Yo!

Micro Web Crawler in PHP & Manticore

Yo! is the super thin layer for Manticore search server that extends official manticoresearch-php client with CLI tools and simple JS-less WebUI.

Features

• MIME-based crawler with flexible filter settings by regular expressions, selectors, external links etc
• Page snap history with local and remote mirrors support (including FTP protocol)
• CLI tools for index administration and crontab tasks
• JS-less frontend to run local or public search web portal
• API tools to make search index distributed

Components

• Manticore Server
• PHP library for Manticore
• Symfony DOM crawler
• Symfony CSS selector
• FTP client for snap mirrors
• Hostname ident icons
• Bootstrap icons

Install

Environment

Debian

• wget https://repo.manticoresearch.com/manticore-repo.noarch.deb
• dpkg -i manticore-repo.noarch.deb
• apt update
• apt install git composer manticore manticore-extra php-fpm php-curl php-mbstring

Yo search engine uses Manticore as the primary database. If your server sensitive to power down, change default binlog flush strategy to binlog_flush = 1

Deployment

Project in development, to create new search project, use dev-main branch:

• composer create-project yggverse/yo:dev-main

Development

• git clone https://github.com/YGGverse/Yo.git
• cd Yo
• composer update
• git checkout -b pr-branch
• git commit -m 'new fix'
• git push

Update

• cd Yo
• git pull
• composer update

Init

• cp example/config.json config.json
• php src/cli/index/init.php

Usage

• php src/cli/document/add.php URL
• php src/cli/document/crawl.php
• php src/cli/document/search.php '*'

Web UI

1. cd src/webui
2. php -S 127.0.0.1:8080
3. open http://127.0.0.1:8080 in browser

Documentation

CLI

Index

Init

Create initial index

php src/cli/index/init.php [reset]

• reset - optional, reset existing index

Alter

Change existing index

php src/cli/index/alter.php {operation} {column} {type}

• operation - operation name, supported values: add|drop
• column - target column name
• type - target column type, supported values: text|integer

Document

Add

php src/cli/document/add.php URL

• URL - add new URL to the crawl queue

Crawl

php src/cli/document/crawl.php

Clean

Make index optimization, apply new configuration rules

php src/cli/document/clean.php [limit]

• limit - integer, documents quantity per queue

Search

php src/cli/document/search.php '@title "*"' [limit]

• query - required
• limit - optional search results limit

Migration

YGGo

Import index from YGGo database

php src/cli/yggo/import.php 'host' 'port' 'user' 'password' 'database' [unique=off] [start=0] [limit=100]

Source DB fields required:

• host
• port
• user
• password
• database
• unique - optional, check for unique URL (takes more time)
• start - optional, offset to start queue
• limit - optional, limit queue

Backup

Logical

SQL text dumps could be useful for public index distribution, but requires more computing resources.

Read more

Physical

Better for infrastructure administration and includes original data binaries.

Read more

Instances

Yggdrasil

• http://[201:23b4:991a:634d:8359:4521:5576:15b7]/yo/ - IPv6 0200::/7 addresses only | index

Alfis DNS

• http://yo.ygg - .ygg domain zone search only | index
• http://ygg.yo.index - alias of http://yo.ygg | index

**.yo.index reserved for domain-oriented instances e.g. .btn, .conf, .mirror - feel free to request the address