Skip to contents

cellsight 0.0.1

First release of CellSight, derived from ShinyCell2 under GPL-3.0. See vignette("credits") for upstream attribution.

App generation

  • makeShinyFiles() dispatches on the object it is given, so a single call writes every modality present: gene expression for Seurat objects and .h5ad files, spatial data when obj@images is populated, ATAC coverage when a peaks assay and bigWigGroup are supplied, and a DEG page when precomputed.deg and clusters are given. ArchRProject inputs route to makeShinyFilesATACarchr().

  • makeShinyCodes() assembles server.R, ui.R and shinyFunc.R from jinjar templates shipped in inst/templates, replacing the previous hand-written wr* code writers. The generated files are run through styler so their layout is consistent regardless of how the template fragments were indented. See vignette("templates") for the template map.

  • Both single-dataset and multi-dataset apps are supported; multi-dataset apps group each dataset’s tabs under a per-dataset dropdown header.

  • .h5ad (AnnData / Scanpy) files are accepted wherever an object is, read via reticulate. makeH5fromAnndata() writes the expression matrix — reading h5ad.X by default, or a named entry of h5ad.layers.

Configuration

Tooling

  • tools/render-and-lint.R renders a fully-featured app from the templates and lints the generated R, and runs on every pull request.

  • Container assets under tools/docker/CellSight/, including the ingest_10x.R and seurat_inspector.R helper scripts. See vignette("docker").

  • A bundled VS Code syntax extension for *.R.jinja files under inst/extdata/vscode-rjinja/.