33 adapt write dwc for use by opencpu

[PietrH]

Closes #33
Resolves #35

This function does not mimic the behaviour of etn's write_dwc() one to one, instead will only return a list of dataframes. The handling of writing to a local file, including creating a directory if necessary will need to handled by the client.

---

Motivation of changes + outlook on future

This function actually has two steps,

1. fetch project info
2. query (including mapping to dwc)

With a progress message in the middle. Because the API will handle this in a single request, the progress message doesn't make much sense and was removed.

We could retain this message if we have the client handle the first call to get_animal_projects() and change the inputs to etnservice's write_dwc() to the parsed output of get_animal_projects(). However this would result in a mismatch between the arguments that write_dwc() accepts between etn and etnservice, and thus break a future possible generic sql helper or the option of inheriting documentation between the packages. (see comments in #33)

This doesn't seem worth it to me. Alternatively we could create a new function just for etnservice, eg. get_dwc() with input arguments based on get_animal_projects, and retain a write_dwc() in etnservice that isn't actually used by the client but does allow for inheritance of documentation or use by a (possible future) generic sql helper. For discussion on this last helper, see inbo/etn#280

TODO

• add postman test
• pass R CMD Check
• fix broken examples