Skip to content

Releases: byzer-org/byzer-lang

MLSQL v1.1.6 Release

18 Dec 02:12
@allwefantasy allwefantasy
v1.1.6
3bda721
Compare
Choose a tag to compare
MLSQL v1.1.6 Release

We are glad to announce that MLSQL v1.1.6 is released. There are huge new features, bug fix and code refractor in this half month' development:

  1. Almost 163 commits.
  2. 28 ISSUE
  3. 41 PR. 25 of them are BugFix, 26 of them are improvements.

Improvements：

  • PR-747: Load balance, Dynamic Resource Allocation,Instances manager based on labels。More details please check：mlsql-cluster doc.
  • PR-735: Supports Using load statement to view API doc and Configuration of MLSQL。
  • PR-728: Upgrade to latest carbondata
  • PR-781: A new transformer supports complex father-children relation computing which is hard if you use SQL to do this.
  • PR-794: Provides shell/cmd script to make you compile/package MLSQL more easy.

BugFix：

  • PR-773： The last As keyword is upper will cause exception。
  • PR-777: Save partition by can not support multi columns.
  • PR-760: conflict in multi kafka-client.（MLSQLsupport kafka 0.8/0.9/1.x at the same time）
  • PR-757: MLSQL Auth bug
  • PR-795： PS cluster enabled will fail the exetuor in yarn mode.

Code refractor：

  • PR-727： Remove the support for Spark 1.6.x
  • PR-732： Move some non-core modules to external/contri directory.
Assets 5
Loading

MLSQL v1.1.5 Release

29 Nov 07:12
@allwefantasy allwefantasy
v1.1.5
8ccf38a
Compare
Choose a tag to compare
MLSQL v1.1.5 Release

We are glad to announce that MLSQL-v1.1.5 version released.

V1.1.5 is a maintainable release of 1.1.x which is based on branch b1.1.5.

This release we have done a lot of things to make MLSQL work on spark 2.4.0. The changes applied in this release can be tracked by All Issue tagged with MSLQL-1.1.5 and All PR tagged with MSLQL-1.1.5

MLSQL 1.1.5 still not contains following jars by default:

  1. ansj_seg-5.1.6.jar
  2. nlp-lang-1.7.8.jar

If you want to analyze text content, please download it manually and add them to --jars.

We plan to release v1.1.6 in next two weeks.

Usage:

  1. download distribution based on your spark version.
  2. unzip and make sure SPARK_HOME have be confgiured.
  3. use ./start-local.sh to start MLSQL server.
Assets 5
Loading

MLSQL v1.1.4 Release

18 Nov 08:02
@allwefantasy allwefantasy
b1.1.4
5d4232a
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MLSQL v1.1.4 Release

V1.1.4 is a maintainable release of 1.1.x which is based on branch b1.1.4.

The changes applied in this release can be tracked by All Issue tagged with MSLQL-1.1.4 and All PR tagged with MSLQL-1.1.4

Notice that we do not tag the issue before, so there are many bugfix and features may not be tracked by the tag MLSQL-1.1.4. However, since this release, we will make sure that all Issue And PR tracked.

MLSQL v1.1.4 still not contains by default:

  1. ansj_seg-5.1.6.jar
  2. nlp-lang-1.7.8.jar

If you want to analyze text content, please download it manually and add them to --jars.

We plan to release v1.1.5 in next two weeks.

Usage:

  1. download distribution based on your spark version.
  2. unzip and make sure SPARK_HOME have be confgiured.
  3. use ./start-local.sh to start MLSQL server.
Assets 5
Loading

v1.1.3-dev（dev preview version）

13 Oct 09:19
@allwefantasy allwefantasy
v1.1.3-dev
e2403f8
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.1.3-dev（dev preview version） Pre-release
Pre-release

https://pan.baidu.com/s/1YCiKn0p65g8X1Gp6Yb7k1g

Assets 2
Loading

StremaingPro V1.1.3 Released

29 Aug 08:06
@allwefantasy allwefantasy
v1.1.3
830327a
Compare
Choose a tag to compare
StremaingPro V1.1.3 Released

There are really a bunch of features added to this release.

Main features:

  1. MLSQL syntax check, any grammar error will be thrown as an exception.
  2. Enhance the stream job support mlsql-stream
  3. JDBC in stream/bach mlsql-jdbc
  4. Scala/Python script UDF support mlsql-script-udf
  5. include statement make MLSQL snippet is reusable.
  6. job manager in build-in web UI.
Assets 3
Loading

StreamingPro v1.1.2

19 Jul 02:52
@allwefantasy allwefantasy
1.1.2
5225c5a
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
StreamingPro v1.1.2

功能特性：

  1. 初步增加tensorflow 分布式训练支持 #352
  2. 模型支持多版本 #331
  3. 优化了PythonAlg模块的功能，包括训练数据分发等。
  4. 修正了很多bug

NLP分词相关的功能依赖ansj分词包，可到最早的那个版本中下载附件：https://github.com/allwefantasy/streamingpro/releases

Assets 3
Loading

StreamingPro v1.1.1版本发布

03 Jul 02:50
@allwefantasy allwefantasy
v1.1.1
4bdac6d
Compare
Choose a tag to compare
StreamingPro v1.1.1版本发布

StreamingPro 经过实际生产环境的验证，添加了很多针对算法的功能，近期大部分PR都是针对算法的。

更多细节可参看PR列表

有用的功能有：

  1. PythonAlg模块具备同步资源文件的能力自定义python算法支持自定义资源文
  2. 更新了交互界面，使用更友好。web交互
  3. 强化了词向量预处理功能，可以更好的接入深度学习模型SQLWord2VecInPlace 功能扩展
  4. 可以对训练集各个分类的数据进行统计，计算分类权重，可以查看应训练好的sklearn模型参数，支持混淆矩阵等
  5. 大量矩阵运算的UDF函数加入
  6. 准备了两个新的模块，方便支持spark 2.3.x
  7. 添加了图片数据源，实现了java/opencv对图片的缩放功能
  8. more ....

NLP分词相关的功能依赖ansj分词包，可到之前发布的版本的附件中下载：https://github.com/allwefantasy/streamingpro/releases

Assets 3
Loading

StreamingPro v1.1.0版本发布

25 Apr 08:27
@allwefantasy allwefantasy
v1.1.0
ee31da1
Compare
Choose a tag to compare
StreamingPro v1.1.0版本发布

StreamingPro支持通过配置文件的方式编写Spark应用，包括批处理和流式。

StreamingPro同时也支持将Spark Application作为服务，然后暴露出Rest接口，使用SQL Script作为交互格式。目前支持：

  1. 爬虫，通过SQL脚本完成爬取网页功能。
  2. 支持流式计算SQL脚本
  3. 支持批处理SQL脚本
  4. 支持用SQL做算法训练和预测，以及部署成API服务的能力。
  5. 剥离dl4j，减少jar包大小
  6. 支持carbondata 1.3.1,支持流式写入carbondata
  7. 启动时通过--jars带上 ansj_seg-5.1.6.jar 和nlp-lang-1.7.8.jar 可在MLSQL中运用分词抽词工具

注意： 该预编译版本仅测试过spark 2.2.0。如果需要2.1.0 或者2.3.0 则需要自己进行编译。

Assets 5
Loading