Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/netease-im/electron-fulltext-search-demo

electron 全文搜索方案的一个 demo
https://github.com/netease-im/electron-fulltext-search-demo

Last synced: about 1 month ago
JSON representation

electron 全文搜索方案的一个 demo

Host: GitHub
URL: https://github.com/netease-im/electron-fulltext-search-demo
Owner: netease-im
License: cc0-1.0
Created: 2021-02-25T02:37:56.000Z (almost 4 years ago)
Default Branch: master
Last Pushed: 2023-04-04T13:59:04.000Z (over 1 year ago)
Last Synced: 2024-08-05T17:37:45.271Z (4 months ago)
Language: JavaScript
Size: 322 KB
Stars: 8
Watchers: 5
Forks: 0
Open Issues: 3
Metadata Files:
- Readme: README.md
- License: LICENSE.md

Awesome Lists containing this project

awesome-hacking-lists - netease-im/electron-fulltext-search-demo - electron 全文搜索方案的一个 demo (JavaScript)

README

        # ELECTRON-FULLTEXT-SEARCH-DEMO

electron fulltext demo by [search-index](https://github.com/fergiemcdowall/search-index), support Chinese

node 12+

## how to start

git clone 后

```

> npm install

> npm run start

```

## 结合 NIM 接入全文索引

参照 renderer.js

1. 引入并初始化，打开 search-index 数据库

```js

var nodejieba = window.require('nodejieba');

const si = require('search-index')

var SDK = require('./lib/NIM_Web_SDK_v8.3.0_test')

let searchDB

si({

  name: 'ELECTRON-FULLTEXT-SEARCH-DB',

  storeVectors: true

}).then((result) => {

  searchDB = result;

})

```

2. 定义存、取函数

```js

// 用于过滤无意义的符号

const ignoreChars = " \t\r\n~!@#$%^&*()_+-=【】、{}|;':\"，。、《》？αβγδεζηθικλμνξοπρστυφχψωΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩ。，、；：？！…—·ˉ¨‘’“”々～‖∶＂＇｀｜〃〔〕〈〉《》「」『』．〖〗【】（）［］｛｝ⅠⅡⅢⅣⅤⅥⅦⅧⅨⅩⅪⅫ⒈⒉⒊⒋⒌⒍⒎⒏⒐⒑⒒⒓⒔⒕⒖⒗⒘⒙⒚⒛㈠㈡㈢㈣㈤㈥㈦㈧㈨㈩①②③④⑤⑥⑦⑧⑨⑩⑴⑵⑶⑷⑸⑹⑺⑻⑼⑽⑾⑿⒀⒁⒂⒃⒄⒅⒆⒇≈≡≠＝≤≥＜＞≮≯∷±＋－×÷／∫∮∝∞∧∨∑∏∪∩∈∵∴⊥∥∠⌒⊙≌∽√§№☆★○●◎◇◆□℃‰€■△▲※→←↑↓〓¤°＃＆＠＼︿＿￣―♂♀┌┍┎┐┑┒┓─┄┈├┝┞┟┠┡┢┣│┆┊┬┭┮┯┰┱┲┳┼┽┾┿╀╁╂╃└┕┖┗┘┙┚┛━┅┉┤┥┦┧┨┩┪┫┃┇┋┴┵┶┷┸┹┺┻╋╊╉╈╇╆╅╄";

// 增，改

async function putFts(msgs) {

  if (Object.prototype.toString.call(msgs) !== '[object Array]') {

    msgs = [msgs]

  }

  // 分词，并过滤无意义的符号

  var fts = msgs.filter(msg => msg.text && msg.idClient).map(msg => ({

    idx: nodejieba.cut(msg.text).filter(word => !ignoreChars.includes(word)),

    _id: msg.idClient

  }))

  await searchDB.PUT(fts);

  console.log('search-index save success ', fts)

}

// 查

async function queryFts(text, limit = 100) {

  var searchParams = nodejieba.cut(text).filter(word => !ignoreChars.includes(word))

  let records = await searchDB.QUERY({

    SEARCH: searchParams

  })

  let idClients = records.RESULT.map(item => item._id).slice(0, limit);

  if (!idClients) {

    console.log('查询本地消息，无匹配词')

    return;

  }

  window.nim.getLocalMsgsByIdClients({

    idClients,

    done: function (err, obj) {

      console.log('查询本地消息' + (!err ? '成功' : '失败'), err, obj);

    }

  })

}

// 删

async function deleteFts(_ids) {

  if (Object.prototype.toString.call(_ids) !== '[object Array]') {

    _ids = [_ids]

  }

  await searchDB.DELETE(_ids)

  console.log('成功删除：', _ids)

}

```

3. SDK 初始化，接收消息事件拦截

```js

window.nim = SDK.NIM.getInstance({

  debug: true,

  appKey: '******',

  account: '******',

  token: '******',

  // ...... 省略其他参数声明

  /* 消息 */

  onroamingmsgs: function(obj) {

    putFts(obj)

  },

  onofflinemsgs: function(obj) {

    putFts(obj)

  },

  onmsg: function(obj) {

    putFts(obj)

  },

  // ...... 省略其他参数声明

});

```

4. 处理调用会更改本地 msgs 的逻辑

```js

// 发送逻辑

window.nim.sendText({

  scene: 'p2p',

  to: '******',

  text: '春眠不觉晓',

  done(err, obj) {

    if (err) return

    // 发送失败的时候可能无 idClient

    if (!obj.idClient) return

    putFts(obj)

  }

})

// 直接插入本地消息

window.nim.saveMsgsToLocal({

  msgs: msgs,

  done(err, obj) {

    if (err) return

    putFts(obj)

  }

})

// 撤回消息

window.nim.deleteMsg({

  msg: {

    scene: "p2p",

    sessionId: "p2p-******",

    idClient: 'ef1a89e0aaa*******12391d7db22849'

  },

  done(err, obj) {

    if (err) return

    // 删除该 idClient 的记录

    deleteFts('ef1a89e0aaa*******12391d7db22849')

  }

})

// 本地删除消息

window.nim.deleteLocalMsg({

  msg: {

    scene: "p2p",

    sessionId: "p2p-******",

    idClient: 'ef1a89e0aaa*******12391d7db22849'

  },

  done(err, obj) {

    if (err) return

    // 删除该 idClient 的记录

    deleteFts('ef1a89e0aaa*******12391d7db22849')

  }

});

```

5. 若是希望拉取数据库存量的 msg 记录做索引，可尝试使用 getLocalMsgs 获取本地消息。

```js

window.nim.getLocalMsgs({

  // 30 天前

  start: new Date().getTime() - 1000 * 60 * 60 * 24 * 30,

  end: new Date().getTime(),

  limit: Infinity,

  done(err, obj) {

    if (err) return

    // 删除该 idClient 的记录

    putFts(obj.msgs)

  }

});

```

附 [search-index API 文档](https://github.com/fergiemcdowall/search-index) 其他常用的 api 说明

```

// 清除 search-index 中所有记录项

searchDB.FLUSH().then(console.log)

// 打印所有 searchDB 中的记录项

searchDB.ALL_DOCUMENTS().then(console.log)

```

## Question

1. 加入切换了账号登录，账号间的数据怎么做隔离

si 的 name 代表这打开哪个数据库，换个数据库名存储。

```js

const si = require('search-index')

let searchDB

si({

  name: 'ELECTRON-FULLTEXT-SEARCH-DB-WITH-YOUR-ACCOUNT',

  storeVectors: true

}).then((result) => {

  searchDB = result;

})

```

## reference resources

* [nodejieba](https://github.com/yanyiwu/nodejieba)

* [search-index](https://github.com/fergiemcdowall/search-index)

* [How to access Google Chrome's IndexedDB/LevelDB files?](https://stackoverflow.com/questions/35074659/how-to-access-google-chromes-indexeddb-leveldb-files)