Project status
raft project-status is a read-only command for summarizing the current
state of one RAFT project.
It does not repair, download, delete, rerun, or modify project files. It is safe to run repeatedly while inspecting a project.
Basic usage
Run:
$ raft project-status --project-id my-project
The command reports:
project path and detected workflow context
an overall status label such as
setup complete, not run yet,run failed, orcomplete, reports availablethe manifest RAFT believes the project is using
manifest row, patient, dataset, run, sample-prefix,
Group, andAllelessummarieslatest Nextflow trace status counts and failed task count
presence of
outputs/,outputs/samples/,outputs/qc/, andoutputs/reports/simple output file counts
workflow-specific checks when RAFT recognizes the workflow
a
Nextsection with suggested follow-up commands
Use JSON for automation:
$ raft project-status --project-id my-project --json
Next actions
The Next section is intended to answer “what should I do now?” without
changing project files. The suggestions are derived from detected project
state.
After --setup-only, project-status suggests starting the workflow:
Status: setup complete, not run yet
Next:
Start the workflow:
raft run --project-id my-project
Open the viewer without preloaded outputs:
raft generate-reports --project-id my-project --no-load
After a failed run, it points to the latest trace, the Nextflow log, and the rerun command:
Status: run failed
Next:
Inspect the latest trace:
less /path/to/projects/my-project/outputs/reports/trace-20260419.txt
Inspect the Nextflow log:
less /path/to/projects/my-project/logs/.nextflow.log
After fixing the problem, rerun:
raft run --project-id my-project
When reportable outputs are present, it suggests opening the report viewer and packaging the project:
Status: complete, reports available
Next:
Open the report viewer:
raft generate-reports --project-id my-project
Package the project:
raft package-project --project-id my-project
Manifest detection
A project can contain multiple files under inputs/metadata. project-status
therefore does not assume that the newest or only manifest-like file is the one
used by the workflow.
Detection order is:
Read
workflow/main.nf.Look for manifest-like
Channel.fromPath(...)entries.Resolve
${params.*}expressions using parameter assignments inmain.nf. For example, RAFT resolves${params.metadata_dir}/my.manifest.tsvusing the project’s actualinputs/metadatadirectory.If exactly one referenced manifest exists, report that file as the manifest source.
If
main.nfis ambiguous, fall back to files under the project’sinputs/metadatadirectory.
The output includes the manifest source, for example:
Manifest: inputs/metadata/lens.human.demo.manifest
Manifest source: workflow/main.nf
If multiple manifest-like files exist and main.nf does not identify one
unambiguously, project-status reports a warning.
Workflow-specific status
project-status is designed so workflow-specific checks can be added without
changing the user-facing command. Workflow-owned status files use the same
repository pattern as manifest and report UI templates.
RAFT looks for status configs in this order:
project-local workflow config checkouts, such as
<project>/lens-human-fastqs/lens.human.fastqs.status.jsonpackaged RAFT defaults under
src/raft/ui/static/workflows/<workflow>/
The most specific filename is preferred:
<workflow>.<species>.<input>.status.json
<workflow>.<species>.status.json
<workflow>.status.json
status.json
This means a workflow config repository such as
lens-human-fastqs can own lens.human.fastqs.status.json beside:
lens.human.fastqs.config
lens.human.fastqs.json
lens.human.fastqs.manifest-ui.json
lens.human.fastqs.report-ui.json
Status config files can define:
expected output paths
table/report glob patterns
fields to count
required columns
columns that should warn when blank in every row
For LENS projects, it reports:
number of final LENS report files
row count for the first final report
candidate counts by
antigen_sourcewarnings for missing configured required columns, such as
candidate_idand review columnswarnings for final-report columns that are blank in all rows for selected tumor-level fields
Interpreting warnings
Warnings are meant to identify conditions that deserve review. They are not always fatal errors.
Examples:
latest trace contains failed tasks
multiple manifest-like files were found but
main.nfdid not identify onefinal LENS report lacks
candidate_idreview columns are missing from the final LENS report
Use these warnings to decide whether to inspect logs, trace files, intermediate outputs, or workflow-specific documentation before trusting the final report.