Ingesting Data and Models¶
TruEra's Python SDK supports two types of data ingestion: Real-time and batch. Real-time data ingestion entails importing each data item as the source emits it. With batch ingestion, data items are imported in discrete chunks at periodic intervals. This also means that individual files must be validated and data items routed to the intended destination.
Models — well-defined computations formed as a result of an algorithm that uses ingested data as input to produce a result — also need to be ingested for TruEra to properly use them. Importing your model to TruEra over the appropriate pathway depends on whether your model is already packaged and how. In the context of Python, a model bundled as software to be installed like a library or application is a package. In other words, a packaged Python model adheres to a standard structure and thus can be shipped and installed by others. If your model is written in Java or another non-Python language, you'll need to "wrap" it so TruEra can recognize it as a package (see Python Model Wrappers).