File indexing service API

File indexing service is implemented using 2 services:

• XSODATA service
  • Used for full-text search in files
  • Also possible to create files and rows, but for more convenient API xsjs service is more suitable
• XS javascript (XSJS) service
  • Used for creating files, appending rows, etc.

Service authorizations

Indexing agent application should use database user with assigned role 'dvd.filesearch.roles::agent'.

Technical information

Implementation of File indexing agent

Analyzed file should be parsed by file index agent in way where one row contains information (string) about one event from analyzed file. File agent should be able to identify time of particular event and get information for this event. 

In example: We have following file content:

2017-08-16 16:21:21 System is overloaded
2017-08-16 16:21:56 System is OK again

File indexing agent should be able parse file into 2 rows where first row will be parsed into:

• timestamp: 20170816162156
• content: 2017-08-16 16:21:21 System is overloaded

Workflow for implementation File Indexing agent is following:

• Create File origin (id of location where file original resides) using OData entity FileOrigin
  

  dvd.filesearch::FileIndex.FileOrigin

  Origin header table where all origin definition are stored.

  Field	Key	Type	Description
  id	X	NVARCHAR 32	Unique identifier of file origin (i.e. system name where files are coming from)
  name		NVARCHAR 128	Name of origin
  host		NVARCHAR 128	Server host

• Create File (full path of file location) using OData entity File
  

  dvd.filesearch::FileIndex.File

  File header table where all file definitions are stored.

  Field	Key	Type	Description
  id	X	INTEGER	Unique identifier of file automatically assigned during entry creation
  guid		NVARCHAR 32	Unique identifier of file given by File Indexing agent application
  name		NVARCHAR 255	Full name of file (including full path in origin)
  origin.id		32	File origin
  fileType		100	Type of file

• Create File content using XSJS script /dvd/filesearch/addFileRows.xsjs  

  dvd.filesearch::FileIndex.FileData

  Data that represent content of defined files. 

  Field	Key	Type	Description
  fileId	X	INTEGER	Unique identifier of file
  rowNum	X	INTEGER	Row number of record
  timestamp		TIMESTAMP	Time stamp of record
  content		CLOB	String content of record

XSODATA service

CDS are exposed to oData entities using XSODATA definition file. OData are mainly used by SAPUI5 application and File indexing agent.

More information about properties can be found from metadata of service: 

/dvd/filesearch/FileIndex.xsodata/$metadata/

XSJS service

All request should contain following headers and cookies gathered during initial fetch of X-CSRF-Token token.

addFileRows.xsjs

Script path

/dvd/filesearch/addFileRows.xsjs

Request

HTTP method: POST

Request body content example:

{
	"guid": "CB1D315DD10C334792A32702B834DAF6",
	"rows": [	{"timestamp":"20170321103106","content":"2017-03-21 10:31:06 Show me the way to the next whisky bar"},
			{"timestamp":"20170321103106","content":"2017-03-21 10:31:06 and don't ask why, and don't ask why"},
			{"timestamp":"20170321103106","content":"2017-03-21 10:31:06 What if we don't find the next whisky bar?"},
			{"timestamp":"20170321103106","content":"2017-03-21 10:31:06 I'll tell you we must die, I'll tell you we must die!"}
			]
}

Response

Data are successfully added when response contains HTTP code 2* and response body should look as follows:

{"result":"OK"}

deleteFileRows.xsjs

Script path

/dvd/filesearch/deleteFileRows.xsjs

Request

HTTP method: POST

Request body content example:

{
	"guid": "CB1D315DD10C334792A32702B834DAF6"
}

Where guid is identificator of file in which content should be deleted.

Response

Data are successfully deleted when response contains HTTP code 2* and response body should look as follows:

{"result":"OK"}