Manifest Types

A manifest is a file provided by a Stage 1 pipeline or data provider in a project zip that enumerates a related set of data samples. Scripts in the manifestTypes folder:

• declare expected metadata in a specific type of manifest file
• provide handler functions used by sourceFileUpload and related modules

All manifestTypes, e.g., manifestTypes$xyz, must be a list with members:

manifestTypes$xyz <- list(
    patterns = character(),
    load = function(file) ...,
    parse = function(data.frame) ...
)

where:

• $patterns = file suffixes for manifest files that match the manifestType
• $load = a function to read the manifest file from disk
• $parse = a function that takes the data frame from ‘load’ and returns a list of two data frames:
  • $manifest = all input rows, potentially with modified column names/data types
  • $unique = one row per unique sample, often just a repeat of $manifest

In addition to any desired sample value columns, the parsed manifest data frames must ALWAYS have Project, Sample_ID, and Description columns - if necessary, create and fill those columns with a constant value (e.g., ‘all_samples’ or ‘NA’). ‘Project’ and ‘Sample_ID’ are concatenated to create unique sample identifiers, while ‘Description’ is a short, human readable, sample name that is the default value for the editable sample name shown in the UI.

Data frame $parsed$unique may optionally have Yield and Quality columns, when appropriate (they will be displayed as NA if absent).