fix: wikipedia data dump POC improvements
This commit is contained in:
parent
3bed3e0578
commit
8dec198afe
4
.gitignore
vendored
4
.gitignore
vendored
@ -23,8 +23,8 @@ build/
|
||||
bin/
|
||||
cache.json
|
||||
data/dump
|
||||
data/sql/2-pages-inserts.sql
|
||||
data/sql/3-internal-links-inserts.sql
|
||||
data/sql/*
|
||||
!data/sql/0000-tables-create.sql
|
||||
|
||||
# debug
|
||||
npm-debug.log*
|
||||
|
2
TODO.md
2
TODO.md
@ -12,7 +12,7 @@
|
||||
- [x] `page.sql` (`pages` tables)
|
||||
- [x] `pagelinks.sql` (`internal_links` tables)
|
||||
- [ ] Import SQL files => Investigate why there is an error when importing
|
||||
- [ ] Try `SELECT count(*) FROM internal_links il WHERE il.source_id = (SELECT p.id FROM pages p WHERE p.title = 'Linux'); -- Count of internal links for 'Linux' page`
|
||||
- [ ] Try `SELECT count(*) FROM internal_links il WHERE il.from_page_id = (SELECT p.id FROM pages p WHERE p.title = 'Linux'); -- Count of internal links for 'Linux' page`
|
||||
- [ ] Move from POC (Proof of concept) in `data` folder to `apps/cli` folder
|
||||
- [ ] Documentation how to use + Last execution date
|
||||
- [ ] Rewrite bash script to download and extract SQL files from Wikipedia Database Dump to Node.js for better cross-platform support and easier maintenance + automation, preferably one Node.js script to generate everything to create the database
|
||||
|
@ -2,7 +2,7 @@
|
||||
|
||||
```sh
|
||||
./download-wikipedia-dump.sh
|
||||
node --max-old-space-size=10096 database-wikipedia.js
|
||||
node --max-old-space-size=8096 database-wikipedia.js
|
||||
```
|
||||
|
||||
## Utils
|
||||
@ -13,6 +13,12 @@ Show the first 10 characters of sql file: `head -c 10 ./dump/page.sql`
|
||||
|
||||
To inspect volume size used by database: `docker system df -v | grep 'wikipedia-solver-mariadb-data'`
|
||||
|
||||
To enter in the database container: `docker exec -it wikipedia-solver-database sh`
|
||||
|
||||
Then: `mariadb --password="${DATABASE_PASSWORD}" --user="${DATABASE_USER}"`
|
||||
|
||||
And `use wikipedia;`, for example: `SELECT * FROM pages LIMIT 10;` or to execute a SQL script: `source /docker-entrypoint-initdb.d/3-internal-links-inserts.sql;`.
|
||||
|
||||
## Remove a volume
|
||||
|
||||
```sh
|
||||
|
@ -1,9 +1,14 @@
|
||||
import fs from "node:fs"
|
||||
import path from "node:path"
|
||||
import { extractRowsFromSQLValues, swapKeysAndValues } from "./utils.js"
|
||||
import {
|
||||
extractRowsFromSQLValues,
|
||||
swapKeysAndValues,
|
||||
zeroPad,
|
||||
} from "./utils.js"
|
||||
|
||||
const SQL_DUMP_PATH = path.join(process.cwd(), "dump")
|
||||
const SQL_OUTPUT_PATH = path.join(process.cwd(), "sql")
|
||||
const SQL_FILENAME_NUMBER_PAD = 4
|
||||
|
||||
/**
|
||||
* @typedef {Record<string, number>} WikipediaPagesKeyTitle
|
||||
@ -110,7 +115,10 @@ const wikipediaPagesKeyId = await cleanPagesSQL()
|
||||
|
||||
const cleanPagesSQLWriteToFile = async () => {
|
||||
console.log("cleanPagesSQLWriteToFile - Writing to file...")
|
||||
const sqlOutputPath = path.join(SQL_OUTPUT_PATH, "2-pages-inserts.sql")
|
||||
const sqlOutputPath = path.join(
|
||||
SQL_OUTPUT_PATH,
|
||||
`${zeroPad(1, SQL_FILENAME_NUMBER_PAD)}-pages-inserts.sql`,
|
||||
)
|
||||
const INSERT_INTO_START_OUTPUT = "INSERT INTO pages (id, title) VALUES "
|
||||
|
||||
const wikipediaPagesString = Object.entries(wikipediaPagesKeyId)
|
||||
@ -129,12 +137,6 @@ const cleanPagesSQLWriteToFile = async () => {
|
||||
|
||||
await cleanPagesSQLWriteToFile()
|
||||
|
||||
const sqlOutputPath = path.join(SQL_OUTPUT_PATH, "3-internal-links-inserts.sql")
|
||||
const INSERT_INTO_START_OUTPUT =
|
||||
"INSERT INTO internal_links (from_page_id, to_page_id) VALUES "
|
||||
const sqlOutputFile = await fs.promises.open(sqlOutputPath, "w")
|
||||
await sqlOutputFile.appendFile(INSERT_INTO_START_OUTPUT, "utf-8")
|
||||
|
||||
/**
|
||||
* Function to clean the `pagelinks.sql` file by:
|
||||
* - Removing all lines that don't start with `INSERT INTO...`.
|
||||
@ -143,6 +145,10 @@ await sqlOutputFile.appendFile(INSERT_INTO_START_OUTPUT, "utf-8")
|
||||
* @returns {Promise<void>}
|
||||
*/
|
||||
const cleanInternalLinksSQL = async () => {
|
||||
let internalLinksFileCount = 2
|
||||
const INSERT_INTO_START_OUTPUT =
|
||||
"INSERT INTO internal_links (from_page_id, to_page_id) VALUES "
|
||||
|
||||
/**
|
||||
* @type {WikipediaPagesKeyTitle}
|
||||
*/
|
||||
@ -157,17 +163,27 @@ const cleanInternalLinksSQL = async () => {
|
||||
let current = ""
|
||||
let lastPercent = 0
|
||||
|
||||
const BATCH_SIZE = 10_000
|
||||
const BATCH_SIZE = 4_000_000
|
||||
|
||||
/**
|
||||
* @type {string[]}
|
||||
*/
|
||||
let batch = []
|
||||
|
||||
const flushBatch = async (isLast = false) => {
|
||||
const flushBatch = async () => {
|
||||
if (batch.length > 0) {
|
||||
const batchString = batch.join(",") + (isLast ? ";" : ",")
|
||||
await sqlOutputFile.appendFile(batchString, "utf-8")
|
||||
const batchString = batch.join(",")
|
||||
const fileName = `${zeroPad(internalLinksFileCount, SQL_FILENAME_NUMBER_PAD)}-internal-links-inserts.sql`
|
||||
const sqlOutputPath = path.join(SQL_OUTPUT_PATH, fileName)
|
||||
await fs.promises.writeFile(
|
||||
sqlOutputPath,
|
||||
`${INSERT_INTO_START_OUTPUT}${batchString};`,
|
||||
{
|
||||
encoding: "utf-8",
|
||||
},
|
||||
)
|
||||
console.log(`flushBatch - ${fileName}, batch.length: ${batch.length}`)
|
||||
internalLinksFileCount += 1
|
||||
batch = []
|
||||
}
|
||||
}
|
||||
@ -218,7 +234,7 @@ const cleanInternalLinksSQL = async () => {
|
||||
|
||||
if (plFromNamespace === "0" && plTargetNamespace === "0") {
|
||||
const toPageId = wikipediaPagesKeyTitle[plTargetTitle]
|
||||
if (toPageId != null) {
|
||||
if (toPageId != null && wikipediaPagesKeyId[plFromPageId] != null) {
|
||||
/**
|
||||
* @type {WikipediaInternalLink}
|
||||
*/
|
||||
@ -229,19 +245,21 @@ const cleanInternalLinksSQL = async () => {
|
||||
batch.push(
|
||||
`(${wikipediaInternalLink.fromPageId},${wikipediaInternalLink.toPageId})`,
|
||||
)
|
||||
|
||||
if (batch.length >= BATCH_SIZE) {
|
||||
await flushBatch()
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (batch.length >= BATCH_SIZE) {
|
||||
sqlInputFileStream.pause()
|
||||
await flushBatch()
|
||||
sqlInputFileStream.resume()
|
||||
}
|
||||
})
|
||||
.on("error", (error) => {
|
||||
return reject(error)
|
||||
})
|
||||
.on("close", async () => {
|
||||
await flushBatch(true)
|
||||
await flushBatch()
|
||||
console.log("cleanInternalLinksSQL - Bytes read (100%).")
|
||||
return resolve()
|
||||
})
|
||||
|
@ -11,12 +11,12 @@ CREATE TABLE `pages` (
|
||||
-- > War on varchar. Changed all occurrences of varchar(N) and varchar(N) binary to varbinary(N). varchars cause problems ("Invalid mix of collations" errors) on MySQL databases with certain configs, most notably the default MySQL config.
|
||||
|
||||
CREATE TABLE `internal_links` (
|
||||
`id` INT(8) UNSIGNED NOT NULL AUTO_INCREMENT,
|
||||
-- `id` INT(8) UNSIGNED NOT NULL AUTO_INCREMENT,
|
||||
`from_page_id` INT(8) UNSIGNED NOT NULL,
|
||||
`to_page_id` INT(8) UNSIGNED NOT NULL,
|
||||
|
||||
PRIMARY KEY (`id`)
|
||||
-- PRIMARY KEY (`from_page_id`, `to_page_id`),
|
||||
-- FOREIGN KEY (`from_page_id`) REFERENCES `pages` (`id`) ON DELETE CASCADE,
|
||||
-- FOREIGN KEY (`to_page_id`) REFERENCES `pages` (`id`) ON DELETE CASCADE
|
||||
-- PRIMARY KEY (`id`)
|
||||
PRIMARY KEY (`from_page_id`, `to_page_id`),
|
||||
FOREIGN KEY (`from_page_id`) REFERENCES `pages` (`id`) ON DELETE CASCADE,
|
||||
FOREIGN KEY (`to_page_id`) REFERENCES `pages` (`id`) ON DELETE CASCADE
|
||||
) ENGINE=InnoDB DEFAULT CHARSET=binary ROW_FORMAT=COMPRESSED;
|
@ -74,3 +74,15 @@ export const swapKeysAndValues = (object) => {
|
||||
}),
|
||||
)
|
||||
}
|
||||
|
||||
/**
|
||||
*
|
||||
* @param {number} number
|
||||
* @param {number} places
|
||||
* @returns {string}
|
||||
* @example zeroPad(1, 2) // '01'
|
||||
* @example zeroPad(10, 2) // '10'
|
||||
*/
|
||||
export const zeroPad = (number, places = 2) => {
|
||||
return number.toString().padStart(places, "0")
|
||||
}
|
||||
|
Reference in New Issue
Block a user