* Add TsFile Docs, delete Json schema

good suggestion
Co-Authored-By: NJialin Qiao <qjl16@mails.tsinghua.edu.cn>

* Fix words that are inconsistent with the document
* Update some screenshots
* Modify the layout of some content

There are still some other mistaken links.

* update tsfile-spark-connector to support spark 2.4.3

* add time cost statstic

* Add documents in Chinese and fix links in English documents

*  New Storage Engine. BufferwriteProcessor, OverflowProcessor, FileNodeProcessor and FileNodeManager are replaced by UnsealedTsFileProcessor, UnsealedTsFileProcessor, StorageGroupProcessor and StorageGroupManager.

* Async Flush Processor

* QueryDataSource is modified to adapt the new storage engine.

* Data dir is modified. TsFile.restore, FileNode.restore are discarded.

* a new tsfile summary file: tsfile.resource.

* Overflow file is isomorphic with Tsfile.

* A new chunk structure: TVList

* Each tsfile has a WAL file.

* Add some config parameters: memTableNumber, chunkBufferPoolEnable. 

* System_dir is discarded.

a small grammar mistake and should be "a device".

* add sync docu

* modify sync docu according to pr review

The word colomn should be column.

* fix mistakes in doc

* add logical part for global time filters from partition problem

* Update README.md, doc, dependency version pom, .travis.yml

* modify scala version to 2.11.8 and restrict java version to 8, change scalatest version from 2.12 to 2.1

* reimplement MTree.toString() by JSON
add MManager.combineMetadataInStrings(String[] metadatas) for cluster module

* move add TIMESERIES HEADER from MGraph to TSServiceImpl
fix bug when MTree combine JSONObjects

* update timeseries metadata format in docs