532 lines
18 KiB
JavaScript
532 lines
18 KiB
JavaScript
const fs = require('fs-extra');
|
||
|
||
const InpxParser = require('./InpxParser');
|
||
const InpxHashCreator = require('./InpxHashCreator');
|
||
const utils = require('./utils');
|
||
|
||
const emptyFieldValue = '?';
|
||
|
||
class DbCreator {
|
||
constructor(config) {
|
||
this.config = config;
|
||
}
|
||
|
||
async loadInpxFilter() {
|
||
const inpxFilterFile = this.config.inpxFilterFile;
|
||
|
||
if (await fs.pathExists(inpxFilterFile)) {
|
||
let filter = await fs.readFile(inpxFilterFile, 'utf8');
|
||
filter = JSON.parse(filter);
|
||
|
||
if (filter.includeAuthors) {
|
||
filter.includeAuthors = filter.includeAuthors.map(a => a.toLowerCase());
|
||
filter.includeSet = new Set(filter.includeAuthors);
|
||
}
|
||
|
||
if (filter.excludeAuthors) {
|
||
filter.excludeAuthors = filter.excludeAuthors.map(a => a.toLowerCase());
|
||
filter.excludeSet = new Set(filter.excludeAuthors);
|
||
}
|
||
|
||
return filter;
|
||
} else {
|
||
return false;
|
||
}
|
||
}
|
||
|
||
//процедура формировани БД несколько усложнена, в целях экономии памяти
|
||
async run(db, callback) {
|
||
const config = this.config;
|
||
|
||
callback({jobStepCount: 5});
|
||
callback({job: 'load inpx', jobMessage: 'Загрузка INPX', jobStep: 1, progress: 0});
|
||
|
||
//временная таблица
|
||
await db.create({
|
||
table: 'book',
|
||
cacheSize: (config.lowMemoryMode ? 5 : 500),
|
||
});
|
||
|
||
//поисковые таблицы, позже сохраним в БД
|
||
let authorMap = new Map();//авторы
|
||
let authorArr = [];
|
||
let seriesMap = new Map();//серии
|
||
let seriesArr = [];
|
||
let titleMap = new Map();//названия
|
||
let titleArr = [];
|
||
let genreMap = new Map();//жанры
|
||
let genreArr = [];
|
||
let langMap = new Map();//языки
|
||
let langArr = [];
|
||
|
||
//stats
|
||
let authorCount = 0;
|
||
let bookCount = 0;
|
||
let noAuthorBookCount = 0;
|
||
let bookDelCount = 0;
|
||
|
||
//stuff
|
||
let recsLoaded = 0;
|
||
callback({recsLoaded});
|
||
let chunkNum = 0;
|
||
|
||
//фильтр
|
||
const inpxFilter = await this.loadInpxFilter();
|
||
let filter = () => true;
|
||
if (inpxFilter) {
|
||
|
||
let recFilter = () => true;
|
||
if (inpxFilter.filter) {
|
||
if (config.allowUnsafeFilter)
|
||
recFilter = new Function(`'use strict'; return ${inpxFilter.filter}`)();
|
||
else
|
||
throw new Error(`Unsafe property 'filter' detected in ${this.config.inpxFilterFile}. Please specify '--unsafe-filter' param if you know what you're doing.`);
|
||
}
|
||
|
||
filter = (rec) => {
|
||
let author = rec.author;
|
||
if (!author)
|
||
author = emptyFieldValue;
|
||
|
||
author = author.toLowerCase();
|
||
|
||
let excluded = false;
|
||
if (inpxFilter.excludeSet) {
|
||
const authors = author.split(',');
|
||
|
||
for (const a of authors) {
|
||
if (inpxFilter.excludeSet.has(a)) {
|
||
excluded = true;
|
||
break;
|
||
}
|
||
}
|
||
}
|
||
|
||
return recFilter(rec)
|
||
&& (!inpxFilter.includeSet || inpxFilter.includeSet.has(author))
|
||
&& !excluded
|
||
;
|
||
};
|
||
}
|
||
|
||
//вспомогательные функции
|
||
const splitAuthor = (author) => {
|
||
if (!author)
|
||
author = emptyFieldValue;
|
||
|
||
const result = author.split(',');
|
||
if (result.length > 1)
|
||
result.push(author);
|
||
|
||
return result;
|
||
}
|
||
|
||
let totalFiles = 0;
|
||
const readFileCallback = async(readState) => {
|
||
callback(readState);
|
||
|
||
if (readState.totalFiles)
|
||
totalFiles = readState.totalFiles;
|
||
|
||
if (totalFiles)
|
||
callback({progress: (readState.current || 0)/totalFiles});
|
||
};
|
||
|
||
let id = 0;
|
||
const parsedCallback = async(chunk) => {
|
||
let filtered = false;
|
||
for (const rec of chunk) {
|
||
//сначала фильтр
|
||
if (!filter(rec)) {
|
||
rec.id = 0;
|
||
filtered = true;
|
||
continue;
|
||
}
|
||
|
||
rec.id = ++id;
|
||
|
||
if (!rec.del) {
|
||
bookCount++;
|
||
if (!rec.author)
|
||
noAuthorBookCount++;
|
||
} else {
|
||
bookDelCount++;
|
||
}
|
||
|
||
//авторы
|
||
const author = splitAuthor(rec.author);
|
||
|
||
for (let i = 0; i < author.length; i++) {
|
||
const a = author[i];
|
||
const value = a.toLowerCase();
|
||
|
||
let authorRec;
|
||
if (authorMap.has(value)) {
|
||
const authorTmpId = authorMap.get(value);
|
||
authorRec = authorArr[authorTmpId];
|
||
} else {
|
||
authorRec = {tmpId: authorArr.length, author: a, value, bookCount: 0, bookDelCount: 0, bookId: []};
|
||
authorArr.push(authorRec);
|
||
authorMap.set(value, authorRec.tmpId);
|
||
|
||
if (author.length == 1 || i < author.length - 1) //без соавторов
|
||
authorCount++;
|
||
}
|
||
|
||
//это нужно для того, чтобы имя автора начиналось с заглавной
|
||
if (a[0].toUpperCase() === a[0])
|
||
authorRec.author = a;
|
||
|
||
//счетчики
|
||
if (!rec.del) {
|
||
authorRec.bookCount++;
|
||
} else {
|
||
authorRec.bookDelCount++;
|
||
}
|
||
|
||
//ссылки на книги
|
||
authorRec.bookId.push(id);
|
||
}
|
||
}
|
||
|
||
let saveChunk = [];
|
||
if (filtered) {
|
||
saveChunk = chunk.filter(r => r.id);
|
||
} else {
|
||
saveChunk = chunk;
|
||
}
|
||
|
||
await db.insert({table: 'book', rows: saveChunk});
|
||
|
||
recsLoaded += chunk.length;
|
||
callback({recsLoaded});
|
||
|
||
if (chunkNum++ % 10 == 0 && config.lowMemoryMode)
|
||
utils.freeMemory();
|
||
};
|
||
|
||
//парсинг 1
|
||
const parser = new InpxParser();
|
||
await parser.parse(config.inpxFile, readFileCallback, parsedCallback);
|
||
|
||
utils.freeMemory();
|
||
|
||
//отсортируем авторов и выдадим им правильные id
|
||
//порядок id соответствует ASC-сортировке по author.toLowerCase
|
||
callback({job: 'author sort', jobMessage: 'Сортировка авторов', jobStep: 2, progress: 0});
|
||
await utils.sleep(100);
|
||
authorArr.sort((a, b) => a.value.localeCompare(b.value));
|
||
|
||
id = 0;
|
||
authorMap = new Map();
|
||
for (const authorRec of authorArr) {
|
||
authorRec.id = ++id;
|
||
authorMap.set(authorRec.author, id);
|
||
delete authorRec.tmpId;
|
||
}
|
||
|
||
utils.freeMemory();
|
||
|
||
//подготовка к сохранению author_book
|
||
const saveBookChunk = async(authorChunk, callback) => {
|
||
callback(0);
|
||
|
||
const ids = [];
|
||
for (const a of authorChunk) {
|
||
for (const id of a.bookId) {
|
||
ids.push(id);
|
||
}
|
||
}
|
||
|
||
ids.sort();// обязательно, иначе будет тормозить - особенности JembaDb
|
||
|
||
callback(0.1);
|
||
const rows = await db.select({table: 'book', where: `@@id(${db.esc(ids)})`});
|
||
callback(0.6);
|
||
await utils.sleep(100);
|
||
|
||
const bookArr = new Map();
|
||
for (const row of rows)
|
||
bookArr.set(row.id, row);
|
||
|
||
const abRows = [];
|
||
for (const a of authorChunk) {
|
||
const aBooks = [];
|
||
for (const id of a.bookId) {
|
||
const rec = bookArr.get(id);
|
||
aBooks.push(rec);
|
||
}
|
||
|
||
abRows.push({id: a.id, author: a.author, books: JSON.stringify(aBooks)});
|
||
|
||
delete a.bookId;//в дальнейшем не понадобится, authorArr сохраняем без него
|
||
}
|
||
|
||
callback(0.7);
|
||
await db.insert({
|
||
table: 'author_book',
|
||
rows: abRows,
|
||
});
|
||
callback(1);
|
||
};
|
||
|
||
callback({job: 'book sort', jobMessage: 'Сортировка книг', jobStep: 3, progress: 0});
|
||
|
||
//сохранение author_book
|
||
await db.create({
|
||
table: 'author_book',
|
||
});
|
||
|
||
let idsLen = 0;
|
||
let aChunk = [];
|
||
let prevI = 0;
|
||
for (let i = 0; i < authorArr.length; i++) {// eslint-disable-line
|
||
const author = authorArr[i];
|
||
|
||
aChunk.push(author);
|
||
idsLen += author.bookId.length;
|
||
|
||
if (idsLen > 50000) {//константа выяснена эмпирическим путем "память/скорость"
|
||
await saveBookChunk(aChunk, (p) => {
|
||
callback({progress: (prevI + (i - prevI)*p)/authorArr.length});
|
||
});
|
||
|
||
prevI = i;
|
||
idsLen = 0;
|
||
aChunk = [];
|
||
await utils.sleep(100);
|
||
utils.freeMemory();
|
||
await db.freeMemory();
|
||
}
|
||
}
|
||
if (aChunk.length) {
|
||
await saveBookChunk(aChunk, () => {});
|
||
aChunk = null;
|
||
}
|
||
|
||
callback({progress: 1});
|
||
|
||
//чистка памяти, ибо жрет как не в себя
|
||
await db.drop({table: 'book'});
|
||
await db.freeMemory();
|
||
utils.freeMemory();
|
||
|
||
//парсинг 2, подготовка
|
||
const parseField = (fieldValue, fieldMap, fieldArr, authorIds) => {
|
||
if (!fieldValue)
|
||
fieldValue = emptyFieldValue;
|
||
|
||
const value = fieldValue.toLowerCase();
|
||
|
||
let fieldRec;
|
||
if (fieldMap.has(value)) {
|
||
const fieldId = fieldMap.get(value);
|
||
fieldRec = fieldArr[fieldId];
|
||
} else {
|
||
fieldRec = {id: fieldArr.length, value, authorId: new Set()};
|
||
fieldArr.push(fieldRec);
|
||
fieldMap.set(value, fieldRec.id);
|
||
}
|
||
|
||
for (const id of authorIds) {
|
||
fieldRec.authorId.add(id);
|
||
}
|
||
};
|
||
|
||
const parseBookRec = (rec) => {
|
||
//авторы
|
||
const author = splitAuthor(rec.author);
|
||
|
||
const authorIds = [];
|
||
for (const a of author) {
|
||
const authorId = authorMap.get(a);
|
||
if (!authorId) //подстраховка
|
||
continue;
|
||
authorIds.push(authorId);
|
||
}
|
||
|
||
//серии
|
||
parseField(rec.series, seriesMap, seriesArr, authorIds);
|
||
|
||
//названия
|
||
parseField(rec.title, titleMap, titleArr, authorIds);
|
||
|
||
//жанры
|
||
let genre = rec.genre || emptyFieldValue;
|
||
genre = rec.genre.split(',');
|
||
|
||
for (let g of genre) {
|
||
if (!g)
|
||
g = emptyFieldValue;
|
||
|
||
let genreRec;
|
||
if (genreMap.has(g)) {
|
||
const genreId = genreMap.get(g);
|
||
genreRec = genreArr[genreId];
|
||
} else {
|
||
genreRec = {id: genreArr.length, value: g, authorId: new Set()};
|
||
genreArr.push(genreRec);
|
||
genreMap.set(g, genreRec.id);
|
||
}
|
||
|
||
for (const id of authorIds) {
|
||
genreRec.authorId.add(id);
|
||
}
|
||
}
|
||
|
||
//языки
|
||
parseField(rec.lang, langMap, langArr, authorIds);
|
||
};
|
||
|
||
callback({job: 'search tables create', jobMessage: 'Создание поисковых таблиц', jobStep: 4, progress: 0});
|
||
|
||
//парсинг 2, теперь можно создавать остальные поисковые таблицы
|
||
let proc = 0;
|
||
while (1) {// eslint-disable-line
|
||
const rows = await db.select({
|
||
table: 'author_book',
|
||
where: `
|
||
let iter = @getItem('parse_book');
|
||
if (!iter) {
|
||
iter = @all();
|
||
@setItem('parse_book', iter);
|
||
}
|
||
|
||
const ids = new Set();
|
||
let id = iter.next();
|
||
while (!id.done) {
|
||
ids.add(id.value);
|
||
if (ids.size >= 10000)
|
||
break;
|
||
id = iter.next();
|
||
}
|
||
|
||
return ids;
|
||
`
|
||
});
|
||
|
||
if (rows.length) {
|
||
for (const row of rows) {
|
||
const books = JSON.parse(row.books);
|
||
for (const rec of books)
|
||
parseBookRec(rec);
|
||
}
|
||
|
||
proc += rows.length;
|
||
callback({progress: proc/authorArr.length});
|
||
} else
|
||
break;
|
||
|
||
await utils.sleep(100);
|
||
if (config.lowMemoryMode) {
|
||
utils.freeMemory();
|
||
await db.freeMemory();
|
||
}
|
||
}
|
||
|
||
//чистка памяти, ибо жрет как не в себя
|
||
authorMap = null;
|
||
seriesMap = null;
|
||
titleMap = null;
|
||
genreMap = null;
|
||
|
||
utils.freeMemory();
|
||
|
||
//config
|
||
callback({job: 'config save', jobMessage: 'Сохранение конфигурации', jobStep: 5, progress: 0});
|
||
await db.create({
|
||
table: 'config'
|
||
});
|
||
|
||
const stats = {
|
||
recsLoaded,
|
||
authorCount,
|
||
authorCountAll: authorArr.length,
|
||
bookCount,
|
||
bookCountAll: bookCount + bookDelCount,
|
||
bookDelCount,
|
||
noAuthorBookCount,
|
||
titleCount: titleArr.length,
|
||
seriesCount: seriesArr.length,
|
||
genreCount: genreArr.length,
|
||
langCount: langArr.length,
|
||
};
|
||
//console.log(stats);
|
||
|
||
const inpxHashCreator = new InpxHashCreator(config);
|
||
|
||
await db.insert({table: 'config', rows: [
|
||
{id: 'inpxInfo', value: (inpxFilter && inpxFilter.info ? inpxFilter.info : parser.info)},
|
||
{id: 'stats', value: stats},
|
||
{id: 'inpxHash', value: await inpxHashCreator.getHash()},
|
||
]});
|
||
|
||
//сохраним поисковые таблицы
|
||
const chunkSize = 10000;
|
||
|
||
const saveTable = async(table, arr, nullArr, authorIdToArray = true) => {
|
||
|
||
arr.sort((a, b) => a.value.localeCompare(b.value));
|
||
|
||
await db.create({
|
||
table,
|
||
index: {field: 'value', unique: true, depth: 1000000},
|
||
});
|
||
|
||
//вставка в БД по кусочкам, экономим память
|
||
for (let i = 0; i < arr.length; i += chunkSize) {
|
||
const chunk = arr.slice(i, i + chunkSize);
|
||
|
||
if (authorIdToArray) {
|
||
for (const rec of chunk)
|
||
rec.authorId = Array.from(rec.authorId);
|
||
}
|
||
|
||
await db.insert({table, rows: chunk});
|
||
|
||
if (i % 5 == 0) {
|
||
await db.freeMemory();
|
||
await utils.sleep(100);
|
||
}
|
||
|
||
callback({progress: i/arr.length});
|
||
}
|
||
|
||
nullArr();
|
||
await db.close({table});
|
||
utils.freeMemory();
|
||
await db.freeMemory();
|
||
};
|
||
|
||
//author
|
||
callback({job: 'author save', jobMessage: 'Сохранение индекса авторов', jobStep: 6, progress: 0});
|
||
await saveTable('author', authorArr, () => {authorArr = null}, false);
|
||
|
||
//series
|
||
callback({job: 'series save', jobMessage: 'Сохранение индекса серий', jobStep: 7, progress: 0});
|
||
await saveTable('series', seriesArr, () => {seriesArr = null});
|
||
|
||
//title
|
||
callback({job: 'title save', jobMessage: 'Сохранение индекса названий', jobStep: 8, progress: 0});
|
||
await saveTable('title', titleArr, () => {titleArr = null});
|
||
|
||
//genre
|
||
callback({job: 'genre save', jobMessage: 'Сохранение индекса жанров', jobStep: 9, progress: 0});
|
||
await saveTable('genre', genreArr, () => {genreArr = null});
|
||
|
||
//lang
|
||
callback({job: 'lang save', jobMessage: 'Сохранение индекса языков', jobStep: 10, progress: 0});
|
||
await saveTable('lang', langArr, () => {langArr = null});
|
||
|
||
//кэш-таблицы запросов
|
||
await db.create({table: 'query_cache'});
|
||
await db.create({table: 'query_time'});
|
||
|
||
//кэш-таблица имен файлов и их хешей
|
||
await db.create({table: 'file_hash'});
|
||
|
||
callback({job: 'done', jobMessage: ''});
|
||
}
|
||
}
|
||
|
||
module.exports = DbCreator; |