Model

Predict

Used to send prediction request and get result-id as response.

POST URL:{{base_url}}/api/{{version}}/model/:id/predict

Path Parameters

Name	Type	Description
id*	String	Model id of a particular ml model.

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

Request Body

Name	Type	Description
file	multipart/form-data	File with extensions png , jpeg and pdf. This field is used to send file for getting prediction from ml models.
version*	multipart/form-data	Used to specify the version of ml model, using which we need to get the prediction. To get prediction using latest version of the model without specifying the correct version , give version value as 0.0. Then the api will use latest available version of the model to get prediction.
splitDocument*	multipart/form-data	It can be "true" or "false". But currently "true" value is only available for some ml models. If the chosen model could not support "true", api will send bad request response. "true" value can give more accurate result if the give pdf file contains more than one page. Please use "true" value in the case of invoice model to increase accuracy.
text	multipart/form-data	Used to get prediction from NER type ml models.

200: OK Returns result_id.

{
    "data": {
        "result_id": "String"
    },
    "message": "String",
    "status": 200
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

401: Unauthorized Invalid API key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

500: Internal Server Error

{
    // Response
}

Used to get result of a particular prediction using result id.

GET URL:{{base_url}}/api/{{version}}/model/result/:id

Path Parameters

Name	Type	Description
id *	String	Result id of a particular prediction.

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

200: OK Value of result field is a dictionary that contains prediction result. Structure of result dictionary may vary with different ml models.

{
    "data": {
        "message":"String",
        "status":"SUCCESS",
        "result":{}
    },
    "message": "String",
    "status": 200
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

401: Unauthorized Invalid api key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

500: Internal Server Error

{
    // Response
}

202: Accepted IN PROGRESS status indicates that the ml model is predicting the result. May take some time to produce the result. Try to call the API after some time to get the result.

{
    "data": {
        "message":"String",
        "status":"IN PROGRESS",
        "result":{}
    },
    "message": "String",
    "status": 202
}

204: No Content FAILED status indicates that the given prediction request is failed due to some reasons. Please contact Visualyze.AI for more information.

{
    "data": {
        "message":"String",
        "status":"FAILED",
        "result":{}
    },
    "message": "String",
    "status": 204
}

Model

Used to list available models.

GET URL:{{base_url}}/api/{{version}}/model/view

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

200: OK Models field contains list of models. Each dictionary in the list is used to represent a unique ml model. Version details of pretrained models will be unavailable to the user.

{
    "data": {
        "models":[]     
    },
    "message": "String",
    "status": 200
}

401: Unauthorized Invalid API key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

Used to get model details.

GET URL:{{base_url}}/api/{{version}}/model/:id/view

Path Parameters

Name	Type	Description
id*	String	Model id of a particular ml model.

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

200: OK Model details dictionary will have the details of ml model.

{
    "data": {
        "model_details":{}   
    },
    "message": "String",
    "status": 200
}

401: Unauthorized Invalid API key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

500: Internal Server Error

{
    // Response
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

Used to get version details of particular model.

GET URL:{{base_url}}/api/{{version}}/model/:id/:version/info

Path Parameters

Name	Type	Description
id*	String	Model id of a particular ml model.
version*	Float	Version of a particular ml model.

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

200: OK Info field contains version based details of a particular ml model. But we do not provide version details of PRETRINED ml models. You may get an empty dictionary in the case of PRETRINED models.

{
    "data": {
        "info": {}
    },
    "message": "successfully retrieved details",
    "status": 200
}

401: Unauthorized Invalid API key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

500: Internal Server Error

{
    // Response
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

Used to delete latest model version.

DELETE URL:{{base_url}}/api/{{version}}/model/:id/delete

Path Parameters

Name	Type	Description
id *	String	Model id of a particular ml model.

Headers

Name	Type	Description
Authorization*	API key
Content-Type*	application/json

200: OK Deleted the latest version of model.

{
    "data": {},
    "message": "deleted successfully",
    "status": 200
}

401: Unauthorized Invalid API key.

{
    "data": {},
    "message": "ERROR: Unauthorized",
    "status": 401
}

500: Internal Server Error

{
    // Response
}

400: Bad Request Message field can be used to identify the issue.

{
    "data": {},
    "message": "String",
    "status": 400
}

You cannot delete PRETRIAINED models using Delete API