memories/lib/Command/PlacesSetup.php

416 lines
15 KiB
PHP
Raw Normal View History

2023-02-06 01:15:18 +00:00
<?php
declare(strict_types=1);
/**
* @copyright Copyright (c) 2022, Varun Patil <radialapps@gmail.com>
* @author Varun Patil <radialapps@gmail.com>
* @license AGPL-3.0-or-later
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
namespace OCA\Memories\Command;
use OCP\IConfig;
use OCP\IDBConnection;
use Symfony\Component\Console\Command\Command;
use Symfony\Component\Console\Input\InputInterface;
use Symfony\Component\Console\Output\OutputInterface;
const GIS_TYPE_NONE = 0;
const GIS_TYPE_MYSQL = 1;
const GIS_TYPE_POSTGRES = 2;
2023-02-06 15:50:07 +00:00
const APPROX_PLACES = 635189;
2023-02-06 01:15:18 +00:00
2023-02-06 15:50:07 +00:00
const PLANET_URL = 'https://github.com/pulsejet/memories-assets/releases/download/geo-0.0.2/planet_coarse_boundaries.zip';
2023-02-06 06:42:34 +00:00
class PlacesSetup extends Command
2023-02-06 01:15:18 +00:00
{
protected IConfig $config;
protected OutputInterface $output;
protected IDBConnection $connection;
protected int $gisType = GIS_TYPE_NONE;
public function __construct(
IConfig $config,
IDBConnection $connection
) {
parent::__construct();
$this->config = $config;
$this->connection = $connection;
}
protected function configure(): void
{
$this
2023-02-06 06:42:34 +00:00
->setName('memories:places-setup')
2023-02-06 01:15:18 +00:00
->setDescription('Setup reverse geocoding')
;
}
protected function execute(InputInterface $input, OutputInterface $output): int
{
$this->output = $output;
$this->output->writeln('Attempting to set up reverse geocoding');
// Detect the GIS type
$this->detectGisType();
$this->config->setSystemValue('memories.gis_type', $this->gisType);
// Make sure we support something
if (GIS_TYPE_NONE === $this->gisType) {
$this->output->writeln('<error>No supported GIS type detected</error>');
return 1;
}
// Check if the database is already set up
2023-02-06 15:29:24 +00:00
$hasDb = false;
2023-02-06 01:15:18 +00:00
try {
2023-02-06 06:42:34 +00:00
$this->output->writeln('');
2023-02-06 01:15:18 +00:00
$this->connection->executeQuery('SELECT osm_id FROM memories_planet_geometry LIMIT 1')->fetch();
2023-02-06 06:42:34 +00:00
$this->output->writeln('<error>Database is already set up</error>');
2023-02-06 15:29:24 +00:00
$this->output->writeln('<error>This will drop and re-download the planet database</error>');
2023-02-06 06:42:34 +00:00
$this->output->writeln('<error>This is generally not necessary to do frequently </error>');
2023-02-06 15:29:24 +00:00
$hasDb = true;
2023-02-06 01:15:18 +00:00
} catch (\Exception $e) {
}
2023-02-06 06:42:34 +00:00
// Ask confirmation
$tempdir = sys_get_temp_dir();
$this->output->writeln('');
$this->output->writeln('Are you sure you want to download the planet database?');
$this->output->writeln("This will take a very long time and use some disk space in {$tempdir}");
$this->output->write('Proceed? [y/N] ');
$handle = fopen('php://stdin', 'r');
$line = fgets($handle);
if ('y' !== trim($line)) {
$this->output->writeln('Aborting');
return 1;
}
2023-02-06 15:29:24 +00:00
// Drop the table
$p = $this->connection->getDatabasePlatform();
if ($hasDb) {
$this->output->writeln('');
$this->output->write('Dropping table ... ');
$this->connection->executeStatement($p->getDropTableSQL('memories_planet_geometry'));
$this->output->writeln('OK');
}
// Setup the database
$this->output->write('Setting up database ... ');
$this->setupDatabase();
2023-02-06 06:42:34 +00:00
$this->output->writeln('OK');
2023-02-06 01:15:18 +00:00
// Truncate tables
2023-02-06 06:42:34 +00:00
$this->output->write('Truncating tables ... ');
2023-02-06 02:46:39 +00:00
$this->connection->executeStatement($p->getTruncateTableSQL('*PREFIX*memories_planet', false));
$this->connection->executeStatement($p->getTruncateTableSQL('memories_planet_geometry', false));
2023-02-06 06:42:34 +00:00
$this->output->writeln('OK');
2023-02-06 15:29:24 +00:00
// Download the data
$this->output->write('Downloading data ... ');
$datafile = $this->downloadPlanet();
$this->output->writeln('OK');
2023-02-06 06:42:34 +00:00
// Start importing
$this->output->writeln('');
$this->output->writeln('Importing data (this will take a while) ...');
2023-02-06 01:15:18 +00:00
// Start time
$start = time();
// Create place insertion statement
$query = $this->connection->getQueryBuilder();
$query->insert('memories_planet')
->values([
'osm_id' => $query->createParameter('osm_id'),
'admin_level' => $query->createParameter('admin_level'),
'name' => $query->createParameter('name'),
])
;
$insertPlace = $this->connection->prepare($query->getSQL());
// Create geometry insertion statement
$query = $this->connection->getQueryBuilder();
$geomParam = $query->createParameter('geometry');
if (GIS_TYPE_MYSQL === $this->gisType) {
$geomParam = "ST_GeomFromText({$geomParam})";
} elseif (GIS_TYPE_POSTGRES === $this->gisType) {
$geomParam = "POLYGON({$geomParam})";
}
$query->insert('memories_planet_geometry')
->values([
'id' => $query->createParameter('id'),
'poly_id' => $query->createParameter('poly_id'),
'type_id' => $query->createParameter('type_id'),
'osm_id' => $query->createParameter('osm_id'),
'geometry' => $query->createFunction($geomParam),
])
;
$sql = str_replace('*PREFIX*memories_planet_geometry', 'memories_planet_geometry', $query->getSQL());
$insertGeometry = $this->connection->prepare($sql);
// Iterate over the data file
2023-02-06 01:15:18 +00:00
$handle = fopen($datafile, 'r');
if ($handle) {
$count = 0;
while (($line = fgets($handle)) !== false) {
// Skip empty lines
if ('' === trim($line)) {
continue;
}
++$count;
// Decode JSON
$data = json_decode($line, true);
if (null === $data) {
$this->output->writeln('<error>Failed to decode JSON</error>');
continue;
}
// Extract data
$osmId = $data['osm_id'];
$adminLevel = $data['admin_level'];
$name = $data['name'];
$boundaries = $data['geometry'];
2023-02-06 07:24:29 +00:00
// Skip some places
if ($adminLevel >= 10) {
// These are too specific, e.g. "Community Board"
continue;
}
2023-02-06 01:15:18 +00:00
// Insert place into database
$insertPlace->bindValue('osm_id', $osmId);
$insertPlace->bindValue('admin_level', $adminLevel);
$insertPlace->bindValue('name', $name);
$insertPlace->execute();
2023-02-06 01:15:18 +00:00
// Insert polygons into database
$idx = 0;
foreach ($boundaries as &$polygon) {
2023-02-06 15:29:24 +00:00
// $polygon is a struct as
// [ "t" => "e", "c" => [lon, lat], [lon, lat], ... ] ]
$polyid = $polygon['i'];
$typeid = $polygon['t'];
$pkey = $polygon['k'];
$coords = $polygon['c'];
// Create parameters
2023-02-06 01:15:18 +00:00
++$idx;
$geometry = '';
2023-02-06 15:29:24 +00:00
if (\count($coords) < 3) {
2023-02-06 01:15:18 +00:00
$this->output->writeln('<error>Invalid polygon</error>');
continue;
}
if (GIS_TYPE_MYSQL === $this->gisType) {
2023-02-06 18:50:48 +00:00
$points = implode(',', array_map(function (&$point) {
$x = $point[0];
$y = $point[1];
2023-02-06 01:15:18 +00:00
2023-02-06 18:50:48 +00:00
return "{$x} {$y}";
}, $coords));
$geometry = "POLYGON(({$points}))";
2023-02-06 01:15:18 +00:00
} elseif (GIS_TYPE_POSTGRES === $this->gisType) {
$geometry = implode(',', array_map(function (&$point) {
2023-02-06 18:50:48 +00:00
$x = $point[0];
$y = $point[1];
return "({$x},{$y})";
}, $coords));
2023-02-06 01:15:18 +00:00
}
try {
$insertGeometry->bindValue('id', $pkey);
$insertGeometry->bindValue('poly_id', $polyid);
$insertGeometry->bindValue('type_id', $typeid);
$insertGeometry->bindValue('osm_id', $osmId);
$insertGeometry->bindValue('geometry', $geometry);
$insertGeometry->execute();
2023-02-06 01:15:18 +00:00
} catch (\Exception $e) {
$this->output->writeln('<error>Failed to insert into database</error>');
$this->output->writeln($e->getMessage());
continue;
}
2023-02-06 16:01:47 +00:00
}
2023-02-06 01:15:18 +00:00
2023-02-06 16:01:47 +00:00
// Print progress
if (0 === $count % 500) {
$end = time();
$elapsed = $end - $start;
$rate = $count / $elapsed;
$remaining = APPROX_PLACES - $count;
$eta = round($remaining / $rate);
$rate = round($rate, 1);
$this->output->writeln("Inserted {$count} places, {$rate}/s, ETA: {$eta}s, Last: {$name}");
2023-02-06 01:15:18 +00:00
}
}
fclose($handle);
}
2023-02-06 06:42:34 +00:00
// Delete file
unlink($datafile);
// Done
$this->output->writeln('');
$this->output->writeln('Planet database imported successfully!');
$this->output->writeln('If this is the first time you did this, you should now run:');
$this->output->writeln('occ memories:index -f');
// Mark success
$this->config->setSystemValue('memories.gis_type', $this->gisType);
2023-02-06 01:15:18 +00:00
return 0;
}
protected function detectGisType()
{
// Test MySQL-like support in databse
try {
$res = $this->connection->executeQuery("SELECT ST_GeomFromText('POINT(1 1)')")->fetch();
if (0 === \count($res)) {
throw new \Exception('Invalid result');
}
$this->output->writeln('MySQL-like support detected!');
// Make sure this is actually MySQL
$res = $this->connection->executeQuery('SELECT VERSION()')->fetch();
if (0 === \count($res)) {
throw new \Exception('Invalid result');
}
if (false === strpos($res['VERSION()'], 'MariaDB') && false === strpos($res['VERSION()'], 'MySQL')) {
throw new \Exception('MySQL not detected');
}
$this->gisType = GIS_TYPE_MYSQL;
} catch (\Exception $e) {
$this->output->writeln('No MySQL-like support detected');
}
// Test Postgres native geometry like support in database
if (GIS_TYPE_NONE === $this->gisType) {
try {
$res = $this->connection->executeQuery("SELECT POINT('1,1')")->fetch();
if (0 === \count($res)) {
throw new \Exception('Invalid result');
}
$this->output->writeln('Postgres native geometry support detected!');
$this->gisType = GIS_TYPE_POSTGRES;
} catch (\Exception $e) {
$this->output->writeln('No Postgres native geometry support detected');
}
}
}
protected function setupDatabase(): void
{
try {
$sql = 'CREATE TABLE memories_planet_geometry (
id varchar(255) NOT NULL PRIMARY KEY,
2023-02-06 15:29:24 +00:00
poly_id varchar(255) NOT NULL,
type_id int NOT NULL,
2023-02-06 01:15:18 +00:00
osm_id int NOT NULL,
geometry polygon NOT NULL
);';
$this->connection->executeQuery($sql);
// Add indexes
$this->connection->executeQuery('CREATE INDEX planet_osm_id_idx ON memories_planet_geometry (osm_id);');
// Add spatial index
if (GIS_TYPE_MYSQL === $this->gisType) {
$this->connection->executeQuery('CREATE SPATIAL INDEX planet_osm_polygon_geometry_idx ON memories_planet_geometry (geometry);');
} elseif (GIS_TYPE_POSTGRES === $this->gisType) {
$this->connection->executeQuery('CREATE INDEX planet_osm_polygon_geometry_idx ON memories_planet_geometry USING GIST (geometry);');
}
} catch (\Exception $e) {
$this->output->writeln('Failed to create planet table');
$this->output->writeln($e->getMessage());
exit;
}
}
2023-02-06 15:58:36 +00:00
protected function ensureDeleted(string $filename)
{
2023-02-06 15:56:02 +00:00
unlink($filename);
if (file_exists($filename)) {
$this->output->writeln('<error>Failed to delete data file</error>');
$this->output->writeln("Please delete {$filename} manually");
2023-02-06 15:54:25 +00:00
exit;
}
2023-02-06 15:56:02 +00:00
}
2023-02-06 15:54:25 +00:00
2023-02-06 15:56:02 +00:00
protected function downloadPlanet(): string
{
2023-02-06 06:42:34 +00:00
$filename = sys_get_temp_dir().'/planet_coarse_boundaries.zip';
2023-02-06 15:56:02 +00:00
$this->ensureDeleted($filename);
$txtfile = sys_get_temp_dir().'/planet_coarse_boundaries.txt';
$this->ensureDeleted($txtfile);
2023-02-06 06:42:34 +00:00
$fp = fopen($filename, 'w+');
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, PLANET_URL);
curl_setopt($ch, CURLOPT_FILE, $fp);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
curl_setopt($ch, CURLOPT_TIMEOUT, 60000);
curl_exec($ch);
curl_close($ch);
fclose($fp);
// Unzip
$zip = new \ZipArchive();
$res = $zip->open($filename);
if (true === $res) {
$zip->extractTo(sys_get_temp_dir());
$zip->close();
} else {
$this->output->writeln('Failed to unzip planet file');
exit;
2023-02-06 01:15:18 +00:00
}
2023-02-06 06:42:34 +00:00
// Check if file exists
if (!file_exists($txtfile)) {
$this->output->writeln('Failed to find planet data file after unzip');
exit;
}
// Delete zip file
unlink($filename);
return $txtfile;
2023-02-06 01:15:18 +00:00
}
}