| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238 |
- <?php
- /**
- * 计算章节的父子,前后关系
- * 输入: csv文件
- */
- namespace App\Console\Commands;
- use Illuminate\Console\Command;
- use App\Models\PaliText;
- use App\Models\BookTitle;
- use Illuminate\Support\Facades\DB;
- use Illuminate\Support\Facades\Log;
- class UpgradePaliText extends Command
- {
- /**
- * The name and signature of the console command.
- *
- * @var string
- */
- protected $signature = 'upgrade:palitext {from?} {to?}';
- /**
- * The console command description.
- *
- * @var string
- */
- protected $description = 'upgrade pali_texts paragraph infomation';
- /**
- * Create a new command instance.
- *
- * @return void
- */
- public function __construct()
- {
- parent::__construct();
- }
- /**
- * Execute the console command.
- *
- * @return int
- */
- public function handle()
- {
- $this->info("upgrade pali text");
- $startTime = time();
- $_from = $this->argument('from');
- $_to = $this->argument('to');
- if(empty($_from) && empty($_to)){
- $_from = 1;
- $_to = 217;
- }else if(empty($_to)){
- $_to = $_from;
- }
- #载入文件列表
- $fileListFileName = config("app.path.palitext_filelist");
- $filelist = array();
- if (($handle = fopen($fileListFileName, 'r')) !== false) {
- while (($filelist[] = fgetcsv($handle, 0, ',')) !== false) {
- }
- }
- $bar = $this->output->createProgressBar($_to-$_from+1);
- for ($from=$_from; $from <= $_to; $from++) {
- $inputRow = 0;
- $arrInserString = array();
- #载入csv数据
- $FileName = $filelist[$from-1][1];
- $csvFile = config("app.path.pali_title") .'/'. $from.'_pali.csv';
- if (($fp = fopen($csvFile, "r")) !== false) {
- Log::info("csv load:" . $csvFile);
- while (($data = fgetcsv($fp, 0, ',')) !== false) {
- if ($inputRow > 0) {
- array_push($arrInserString, $data);
- }
- $inputRow++;
- }
- fclose($fp);
- } else {
- $this->error( "can not open csv file. filename=" . $csvFile. PHP_EOL) ;
- Log::error( "can not open csv file. filename=" . $csvFile) ;
- continue;
- }
- $title_data = PaliText::select(['book','paragraph','level','parent','toc','lenght'])
- ->where('book',$from)->orderby('paragraph','asc')->get();
- {
- $paragraph_count = count($title_data);
- $paragraph_info = array();
- $paragraph_info[] = array($from, -1, $paragraph_count, -1, -1, -1);
- for ($iPar = 0; $iPar < count($title_data); $iPar++) {
- $title_data[$iPar]["level"] = $arrInserString[$iPar][3];
- }
- for ($iPar = 0; $iPar < count($title_data); $iPar++) {
- $book = $from ;
- $paragraph = $title_data[$iPar]["paragraph"];
- if ((int) $title_data[$iPar]["level"] == 8) {
- $title_data[$iPar]["level"] = 100;
- }
- $curr_level = (int) $title_data[$iPar]["level"];
- # 计算这个chapter的段落数量
- $length = -1;
- for ($iPar1 = $iPar + 1; $iPar1 < count($title_data); $iPar1++) {
- $thislevel = (int) $title_data[$iPar1]["level"];
- if ($thislevel <= $curr_level) {
- $length = (int) $title_data[$iPar1]["paragraph"] - $paragraph;
- break;
- }
- }
- if ($length == -1) {
- $length = $paragraph_count - $paragraph + 1;
- }
- /*
- 上一个段落
- 算法:查找上一个标题段落。而且该标题段落的下一个段落不是标题段落
- */
- $prev = -1;
- if ($iPar > 0) {
- for ($iPar1 = $iPar - 1; $iPar1 >= 0; $iPar1--) {
- if ($title_data[$iPar1]["level"] < 8 && $title_data[$iPar1+1]["level"]==100) {
- $prev = $title_data[$iPar1]["paragraph"];
- break;
- }
- }
- }
- /*
- 下一个段落
- 算法:查找下一个标题段落。而且该标题段落的下一个段落不是标题段落
- */
- $next = -1;
- if ($iPar < count($title_data) - 1) {
- for ($iPar1 = $iPar + 1; $iPar1 < count($title_data)-1; $iPar1++) {
- if ($title_data[$iPar1]["level"] <8 && $title_data[$iPar1+1]["level"]==100) {
- $next = $title_data[$iPar1]["paragraph"];
- break;
- }
- }
- }
- $parent = -1;
- if ($iPar > 0) {
- for ($iPar1 = $iPar - 1; $iPar1 >= 0; $iPar1--) {
- if ($title_data[$iPar1]["level"] < $curr_level) {
- $parent = $title_data[$iPar1]["paragraph"];
- break;
- }
- }
- }
- //计算章节包含总字符数
- $iChapter_strlen = 0;
- for ($i = $iPar; $i < $iPar + $length; $i++) {
- $iChapter_strlen += $title_data[$i]["lenght"];
- }
- $newData = [
- 'level' => $arrInserString[$iPar][3],
- 'toc' => $arrInserString[$iPar][5],
- 'chapter_len' => $length,
- 'next_chapter' => $next,
- 'prev_chapter' => $prev,
- 'parent' => $parent,
- 'chapter_strlen'=> $iChapter_strlen,
- ];
- $path = [];
- $title_data[$iPar]["level"] = $newData["level"];
- $title_data[$iPar]["toc"] = $newData["toc"];
- $title_data[$iPar]["parent"] = $newData["parent"];
- /*
- *获取路径
- */
- $currParent = $parent;
- $iLoop = 0;
- while ($currParent != -1 && $iLoop<7) {
- # code...
- $pathTitle = $title_data[$currParent-1]["toc"];
- $pathLevel = $title_data[$currParent-1]['level'];
- $path[] = ["book"=>$book,"paragraph"=>$currParent,"title"=>$pathTitle,"level"=>$pathLevel];
- $currParent = $title_data[$currParent-1]["parent"];
- $iLoop++;
- }
- if(count($path)>0){
- //插入书名
- $bookTitle = BookTitle::where('book',$book)
- ->where('paragraph',end($path)['paragraph'])
- ->value('title');
- $path[] = ["book"=>0,"paragraph"=>0,"title"=>$bookTitle,"level"=>0];
- }
- # 将路径反向
- $path1 = [];
- for ($i=count($path)-1; $i >=0 ; $i--) {
- # code...
- $path1[] = $path[$i];
- }
- $newData['path'] = $path1;
- PaliText::where('book',$book)
- ->where('paragraph',$paragraph)
- ->update($newData);
- if ($curr_level > 0 && $curr_level < 8) {
- $paragraph_info[] = array($book, $paragraph, $length, $prev, $next, $parent);
- }
- }
- }
- $bar->advance();
- }
- $bar->finish();
- $this->info("instert pali text finished. in ". time()-$startTime . "s" );
- Log::info("instert pali text finished. in ". time()-$startTime . "s");
- return 0;
- }
- }
|