Sources¶
A source is where the files come from. The collate stage reads them from a source
and collates them into accessions, acquisitions and resources (see
Accessions, acquisitions and resources) in a staging directory, ready to be labelled. Each source
provides its own collate subcommand:
- Folders
collate folderreads files from directories on a local path or a mounted file system, e.g. where an instrument exports to. It reads each file's metadata to decide which accession, acquisition and resource it belongs to, and works with any file format that FileFormats can read metadata from.- Orthanc
collate orthancpulls the DICOM studies held by an Orthanc server, for scanners that push to a PACS rather than exporting to a directory. It hardlinks the files out of Orthanc's storage rather than copying them.
Both write the same layout (see Staging directories), so everything after
collate works the same whichever source the data came from.
The sources share these options:
--loopand--wait-period— run continuously, skipping anything still being transferred (see Running continuously and handling errors)--copy-mode— how files are copied into the staging directory, e.g. as hardlinks where possible (the default) or as copies--convert <src-mime-like> <tgt-spec>— convert resources of one format to another as they are collated, e.g.--convert medimage/dicom-series medimage/nifti-gz-x. Converter options go after the target type, e.g.application/zip:compression=ZIP_STORED--logger,--additional-loggerand--raise-errors— logging and error handling (see Running continuously and handling errors)
Run accessionflow collate --help to list the sources that are installed. Sources are
plugins, so a site can add its own (see Writing source and store plugins).