A command line for bioproject.
bioproject is a single pure-Go binary. It reads public bioproject data
over plain HTTPS, shapes it into clean records, and prints output that pipes
into the rest of your tools. No API key, nothing to run alongside it.
The same package is also a resource-URI driver,
so a host program like ant can address
bioproject as bioproject:// URIs.
go install github.com/tamnd/bioproject-cli/cmd/bioproject@latestOr grab a prebuilt binary from the releases, or run the container image:
docker run --rm ghcr.io/tamnd/bioproject:latest --helpbioproject page <path> # fetch one page as a record
bioproject page <path> -o json # as JSON, ready for jq
bioproject page <path> --template '{{.Body}}' # just the readable body text
bioproject links <path> # the pages it links to, one per line
bioproject --help # the whole command treeEvery command shares one output contract: -o table|json|jsonl|csv|tsv|url|raw,
--fields to pick columns, --template for a custom line, and -n to limit.
The default adapts to where output goes (a table on a terminal, JSONL in a
pipe), so the same command reads well by hand and parses cleanly downstream.
This is a fresh scaffold. It ships one example resource type, page, wired end
to end. Model the real bioproject records in bioproject/ and declare their
operations in bioproject/domain.go; each one becomes a command, an HTTP
route, and an MCP tool at once.
The same operations are available over HTTP and as an MCP tool set for agents, with no extra code:
bioproject serve --addr :7777 # GET /v1/page/<path> returns NDJSON
bioproject mcp # speak MCP over stdiobioproject registers a bioproject domain the way a program registers a
database driver with database/sql. A host enables it with one blank import:
import _ "github.com/tamnd/bioproject-cli/bioproject"Then ant (or any program that links the package)
dereferences bioproject:// URIs without knowing anything about bioproject:
ant get bioproject://page/<path> # fetch the record
ant cat bioproject://page/<path> # just the body text
ant ls bioproject://page/<path> # the pages it links to, each addressable
ant url bioproject://page/<path> # the live https URLcmd/bioproject/ thin main: hands cli.NewApp to kit.Run
cli/ assembles the kit App from the bioproject domain
bioproject/ the library: HTTP client, data models, and domain.go (the driver)
docs/ tago documentation site
make build # ./bin/bioproject
make test # go test ./...
make vet # go vet ./...Push a version tag and GitHub Actions runs GoReleaser, which builds the archives, Linux packages, the multi-arch GHCR image, checksums, SBOMs, and a cosign signature:
git tag v0.1.0
git push --tagsThe Homebrew and Scoop steps self-disable until their tokens exist, so the first release works with no extra secrets.
Apache-2.0. See LICENSE.