assemble support for pandas

[daiwaid]

• add assemble support for most transform operations for pandas
• create a custom Map class for mapping info to be more intuitive to use
• add parameter for map to toggle reasoning

NOTE: requires code from Libem #111 to function

[zenodflow]

This is a very neat / nice example for schema alignment for static datasets (known records with fixed size).

Could we have perhaps another example demonstrating this in the message broker setting?

E.g., having a continuous sending data source and the assembled data flow - once "in stalled" to the data path - able to process incoming data.

[zenodflow]

One change: ensure the assemble method reflects the intended semantics, i.e., generating the dataflow that translates incoming data records. E.g., throw a nested function. In next PRs we should support native/non-python dataflow.