Workflow let you create a pipeline of subfeature by inputing an entry data, letting it pass through all the subfeatures and retreiving the result
Note: You can try Workflow on our developer portal and a description
snippet will automatically get generated as you build your workflow
Example:
Schema: ocr --> automatic_translation --> summarize
In this workflow we pass a file as input, ocr parse the text, pass it to translate, and summarize will return a summary of the translated text
Inputs:
Depending on the first subfeature you have three choice for the inital input data: file
, text
or texts
.
To create a workflow you have to pass an object to the description
parameter, the object should look like this:
[
{
"feature": "ocr",
"subfeature": "ocr",
"params": {
"language": "auto-detect",
"providers": "google"
}
},
{
"feature": "translation",
"subfeature": "automatic_translation",
"params": {
"source_language": "auto-detect",
"target_language": "fr",
"providers": "google"
}
},
{
"feature": "text",
"subfeature": "summarize",
"params": {
"providers": "openai"
}
}
]
description
should be a list of nodes each node representing a subfeature.Inside each node, enter the desired feature and subfeature and its params.params
are the parameters you should normally send to the subfeature as if you were doing a direct call but with a few constraints:providers
should take only one provider name, if you input multiple providers, the response of all the other providers will be ignoredfile
,text
,texts
shouldn't be present inparams
as these are initial inputs
Important!: description should be sent as a string since the content-type is a form-data
return_only_last
allows you to chose wether you want a list of all the response returned by each providers or just getting the last response. If set totrue
the response will be the last subfeature result, if set tofalse
the reponse will be a list of all subfeature results.