTable} select * from pagestmp where pageid >= {$minid} and pageid <= {$maxid} and maxDomainReqs != 0;";
tprint("Replacing " . ($maxid - $minid) . " rows from pagestmp to {$pagesTable}:\n {$cmd}");
doSimpleCommand($cmd);
tprint("...done copying rows.");
}
}
// 5. Recalculate stats
tprint("\n5. Recalculate stats...");
// TODO - This script doesn't detect & skip this step if it's already been done, but it's very fast (20 seconds) so we won't worry.
// TODO - could test for bytesFont or perFonts
tprint("Update numPages and numRequests in crawls:");
// It's possible that while removing orphans and pages with wptid="" some meta-crawl information has changed:
$row = doRowQuery("select count(*) as numPages, min(pageid) as minPageid, max(pageid) as maxPageid from {$pagesTable} where {$pageidCond};");
$numRequests = doSimpleQuery("select count(*) from {$requestsTable} where {$pageidCond};");
doSimpleCommand("update {$gCrawlsTable} set numPages = " . $row['numPages'] . ", minPageid = " . $row['minPageid'] . ", maxPageid = " . $row['maxPageid'] . ", numRequests = {$numRequests} where label = '{$label}' and location='{$device}';");
tprint("Compute stats:");
removeStats($label, NULL, $device);
computeMissingStats($device, true);
tprint("...done recalculating stats.");
// 6. Mysqldump
tprint("\n6. Mysqldump & wrap up...");
$col = doSimpleQuery("show columns from {$requestsTable} like '%redirectUrlShort%';");
if ($col) {
tprint("You have to remove the redirectUrlShort column before we can do the dumps:\n" . " alter table {$requestsTable} drop column redirectUrlShort;");
tprint("almost done!");
} else {
$labelUnderscore = str_replace(" ", "_", $label);
$tmpdir = "/tmp/{$labelUnderscore}." . time();
See the License for the specific language governing permissions and
limitations under the License.
*/
require_once "utils.inc";
require_once "ui.inc";
require_once "pages.inc";
$gArchive = getParam('a');
$gLabel = getParam('l');
$gPageid = getParam('p');
$gFormat = getParam('format');
if ($gPageid && "csv" == $gFormat) {
// request to download a CSV of an individual website's requests data
header('Content-Type: application/octet-stream; name="httparchive.csv"');
header('Content-Disposition: inline; filename="httparchive_page' . $gPageid . '.csv"');
$aColumns = array("url", "mimeType", "method", "status", "time", "respSize", "reqCookieLen", "respCookieLen", "reqHttpVersion", "respHttpVersion", "req_accept", "req_accept_charset", "req_accept_encoding", "req_accept_language", "req_connection", "req_host", "req_referer", "resp_accept_ranges", "resp_age", "resp_cache_control", "resp_connection", "resp_content_encoding", "resp_content_language", "resp_content_length", "resp_content_location", "resp_content_type", "resp_date", "resp_etag", "resp_expires", "resp_keep_alive", "resp_last_modified", "resp_location", "resp_pragma", "resp_server", "resp_transfer_encoding", "resp_vary", "resp_via", "resp_x_powered_by");
$sRows = implode(",", $aColumns);
$sRows .= "\n";
$row = doRowQuery("select wptid, wptrun from {$gPagesTable} where pageid = {$gPageid};");
$page = pageFromWPT($row['wptid'], $row['wptrun']);
$aResources = $page['resources'];
foreach ($aResources as $resource) {
foreach ($aColumns as $column) {
$sRows .= (array_key_exists($column, $resource) ? '"' . $resource[$column] . '"' : "") . ",";
// wrap in double quotes in case of commas
}
rtrim($sRows, ",");
// remove trailing comma
$sRows .= "\n";
}
echo $sRows;
}
exit;
}
if (0 === totalNotDone()) {
$curPasses++;
updateCrawl($labelFromRun, $gArchive, $locations[0], array("passes" => $curPasses));
if ($curPasses < $maxPasses) {
resubmitFailures();
cprint("Resubmitted failures - going around once again...");
} else {
// We just finished the last pass. Wrap it up...
cprint(date("G:i") . ": DONE with tests. Copying...");
$gParamLabel = $labelFromRun;
// hack!
// IMPORTANT: Update crawl info FIRST because many pieces of code reference this:
$query = "select min(pageid) as minid, max(pageid) as maxid from {$gPagesTable} where label='{$labelFromRun}';";
$row = doRowQuery($query);
$minid = $row['minid'];
$maxid = $row['maxid'];
$numPages = doSimpleQuery("select count(*) from {$gPagesTable} where pageid >= {$minid} and pageid <= {$maxid};");
$numRequests = doSimpleQuery("select count(*) from {$gRequestsTable} where pageid >= {$minid} and pageid <= {$maxid};");
updateCrawl($labelFromRun, $gArchive, $locations[0], array("minPageid" => $minid, "maxPageid" => $maxid, "numErrors" => statusErrors(), "numPages" => $numPages, "numRequests" => $numRequests));
// Copy rows, calc stats, create dump files, etc.
require_once "copy.php";
updateCrawl($labelFromRun, $gArchive, $locations[0], array("finishedDateTime" => time()));
cprint(date("G:i") . ": DONE with crawl!");
exit(0);
}
}
// TODO - Combine "obtain" and "parse"?
// The "crawl" process has multiple distinct tasks. This is because the URLs are
// sent to and queued at WebPagetest which is asynchronous.
