There was something wrong with fetching your inbox. Please contact support here.
Depending on the API, Haven OnDemand provides a set of default pipelines that allow you to provide your input data in different forms:
Text. You provide some text, which the API uses.
File. You provide a file that contains the text or image that you want to analyze.
Reference. You provide a reference obtained from the Haven OnDemand Store Object or Expand Container API. The corresponding object is passed to the API.
URL. You provide a publicly accessible HTTP URL from which the object can be retrieved.
A further distinction is between structured and unstructured input data.
Most of the Haven OnDemand APIs have a central parameter that takes a piece of unstructured information as input for analysis. For many of the APIs, this input is a block of text. For others, such as OCR Document or Image Recognition, you must provide a file.
For APIs that require input text, when you submit a file the API runs the Text Extraction API. The Text Extraction API uses HPE KeyView to extract the text from the API. It returns the extracted text, which the original API uses as text.
For many APIs, you can submit a container file (such as a zip or rar file). In this case, the API runs the Expand Container API to extract the files. For an API that requires input text, the API passes the result files to the Text Extraction API. For APIs that take a file as a direct input (such as the image analysis APIs), they run the API operation on each of the files from the container.