Implement "source agnostic" (fetch|with)_data retrieval functions

[lianos]

We have (fetch|with)_sample_covariates() and (fetch|with)_assay_data() functions.

It might be useful for interactive analysis to create a higher level (fetch|with)_data() set of functions. These functions would allow for the specification of the source of the data (feature data, or sample data) to be encoded in the item itself.

For instance, let's say we have a set of samples in a facile_frame named xsamplesthat we want to fetch expression data for CD3E and GAPDH, and some sample-level covariates, like "sex" and "sample_type"

Some ways to do that:

d1 <- xsamples %>%
  with_data(c("feature:CD3E", "sample:sex", "sample:sample_type"))

But how would we know what feature column we should use to look for CD3E? Maybe we specify to search in the "name" column like so:

d1 <- xsamples %>%
  with_data(c("feature:CD3E|name", "sample:sex", "sample:sample_type"))

Further to specify what assay data matrix to use:

d1 <- xsamples %>%
  with_data(c("feature[rnaseq]:CD3E|name", "sample:sex", "sample:sample_type"))

Combined Retrieval of Similar Data Types

Since both "sex" and "sample_type" are coming from sample covariates, let's enable a simplification of retrieval using a ","

d1 <- xsamples %>%
  with_data(c("feature[rnaseq]:CD3E|name", "sample:sex,sample_type"))

And let's say we wanted a second gene as well, like GAPDH

d1 <- xsamples %>%
  with_data(c("feature[rnaseq]:CD3E|name,GAPDH|name", "sample:sex,sample_type"))