Classes / functions from drug_markers/meta_analysis

[russHyde]

Copy in the contents of

• [x] `meta.extdata_[download_classes, download_funcs] & tests
• [x] `meta.extdata_import_classes, import_funcs, normalise_funcs] & tests
• [ ] functions defined within marray_importer.R script in drug_markers/meta_analysis project
• [ ] set version to 0.1.1

Then shift the microarray import, normalise, filter steps from marray_importer.R into a script in stemness/differ

[russHyde]

[Download classes / functions] To get the tests to pass

• [x] install XML package

To get check to pass:

• [x] fix file paths in @include directives
• [x] install & add ArrayExpress, GEOquery, RCurl, XML, purrr, tibble to "Import"ed packages
• [x] replace [as_]data_frame with [as_]tibble
• [x] replace functions from drug.markers with equivalents from miiq
  • get_or_else from data_manipulation.R (plus tests for the same)
• [x] Ensure all args for exported functions are documented

To ensure tests are self-contained (ie, don't require access to external files or databases):

• ?

To ensure functions are in appropriate files:

• [ ] move is_[gpl|gds|...]_[acc|uid] into id_functions.R fromdownload-functions.R``

To pass (most of) goodpractice:

• [x] remove sapply
• [ ] additional tests on exported functions
  • [x] dl_gds_to_gpl - no IDs in the input (done - but covr doesn't know)
    • [x] replace tests with mocks (since NCBI is blocking me)
  • [ ] dl_geo_processed - [any tests] non-existent dir, non-GEO accession, ...
  • [ ] dl_geo_raw - [any tests] nonexistent dir, non-GSE accession, ...
  • [ ] dl_aryx_raw - [any tests]

To get the NCBI eutils searches to work, modify the eutils search string to include api_key=API_KEY

• should only affect dl_gds_to_gpl

---

# Check dependencies between arrayExpress and XVector / GenomicRanges versions
On use_package("ArrayExpress")
Warning messages:
1: no function found corresponding to methods exports from ‘XVector’ for: ‘concatenateObjects’ 
2: no function found corresponding to methods exports from ‘GenomicRanges’ for: ‘concatenateObjects’ 
3: multiple methods tables found for ‘dims’

[russHyde]

[Import classes / functions] To get tests to pass:

• ...

To get check to pass:

• [x] "Import" packages: oligo, oligoClasses
• [x] remove dependency on drug.markers
  • [x] preprocess_eset_workflow
  • [x] keep.probe.fn
  • [x] gld_fnDefault_esetAnnotation
• Copy in:
  • [x] add_entrez_ids_to_esets
  • [x] gld_fnDefault_transformAndProbeFilter
  • [ ] get_keep_sample_function
    • [ ] can't easily do this

To ensure tests are self-contained:

• [x] mock out the use of org.[Hs|Mm].eg.db in tests for add_entrez_ids_to_esets

To ensure functions are in appropriate files:

• ...

To ensure files are collated in the correct order:

• ...

TODO:

• [x] - ADD PreprocessConfig that encodes the keep_sample_fn, keep_probe_fn, normalise_fn etc

[russHyde]

This is mostly done now