Added ERT forward model to convert from .UNSMRY to Arrow file format

[sigurdp]

This PR adds an ERT forward model that can converts .UNSMRY files to Apache Arrow IPC file format.
The Arrow IPC file format is a columnar data format optimized for table-like datasets, see https://arrow.apache.org/docs/format/Columnar.html#ipc-file-format
This file format is also known as Feather version 2, see https://arrow.apache.org/docs/python/feather.html

The ERT job imports data from an .UNSMRY file and writes output to an .arrow file in the above format.

The following data types are used in the exported data

• DATE column is stored as an Arrow timetamp with millisecond resolution timestamp[ms]
• All numeric columns will be stored as 32 bit floats
• Summary meta data will be attached to each field/column in the exported table's schema under the
  'smry_meta' key

Currently CPI columns in the summary data wil be excluded from the exported .arrow file.

---

Contributor checklist

• 🎉 This PR closes ERT forward model .UNSMRY → (.parquet or .arrow) #673
• 🤖 I have added tests, or extended existing tests, to cover any new features or bugs fixed in this PR.
• 📖 I have considered adding a new entry in CHANGELOG.md, and added it if should be communicated there.

[anders-kiaer]

Nice! 🚀 Minor comments only. This will be a great improvement when in production 🏎️ ...and the plugins start using this data source through the data factory.