Adds support for workflow runs where Directory was used as an input or output parameter.
CWLProv does not keep track of directories as file system entries, but rather as collections of files; more specifically, as dictionaries where values are the file entities and keys are their IDs. This is only reflected in the provenance metadata, while actual files under data/ are all stored at the same level: data/XY, where XY are the first two characters of the file's SHA1 checksum in hex format. Since RO-Crate's Datasets map to file system directories, for each dir entry the converter creates an actual directory in the crate, with name equal to the checksum of the checksums of all files in the directory.
Adds support for workflow runs where
Directorywas used as an input or output parameter.CWLProv does not keep track of directories as file system entries, but rather as collections of files; more specifically, as dictionaries where values are the file entities and keys are their IDs. This is only reflected in the provenance metadata, while actual files under
data/are all stored at the same level:data/XY, where XY are the first two characters of the file's SHA1 checksum in hex format. Since RO-Crate'sDatasets map to file system directories, for each dir entry the converter creates an actual directory in the crate, with name equal to the checksum of the checksums of all files in the directory.