From 1e6ecd824c69b1c1a87a02762d7f5c67d8376777 Mon Sep 17 00:00:00 2001 From: Artur-man Date: Mon, 18 May 2026 11:27:53 +0200 Subject: [PATCH 1/3] small corrections to zarr attr and stores --- vignettes/SpatialData.Rmd | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/vignettes/SpatialData.Rmd b/vignettes/SpatialData.Rmd index a6cbe18..0d72c38 100644 --- a/vignettes/SpatialData.Rmd +++ b/vignettes/SpatialData.Rmd @@ -44,7 +44,7 @@ The Python implementation and core specifications can be found at the The core data structure is the `SpatialData` class, which organizes data into 5 coordinated **layers: images, labels, points, shapes, and tables**. -Each layer is stored as a list of layer-specific objects that carry associated `SpatialDataAttr` (`@meta` slot), which encode `spatialdata`-specific *.zattr*. +Each layer is stored as a list of layer-specific objects that carry associated `SpatialDataAttr` (`@meta` slot), which encode `spatialdata`-specific zarr attributes (*.zattr* for Zarr v2, and `zarr.json` for Zarr v3) Together, these layers provide a unified representation of spatial omics data, combining raster, vector, and tabular data within a single coherent framework. @@ -62,7 +62,7 @@ These are DuckDB-backed by a `duckspatial_df`, enabling efficient lazy handling. **Tables** store functional annotations or information that has been aggregated across layers (e.g., gene $\times$ cell data). They are currently represented as in-memory `r BiocStyle::Biocpkg("SingleCellExperiment")` objects; delayed, -*.zarr*-backed handling of assay data is under active development. +Zarr backed handling of assay data is under active development. ```{r schematic, echo=FALSE, fig.wide=TRUE} knitr::include_graphics("schematic.png") @@ -70,7 +70,7 @@ knitr::include_graphics("schematic.png") # Handling -`SpatialData` are represented on-disk as `.zarr` stores. The package provides +`SpatialData` are represented on-disk as Zarr stores. The package provides the `readSpatialData()` function to ingest an entire store, although arguments to control which layers and elements to read or not to read are also available. @@ -81,7 +81,7 @@ For this demonstration, we use a toy dataset included in the package: library(SpatialData) library(SingleCellExperiment) -# path to 'spatialdata' .zarr store +# path to 'spatialdata' Zarr store zs <- file.path("extdata", "blobs.zarr") zs <- system.file(zs, package="SpatialData") @@ -155,7 +155,7 @@ Every spatial element (tables excluded) is composed of two key slots: or a `duckspatial_df` for shapes/points. - `meta`: a `SpatialDataAttrs` object containing the OME-NGFF metadata - retrieved from the *.zattrs* present in the original *.zarr* store. + retrieved from the zarr attributes present in the original Zarr store. We here demonstrate how to access these slots for a given element @@ -188,7 +188,7 @@ show(d) For single-cell and spatial omics datasets, functional annotations are commonly stored as [AnnData](https://anndata.readthedocs.io) objects in Python. In R, we use `r BiocStyle::Biocpkg("anndataR")` [@Deconinck2025-anndataR] to read these -`.zarr`-backed `AnnData` as `r BiocStyle::Biocpkg("SingleCellExperiment")`(s). +Zarr backed `AnnData` as `r BiocStyle::Biocpkg("SingleCellExperiment")`(s). A `table` can link to one or more `label` or `shape` (but not other layers), whereby internal metadata (`spatialdata_attrs`) are used to keep track of the @@ -210,7 +210,7 @@ instances(se) A key feature of the `SpatialData` framework is its handling of different coordinate systems. Each element can exist in multiple coordinate spaces -simultaneously, defined by transformations in its on-disk *.zattrs*. +simultaneously, defined by transformations in its on-disk Zarr attributes. The relationships between different elements and their respective coordinate spaces can be complex. `SpatialData` provides the `CTgraph()` and `CTplot()` From 27d172191f0082e78077bbe4eecf6fd970cf5bb4 Mon Sep 17 00:00:00 2001 From: Artur-man Date: Mon, 18 May 2026 11:29:42 +0200 Subject: [PATCH 2/3] zarr.json --- vignettes/SpatialData.Rmd | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/vignettes/SpatialData.Rmd b/vignettes/SpatialData.Rmd index 0d72c38..406c710 100644 --- a/vignettes/SpatialData.Rmd +++ b/vignettes/SpatialData.Rmd @@ -44,7 +44,7 @@ The Python implementation and core specifications can be found at the The core data structure is the `SpatialData` class, which organizes data into 5 coordinated **layers: images, labels, points, shapes, and tables**. -Each layer is stored as a list of layer-specific objects that carry associated `SpatialDataAttr` (`@meta` slot), which encode `spatialdata`-specific zarr attributes (*.zattr* for Zarr v2, and `zarr.json` for Zarr v3) +Each layer is stored as a list of layer-specific objects that carry associated `SpatialDataAttr` (`@meta` slot), which encode `spatialdata`-specific zarr attributes (*.zattr* for Zarr v2, and *zarr.json* for Zarr v3) Together, these layers provide a unified representation of spatial omics data, combining raster, vector, and tabular data within a single coherent framework. From dd09ef852114a775c757cf646ad250f534dfdd49 Mon Sep 17 00:00:00 2001 From: Artur-man Date: Mon, 18 May 2026 11:48:23 +0200 Subject: [PATCH 3/3] backed --- vignettes/SpatialData.Rmd | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/vignettes/SpatialData.Rmd b/vignettes/SpatialData.Rmd index 406c710..d9ad754 100644 --- a/vignettes/SpatialData.Rmd +++ b/vignettes/SpatialData.Rmd @@ -62,7 +62,7 @@ These are DuckDB-backed by a `duckspatial_df`, enabling efficient lazy handling. **Tables** store functional annotations or information that has been aggregated across layers (e.g., gene $\times$ cell data). They are currently represented as in-memory `r BiocStyle::Biocpkg("SingleCellExperiment")` objects; delayed, -Zarr backed handling of assay data is under active development. +Zarr-backed handling of assay data is under active development. ```{r schematic, echo=FALSE, fig.wide=TRUE} knitr::include_graphics("schematic.png") @@ -188,7 +188,7 @@ show(d) For single-cell and spatial omics datasets, functional annotations are commonly stored as [AnnData](https://anndata.readthedocs.io) objects in Python. In R, we use `r BiocStyle::Biocpkg("anndataR")` [@Deconinck2025-anndataR] to read these -Zarr backed `AnnData` as `r BiocStyle::Biocpkg("SingleCellExperiment")`(s). +Zarr-backed `AnnData` as `r BiocStyle::Biocpkg("SingleCellExperiment")`(s). A `table` can link to one or more `label` or `shape` (but not other layers), whereby internal metadata (`spatialdata_attrs`) are used to keep track of the