Installation

cd Package/capellini
pip install -e .

External tools must be on PATH before running pip install — the install will abort otherwise.

External dependencies

Tool	Install
`spacepharer`	`conda install -c bioconda spacepharer`
`mmseqs2`	`conda install -c bioconda mmseqs2`
`minced`	`conda install -c bioconda minced`
`prodigal`	`conda install -c bioconda prodigal`
`Rscript` +	`brew install r` then `R -e "BiocManager::install('dada2')"`
`dada2`
`micro`	`brew install micro`

Bypass the build-time check with CAPELLINI_SKIP_DEP_CHECK=1 pip install -e . (intended only for CI / containers where the tools become available later).

Reference FASTAs

Two large reference files are not shipped inside the source repo and must be fetched separately from the GitHub release:

capellini/data/references/progenome16S.fasta (~76 MB)
capellini/data/references/spacers/spacers_CompleteCollection.fasta (~74 MB)

After installing CAPELLINI, run the bundled downloader to place them in the right location inside the installed package:

capellini fetch-references
# or equivalently
capellini-fetch-references

The same action is available from the interactive UI: Main menu → Fetch reference FASTAs from GitHub release.

If you skip this step, CAPELLINI will fall back to downloading the full proGenomes3 source and rebuilding both files at runtime — slower, but useful if you intentionally want to regenerate them (regenerate_16S_reference: true / regenerate_spacers_collection: true).

To pull from a specific release tag, set CAPELLINI_REFERENCES_TAG:

CAPELLINI_REFERENCES_TAG=v0.1.0 capellini fetch-references