Configuration¶

Config files¶

File	Purpose
`.projio/config.yml`	Project-level configuration (required)
`~/.config/projio/config.yml`	User-level defaults (optional)

User config is merged with project config using deep merge. Project values take precedence.

The user config path respects XDG_CONFIG_HOME.

Config schema¶

project_name: my-project
project_kind: generic     # generic | tool | study
description: ""

indexio:
  config: .indexio/config.yaml
  persist_dir: .projio/index

biblio:
  enabled: true
  config: bib/config/biblio.yml

notio:
  enabled: true
  notes_dir: notes/

codio:
  enabled: true
  catalog_path: .codio/catalog.yml
  profiles_path: .codio/profiles.yml
  notes_dir: docs/reference/codelib/libraries/

pipeio:
  enabled: false                 # enable when project has pipelines
  registry_path: .pipeio/registry.yml
  pipelines_dir: code/pipelines  # where to scan for flows

runtime:
  python_bin: null              # absolute path to Python binary (for MCP config generation)

site:
  framework: mkdocs
  output_dir: site
  base_port: 8000
  host: "127.0.0.1"
  mkdocs:
    config_file: mkdocs.yml
    site_dir: site
  sphinx:
    source_dir: docs
    build_dir: docs/_build/html
  vite:
    app_dir: .
    build_dir: site
  chatbot:
    enabled: false
    backend_url: null
    host: "127.0.0.1"
    port: 9100
    title: "Docs Assistant"
    storage_key: "my-project_chat_v1"

automation:
  index:
    on_sync: false           # rebuild stale index sources during projio sync
    background: true         # run index rebuild in background (default: true)

helpers:
  sibling:
    github:
      sibling: github
      credential: github
      project_template: "{project_name}"
    gitlab:
      sibling: gitlab
      credential: gitlab
      project_template: "{project_name}"
    ria:
      sibling: origin
      alias_strategy: basename
      storage_url: ria+file:///path/to/ria-store
      shared: group
  docs:
    mkdocs:
      enabled: false

Runtime config¶

The runtime section configures how projio tools are invoked externally.

Key	Purpose
`runtime.python_bin`	Absolute path to the Python binary used by `projio mcp-config` to generate `.mcp.json`. Set this in your user config (`~/.config/projio/config.yml`) when the projio ecosystem lives in a specific conda env. Falls back to `sys.executable`.

Code & runner config¶

The code section configures how pipeio invokes snakemake and resolves project environments.

code:
  runner: pixi           # "conda" (default) or "pixi"; auto-detected from pixi.toml if omitted
  envs:
    default: cogpy       # environment name for snakemake / notebook execution
    docs: docs-env       # optional: environment for mkdocs
    datalad: labpy       # optional: environment for datalad
  conda_prefix: /opt/conda  # optional: absolute path to conda installation (for conda runner)

Key	Purpose
`code.runner`	Package manager used to wrap snakemake invocations: `"conda"` or `"pixi"`. When omitted, projio auto-detects `pixi.toml` in the project root.
`code.envs.default`	Environment name passed to the runner. For conda: `conda run -n <name> snakemake`. For pixi: `pixi run -e <name> snakemake` (omit or leave empty for pixi's default environment).
`code.envs.docs`	Optional environment for documentation builds.
`code.envs.datalad`	Optional environment for DataLad operations.
`code.conda_prefix`	Absolute path to the conda envs directory. Used with `code.envs.*` to resolve binary paths. Not needed for pixi projects.

Ecosystem sections¶

Each ecosystem package has a top-level section (indexio, biblio, notio, codio) with an enabled flag. Setting enabled: false prevents that subsystem's MCP tools from being registered.

Helper config¶

The helpers.sibling section provides defaults for projio sibling commands. Values here are used when CLI flags are not provided.

Credential names (e.g., github, gitlab-lrz) can safely live in .projio/config.yml. Actual secrets should remain in DataLad's credential store or environment variables.

Site chatbot config¶

site.chatbot controls whether projio site build / projio site serve inject the indexio chatbot widget into supported site frameworks.

site.framework lets projio treat mkdocs, sphinx, and vite as explicit first-class site modes
site.output_dir is the shared output location concept; by default this is site/ for MkDocs and Vite
site.mkdocs, site.sphinx, and site.vite hold framework-specific paths
enabled: true turns the integration on
for projio site serve, MkDocs sites can auto-start an indexio backend locally if backend_url is unset
for projio site build, set backend_url to the deployed chatbot server URL if you want the static site to include the widget
MkDocs chatbot injection is supported today; Sphinx and Vite do not currently get chatbot injection

Automation config¶

The automation section controls automatic maintenance tasks during projio sync.

`automation.index`¶

Key	Type	Default	Description
`on_sync`	bool	`false`	Run `indexio_build(update=True)` during `projio sync` to rebuild stale sources
`background`	bool	`true`	Run the index rebuild in a background process (non-blocking)

When on_sync is true, each projio sync invocation checks which indexio sources have changed files (by comparing mtime and size against the last build manifest) and only re-indexes those sources. Sources that are up to date are skipped.

The CLI flags --index and --no-index override this config for a single invocation.

To automatically rebuild the index after every git commit, run:

projio sync --install-hooks

This installs a lightweight .git/hooks/post-commit that runs projio sync --index in the background.

Managed ignores¶

projio init also maintains a dedicated projio block in .gitignore for generated site artifacts and runtime files, including site/, .projio/site/, .projio.mkdocs.yml, and framework-specific build directories where relevant.