NEWS • nemo

v0.1.0 (2026-07-28)

Major refactor.

Unified schema format (pr48)

schema.yaml replaces the separate raw.yaml + tidy.yaml pair per tool. Each column now carries raw, tidy, type, description, and versions in one place under a flat tables: map. Child packages must migrate their per-tool configs to this format.

Config / Tool / Workflow API overhaul (pr64, pr65, pr66)

Config: get_raw_* / get_tidy_* methods renamed (e.g. get_raw_patterns() → get_patterns(), get_raw_schemas_all() → get_schemas_raw()); are_raw_schemas_valid() removed (validation now in initialize()); config, raw_schemas_all, tidy_schemas_all fields moved to private; pkg added as a public field; cloneable = FALSE
Tool: $files, $tbls, $files_tbl moved to private — use list_files() / get_tbls() instead; schema delegate fields (get_raw_schema, get_tidy_schema) removed — use tool$config$get_schema_raw() / tool$config$get_schema_tidy(); nemofy(diro = ...) → run(output_dir = ...); column group → prefix_suffix; post_process_files hook renamed to refine_files, now applied to the base prefix before the disambiguation passes so subclasses can fold a semantic distinction (e.g. germline/somatic) into the prefix and avoid a spurious _2/_3 suffix; cloneable = FALSE
Workflow: $tools, $files_tbl moved to private — use get_tools(); nemofy(diro = ...) → run(output_dir = ...); list_files(type = ...) — type arg removed; gains metapkg constructor argument; path existence validated at construction; filter_files() now validates names against known tool_parser values; cloneable = FALSE

CLI renames

--out_dir → --output_dir; --pfix_include → --prefix_include; output column input_pfix → input_prefix; --prefix_include is now opt-in (no longer added by default)

Other

nemo_metadata(): parameter input_dir => input_dirs; return value changed from a named list to a single-row tibble with list-columns
RPostgres dropped as a hard dependency - callers now supply dbdrv explicitly
txt ftypes renamed to tsv (pr70, pr71)

New assertion module: nemo_stop(), nemo_assert_scalar_chr(), nemo_assert_chr(), nemo_assert_not_null(), nemo_assert_out_fmt() replacing assertthat calls
Schema scaffold helpers: config_prep_raw_schema(), config_prep_raw(), config_prep_multi(), config_prep_write() — bootstrap a schema.yaml from example raw files
nemo_gha_mermaid(): generates a Mermaid flowchart of the full CI/CD pipeline (pr52)
nemo_uml(): generates a PlantUML SVG from R6 class names
nemo_schema_reactable() / nemo_schemavis_data(): interactive schema explorer
New file types: tsv-nohead, tsv-keyvalue, csv, csv-nohead-long
tidy: --output_id / --ulid flags for tagging outputs with a run identifier; --max flag on list subcommand
Metadata exported to metadata.parquet per run (pr62, iss61)
Tool and Workflow now both support write_metadata boolean
extra_ftypes() subclass hook for registering pkg-specific ftype parsers (pr70, pr71)
Add DVC support, mostly for child packages (pr69)
New Output Naming vignette, shared with child packages via inst/doc-templates/vignettes/_output_naming.qmd (pr73)

File discovery ~20× faster on large directories — benchmarked at 1.9s => 0.1 s on 43k files x 15 patterns

CI/CD: deploy.yaml refactored from a 136-line monolith to reusable workflows via tidywf/actions; now also triggers on dev branch pushes
@claude GHA restricted to repo owners, collaborators, and members
dependabot.yml added for automated GHA dependency updates
conda: aarch64 lock file added; dedicated bump.yaml env added
Panache format vignettes (pr72)

Add key-value parser (pr12, iss11)
Export config writer (pr13)
Add table description to schema getters, contribution docs, pkgdown fixes (pr19)
Add nemo_id + nemo_pfix in class writer (pr21, iss18)
- changed odir to diro arg in nemofy
list_files: str_remove offers more flexibility (pr24, iss23)
Add metadata export functionality (pr25, iss17)
- rename nemo_id + nemo_pfix to input_id + input_pfix, respectively
- add output_id
Bugfix: use I() for vroom literal data in empty_tbl (iss27, pr28)

Initial release of nemo (pr5)

Fixing issues iss2, iss3 and iss4