{ "producers": [ { "id": "base", "label": "Base processing", "children": [ { "id": "file", "label": "File metadata", "description": "Columns from the file-system scan: paths, sizes, extensions and timestamps." }, { "id": "image", "label": "Image metadata", "description": "Image properties the loaders extract into the shared raster-image schema.", "loaders": [ "bioio", "dicom", "geospatial", "h5", "nifti", "tifffile", "zarr" ] }, { "id": "derived", "label": "Derived metadata", "description": "Columns the pipeline computes from image data: ndim and num_pixels are derived from dim_order and shape; size_ columns are extracted from the actual array shape per dimension." }, { "id": "agg", "label": "Aggregation", "description": "Columns added when per-image rows are rolled up into the obs hierarchy (obs_level, per-dimension coordinates)." } ] }, { "id": "processors", "label": "Processors", "children": [ { "id": "processor:nodata-statistics", "label": "nodata-statistics", "description": "Counts pixels matching the declared nodata value." }, { "id": "processor:raster-basic", "label": "raster-basic", "description": "Computes basic per-image intensity statistics (min, max, mean, std), the finite-pixel count, and the NaN fraction, aggregated across chunks and dimensions." }, { "id": "processor:raster-histogram", "label": "raster-histogram", "description": "Computes a per-image intensity histogram, merging per-chunk histograms onto a shared value range." }, { "id": "processor:raster-quality", "label": "raster-quality", "description": "Computes no-reference image quality metrics (sharpness, noise, spectral characteristics, clipping) over the 2D spatial extent of each image." }, { "id": "processor:thumbnail", "label": "thumbnail", "description": "Generates a small downsampled RGBA thumbnail of each image, assembled from per-chunk spatial patches." } ] } ], "columns": [ { "name": "common_base", "dtype": "str", "description": "Name of the longest path prefix shared by all import bases.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "depth", "dtype": "int", "description": "Depth of the path below the imported base directory.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "file_extension", "dtype": "str", "description": "Lower-cased file extension without the leading dot.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "imported_path", "dtype": "str", "description": "Path of the import base (the -p path or base directory) under which this file was found, relative to the project base directory ('.' when no -p paths were specified).", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "imported_path_short", "dtype": "str", "description": "Shortened, disambiguated label for the import base (only when several -p paths were used).", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "modification_date", "dtype": "datetime", "description": "Last modification timestamp of the file.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "name", "dtype": "str", "description": "File or folder name, including extension.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "parent", "dtype": "str", "description": "Name of one directory level of the containing directory, counted from the top: parent0 is the first directory below the base directory, parent1 the one below that; null when the file is not that deep.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ], "regex": "^parent\\d+$" }, { "name": "path", "dtype": "str", "description": "Path of the file (or folder), relative to the project base directory.", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "size_bytes", "dtype": "int", "description": "Size on disk in bytes (aggregated for folders).", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "type", "dtype": "str", "description": "Row kind: 'file', 'folder', or 'sub_file' (a sub-image inside a container).", "producer": "file", "category": "file", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available.", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ] }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files).", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ] }, { "name": "crs_epsg", "dtype": "Optional", "description": "EPSG code of the CRS; None if not EPSG-registered.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "crs_str", "dtype": "Optional", "description": "Coordinate reference system as a WKT string.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes.", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "geospatial", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-geospatial", "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ] }, { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'.", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "geospatial", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-geospatial", "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ] }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32').", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "geospatial", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-geospatial", "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ] }, { "name": "footprint", "dtype": "Optional", "description": "Bounding-box footprint as a GeoJSON polygon in EPSG:4326.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "h5_attributes", "dtype": "dict", "description": "Raw HDF5 attributes of the dataset, merged with file root attributes.", "producer": "image", "category": "image", "creators": [ "h5" ], "packages": [ "pixel-patrol-loader-bio" ] }, { "name": "h5_dataset_path", "dtype": "string", "description": "Path of the dataset within the HDF5 file.", "producer": "image", "category": "image", "creators": [ "h5" ], "packages": [ "pixel-patrol-loader-bio" ] }, { "name": "latitude", "dtype": "Optional", "description": "Centroid latitude in WGS-84 degrees.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "longitude", "dtype": "Optional", "description": "Centroid longitude in WGS-84 degrees.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "nifti_intent", "dtype": "string", "description": "NIfTI intent code describing the data type (e.g. 'NIFTI_INTENT_NONE', 'NIFTI_INTENT_LABEL').", "producer": "image", "category": "image", "creators": [ "nifti" ], "packages": [ "pixel-patrol-loader-medical" ] }, { "name": "nodata_value", "dtype": "Union", "description": "Declared nodata/fill value from the file metadata.", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "pixel_size_", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit.", "producer": "image", "category": "image", "creators": [ "bioio", "dicom", "h5", "nifti", "tifffile", "zarr" ], "packages": [ "pixel-patrol-loader-bio", "pixel-patrol-loader-medical" ], "regex": "^pixel_size_[A-Za-z]$" }, { "name": "shape", "dtype": "list", "description": "Raster dimensions as [bands, height, width].", "producer": "image", "category": "image", "creators": [ "geospatial" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "zarr_attributes", "dtype": "dict", "description": "Raw key-value attributes stored in the Zarr/OME-Zarr group metadata.", "producer": "image", "category": "image", "creators": [ "zarr" ], "packages": [ "pixel-patrol-loader-bio" ] }, { "name": "ndim", "dtype": "int", "description": "Number of image dimensions, derived from dim_order.", "producer": "derived", "category": "derived", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "num_pixels", "dtype": "int", "description": "Number of pixels in this row's spatial extent (full image at obs_level=0, slice at higher levels).", "producer": "derived", "category": "derived", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "size_", "dtype": "int", "description": "Extent (number of elements) of this row along the given axis.", "producer": "derived", "category": "derived", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ], "regex": "^size_[A-Za-z]$" }, { "name": "dim_", "dtype": "int", "description": "Coordinate of this row along the given axis (e.g. dim_z = Z index); null when the row spans the whole axis.", "producer": "agg", "category": "agg", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ], "regex": "^dim_[A-Za-z]+$" }, { "name": "obs_level", "dtype": "int", "description": "Aggregation level of the row: 0 is the whole-image summary; higher levels are per-dimension breakdowns.", "producer": "agg", "category": "agg", "creators": [ "base processing" ], "packages": [ "pixel-patrol-base" ] }, { "name": "bright_clipping_fraction", "dtype": "float32", "description": "Fraction of pixels at the dtype's maximum representable value (integer types only; NaN for float). Detects sensor saturation at the dtype ceiling.", "producer": "processor:raster-quality", "category": "metric", "creators": [ "raster-quality" ], "packages": [ "pixel-patrol-base" ] }, { "name": "dark_clipping_fraction", "dtype": "float32", "description": "Fraction of pixels at the dtype's minimum representable value (0 for unsigned integers; NaN for float). Detects underexposure, background, and nodata at the dark end.", "producer": "processor:raster-quality", "category": "metric", "creators": [ "raster-quality" ], "packages": [ "pixel-patrol-base" ] }, { "name": "finite_pixel_count", "dtype": "uint64", "description": "Number of finite (non-NaN/Inf) pixels contributing to the statistics.", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "histogram_counts", "dtype": "array", "description": "Per-bin pixel counts over the histogram value range (fixed number of bins; integer data uses one-level-wide bins).", "producer": "processor:raster-histogram", "category": "metric", "creators": [ "raster-histogram" ], "packages": [ "pixel-patrol-base" ] }, { "name": "histogram_max", "dtype": "float32", "description": "Upper bound of the histogram's value range.", "producer": "processor:raster-histogram", "category": "metric", "creators": [ "raster-histogram" ], "packages": [ "pixel-patrol-base" ] }, { "name": "histogram_min", "dtype": "float32", "description": "Lower bound of the histogram's value range.", "producer": "processor:raster-histogram", "category": "metric", "creators": [ "raster-histogram" ], "packages": [ "pixel-patrol-base" ] }, { "name": "histogram_nan_count", "dtype": "uint64", "description": "Number of NaN pixels excluded from the histogram.", "producer": "processor:raster-histogram", "category": "metric", "creators": [ "raster-histogram" ], "packages": [ "pixel-patrol-base" ] }, { "name": "laplacian_variance", "dtype": "float32", "description": "High-frequency content measure. Blur and heavy compression push values down; sharpness, noise, and fine texture push values up. A smooth, featureless scene will score low regardless of focus quality. Clipped pixels inflate the score.", "producer": "processor:raster-quality", "category": "metric", "creators": [ "raster-quality" ], "packages": [ "pixel-patrol-base" ] }, { "name": "max_intensity", "dtype": "float32", "description": "Maximum pixel intensity over the covered extent (ignoring NaNs).", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "mean_intensity", "dtype": "float32", "description": "Pixel-count-weighted mean intensity over the covered extent.", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "min_intensity", "dtype": "float32", "description": "Minimum pixel intensity over the covered extent (ignoring NaNs).", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "nan_fraction", "dtype": "float32", "description": "Fraction of pixels excluded as NaN.", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "nodata_count", "dtype": "int", "description": "Number of pixels equal to the declared nodata value.", "producer": "processor:nodata-statistics", "category": "metric", "creators": [ "nodata-statistics" ], "packages": [ "pixel-patrol-geospatial" ] }, { "name": "spectral_slope", "dtype": "float32", "description": "Log-log slope of the radially averaged power spectrum, fit over the mid-frequency band (5\u201340% of Nyquist). Typical range: \u22122 to \u22124. Closer to 0 indicates noise or uniform content. More negative values indicate blur or stronger low-frequency dominance.", "producer": "processor:raster-quality", "category": "metric", "creators": [ "raster-quality" ], "packages": [ "pixel-patrol-base" ] }, { "name": "std_intensity", "dtype": "float32", "description": "Pooled standard deviation of intensity over the covered extent.", "producer": "processor:raster-basic", "category": "metric", "creators": [ "raster-basic" ], "packages": [ "pixel-patrol-base" ] }, { "name": "thumbnail", "dtype": "bytes", "description": "Raw RGBA bytes of the assembled thumbnail sprite (fixed sprite size).", "producer": "processor:thumbnail", "category": "metric", "creators": [ "thumbnail" ], "packages": [ "pixel-patrol-base" ] }, { "name": "thumbnail_dtype", "dtype": "string", "description": "Original pixel dtype of the source image the thumbnail was built from.", "producer": "processor:thumbnail", "category": "metric", "creators": [ "thumbnail" ], "packages": [ "pixel-patrol-base" ] }, { "name": "thumbnail_norm_max", "dtype": "float", "description": "Upper intensity bound used to normalize the thumbnail.", "producer": "processor:thumbnail", "category": "metric", "creators": [ "thumbnail" ], "packages": [ "pixel-patrol-base" ] }, { "name": "thumbnail_norm_min", "dtype": "float", "description": "Lower intensity bound used to normalize the thumbnail.", "producer": "processor:thumbnail", "category": "metric", "creators": [ "thumbnail" ], "packages": [ "pixel-patrol-base" ] } ], "loader_schemas": [ { "id": "raster-image", "label": "Raster Image Loaders", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'.", "required": true }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32').", "required": true }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes.", "required": false }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available.", "required": false }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files).", "required": false } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "loaders": [ "bioio", "dicom", "geospatial", "h5", "nifti", "tifffile", "zarr" ] } ], "loaders": [ { "name": "bioio", "description": "Opens a wide range of microscopy and standard image formats via BioIO, extracting pixel data and image metadata (dimensions, channels, pixel sizes).", "package": "pixel-patrol-loader-bio", "extensions": [ "bmp", "czi", "jpeg", "jpg", "lif", "nd2", "ome.tif", "ome.tiff", "ome.zarr", "png", "tif", "tiff", "zarr" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [], "extra_patterns": [] }, { "name": "dicom", "description": "Loads DICOM images (.dcm, .dicom), assembling multi-slice series from folders into 3D volumes.", "package": "pixel-patrol-loader-medical", "extensions": [ "dcm", "dicom" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [], "extra_patterns": [] }, { "name": "geospatial", "description": "Loads geospatial raster files (GeoTIFF, NetCDF) via rasterio, extracting CRS, spatial extent, and band metadata.", "package": "pixel-patrol-geospatial", "extensions": [ "nc", "tif", "tiff", "txt", "xyz", "xyz.gz" ], "schema": "raster-image", "columns": [ { "name": "crs_str", "dtype": "Optional", "description": "Coordinate reference system as a WKT string." }, { "name": "crs_epsg", "dtype": "Optional", "description": "EPSG code of the CRS; None if not EPSG-registered." }, { "name": "latitude", "dtype": "Optional", "description": "Centroid latitude in WGS-84 degrees." }, { "name": "longitude", "dtype": "Optional", "description": "Centroid longitude in WGS-84 degrees." }, { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "num_pixels", "dtype": "int", "description": "Number of pixels in this row's spatial extent (full image at obs_level=0, slice at higher levels)." }, { "name": "shape", "dtype": "list", "description": "Raster dimensions as [bands, height, width]." }, { "name": "ndim", "dtype": "int", "description": "Number of image dimensions, derived from dim_order." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "footprint", "dtype": "Optional", "description": "Bounding-box footprint as a GeoJSON polygon in EPSG:4326." }, { "name": "nodata_value", "dtype": "Union", "description": "Declared nodata/fill value from the file metadata." } ], "patterns": [], "extra_columns": [ { "name": "crs_str", "dtype": "Optional", "description": "Coordinate reference system as a WKT string." }, { "name": "crs_epsg", "dtype": "Optional", "description": "EPSG code of the CRS; None if not EPSG-registered." }, { "name": "latitude", "dtype": "Optional", "description": "Centroid latitude in WGS-84 degrees." }, { "name": "longitude", "dtype": "Optional", "description": "Centroid longitude in WGS-84 degrees." }, { "name": "shape", "dtype": "list", "description": "Raster dimensions as [bands, height, width]." }, { "name": "footprint", "dtype": "Optional", "description": "Bounding-box footprint as a GeoJSON polygon in EPSG:4326." }, { "name": "nodata_value", "dtype": "Union", "description": "Declared nodata/fill value from the file metadata." } ], "extra_patterns": [] }, { "name": "h5", "description": "Loads HDF5, BigDataViewer/BDV, and modern Imaris IMS files lazily. Format-aware readers exclude pyramid and helper datasets; generic HDF5 files treat each image-like dataset as a sub-image.", "package": "pixel-patrol-loader-bio", "extensions": [ "h5", "hdf5", "ims" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." }, { "name": "h5_dataset_path", "dtype": "string", "description": "Path of the dataset within the HDF5 file." }, { "name": "h5_attributes", "dtype": "dict", "description": "Raw HDF5 attributes of the dataset, merged with file root attributes." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [ { "name": "h5_dataset_path", "dtype": "string", "description": "Path of the dataset within the HDF5 file." }, { "name": "h5_attributes", "dtype": "dict", "description": "Raw HDF5 attributes of the dataset, merged with file root attributes." } ], "extra_patterns": [] }, { "name": "nifti", "description": "Loads NIfTI images (.nii, .nii.gz), reading voxel data and header metadata.", "package": "pixel-patrol-loader-medical", "extensions": [ "nii", "nii.gz" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." }, { "name": "nifti_intent", "dtype": "string", "description": "NIfTI intent code describing the data type (e.g. 'NIFTI_INTENT_NONE', 'NIFTI_INTENT_LABEL')." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [ { "name": "nifti_intent", "dtype": "string", "description": "NIfTI intent code describing the data type (e.g. 'NIFTI_INTENT_NONE', 'NIFTI_INTENT_LABEL')." } ], "extra_patterns": [] }, { "name": "tifffile", "description": "Loads TIFF and OME-TIFF images via tifffile, reading pixel data and OME image metadata.", "package": "pixel-patrol-loader-bio", "extensions": [ "ome.tif", "ome.tiff", "tif", "tiff" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [], "extra_patterns": [] }, { "name": "zarr", "description": "Loads Zarr / OME-Zarr stores, reading pixel data, image metadata, and the store's raw Zarr attributes.", "package": "pixel-patrol-loader-bio", "extensions": [ "ome.zarr", "zarr" ], "schema": "raster-image", "columns": [ { "name": "dim_order", "dtype": "string", "description": "Axis order of the image, e.g. 'TCZYX'." }, { "name": "dtype", "dtype": "string", "description": "Pixel data type of the source image (e.g. 'uint8', 'float32')." }, { "name": "dim_names", "dtype": "list", "description": "Human-readable names of the image axes." }, { "name": "channel_names", "dtype": "list", "description": "Names of the image channels, if available." }, { "name": "child_id", "dtype": "string", "description": "Identifier of a sub-image within a container file (null for single-image files)." }, { "name": "zarr_attributes", "dtype": "dict", "description": "Raw key-value attributes stored in the Zarr/OME-Zarr group metadata." } ], "patterns": [ { "pattern": "^pixel_size_[A-Za-z]$", "dtype": "float", "description": "Physical pixel size along the given axis, in the image's spatial unit." } ], "extra_columns": [ { "name": "zarr_attributes", "dtype": "dict", "description": "Raw key-value attributes stored in the Zarr/OME-Zarr group metadata." } ], "extra_patterns": [] } ], "processors": [ { "name": "nodata-statistics", "description": "Counts pixels matching the declared nodata value.", "package": "pixel-patrol-geospatial", "input": { "axes": [ "X", "Y" ], "kinds": [ "intensity" ], "capabilities": [] }, "output": "features", "columns": [ { "name": "nodata_count", "dtype": "int", "description": "Number of pixels equal to the declared nodata value." } ] }, { "name": "raster-basic", "description": "Computes basic per-image intensity statistics (min, max, mean, std), the finite-pixel count, and the NaN fraction, aggregated across chunks and dimensions.", "package": "pixel-patrol-base", "input": { "axes": [], "kinds": [ "intensity" ], "capabilities": [] }, "output": "features", "columns": [ { "name": "min_intensity", "dtype": "float32", "description": "Minimum pixel intensity over the covered extent (ignoring NaNs)." }, { "name": "max_intensity", "dtype": "float32", "description": "Maximum pixel intensity over the covered extent (ignoring NaNs)." }, { "name": "mean_intensity", "dtype": "float32", "description": "Pixel-count-weighted mean intensity over the covered extent." }, { "name": "std_intensity", "dtype": "float32", "description": "Pooled standard deviation of intensity over the covered extent." }, { "name": "finite_pixel_count", "dtype": "uint64", "description": "Number of finite (non-NaN/Inf) pixels contributing to the statistics." }, { "name": "nan_fraction", "dtype": "float32", "description": "Fraction of pixels excluded as NaN." } ] }, { "name": "raster-histogram", "description": "Computes a per-image intensity histogram, merging per-chunk histograms onto a shared value range.", "package": "pixel-patrol-base", "input": { "axes": [], "kinds": [ "intensity" ], "capabilities": [] }, "output": "features", "columns": [ { "name": "histogram_min", "dtype": "float32", "description": "Lower bound of the histogram's value range." }, { "name": "histogram_max", "dtype": "float32", "description": "Upper bound of the histogram's value range." }, { "name": "histogram_nan_count", "dtype": "uint64", "description": "Number of NaN pixels excluded from the histogram." }, { "name": "histogram_counts", "dtype": "array", "description": "Per-bin pixel counts over the histogram value range (fixed number of bins; integer data uses one-level-wide bins)." } ] }, { "name": "raster-quality", "description": "Computes no-reference image quality metrics (sharpness, noise, spectral characteristics, clipping) over the 2D spatial extent of each image.", "package": "pixel-patrol-base", "input": { "axes": [ "X", "Y" ], "kinds": [ "intensity" ], "capabilities": [ "spatial-2d" ] }, "output": "features", "columns": [ { "name": "laplacian_variance", "dtype": "float32", "description": "High-frequency content measure. Blur and heavy compression push values down; sharpness, noise, and fine texture push values up. A smooth, featureless scene will score low regardless of focus quality. Clipped pixels inflate the score." }, { "name": "spectral_slope", "dtype": "float32", "description": "Log-log slope of the radially averaged power spectrum, fit over the mid-frequency band (5\u201340% of Nyquist). Typical range: \u22122 to \u22124. Closer to 0 indicates noise or uniform content. More negative values indicate blur or stronger low-frequency dominance." }, { "name": "dark_clipping_fraction", "dtype": "float32", "description": "Fraction of pixels at the dtype's minimum representable value (0 for unsigned integers; NaN for float). Detects underexposure, background, and nodata at the dark end." }, { "name": "bright_clipping_fraction", "dtype": "float32", "description": "Fraction of pixels at the dtype's maximum representable value (integer types only; NaN for float). Detects sensor saturation at the dtype ceiling." } ] }, { "name": "thumbnail", "description": "Generates a small downsampled RGBA thumbnail of each image, assembled from per-chunk spatial patches.", "package": "pixel-patrol-base", "input": { "axes": [ "X", "Y" ], "kinds": [ "intensity" ], "capabilities": [] }, "output": "features", "columns": [ { "name": "thumbnail", "dtype": "bytes", "description": "Raw RGBA bytes of the assembled thumbnail sprite (fixed sprite size)." }, { "name": "thumbnail_norm_min", "dtype": "float", "description": "Lower intensity bound used to normalize the thumbnail." }, { "name": "thumbnail_norm_max", "dtype": "float", "description": "Upper intensity bound used to normalize the thumbnail." }, { "name": "thumbnail_dtype", "dtype": "string", "description": "Original pixel dtype of the source image the thumbnail was built from." } ] } ], "widgets": [ { "id": "custom-plot", "group": "Explore", "label": "Custom Plot", "shortLabel": "Explore", "info": "Build your own plot from the columns in your current data.\n\nEach plot has its own **Slice by** toggles and **per image** / **per slice** badge, controlling whether each datapoint is a whole-image aggregate (**per image**) or one (image \u00d7 dimension) combination (**per slice**).\n\n- **Two numerics** \u2192 scatter\n- **Categorical \u00d7 numeric** \u2192 violin or bar (mean \u00b1 sd)\n- **Any column \u00d7 (count)** \u2192 count bar\n- **Two categoricals** \u2192 count heatmap\n\nIf a column has **no variance**, it is shown in a table instead of a plot. For scatter, this checks the Y column only (regardless of X).\n\n**Color by:**\n- **(none)** \u2192 a single, ungrouped trace (no legend), in a color you pick with the swatch next to \"Color by\".\n- **(global group)** (default) \u2192 colors/splits by whatever column the app-wide grouping is currently set to.\n- **any other column** \u2192 colors/splits by that column instead, using the chosen palette.\n\n**Scatter plots** additionally support coloring by a *numeric* column: each point is colored individually on a continuous colormap (with a colorbar) instead of being split into groups \u2014 pick the colormap from the palette dropdown.\n\nColor-by is automatically switched to **(none)** when:\n- the chosen X/Y column is already the active global-grouping column (avoids \"double grouping\", e.g. the same category split into both the x-axis and the legend), or\n- the chosen color-by column has more than 12 unique values and isn't used as a continuous colormap (too many to color distinctly) \u2014 a warning is shown when this happens.\n\n**Heatmaps** (two categoricals) never use color-by/hue grouping \u2014 instead, pick a single base color with the swatch; cell counts are shown on a white-to-color scale (white = 0, your color = the maximum). Check **Invert** to reverse this (your color = 0, white = the maximum).\n\n**Missing values**: for violin/bar/heatmap, rows where a *categorical* axis is null are grouped into their own \"(missing)\" category (shown last). For scatter, rows where X or Y is null can't be placed on a numeric axis and are excluded \u2014 a warning shows how many.\n\n**\u2193 Export plugin** downloads the plot as a standalone `plugin_*.js` file - if you have access to the source code, drop it into a viewer folder of a PixelPatrol package (e.g. `pixel-patrol-base/src/pixel_patrol_base/viewer/`) to keep this exact plot as a permanent widget.\n\n**\uff0b Add plot** adds an independent plot below.", "inputs": [ "any metric column" ] }, { "id": "dim-size", "group": "Metadata", "scope": "image", "label": "Dimensionality", "info": "Shows how **image dimension sizes** (e.g. X, Y, Z, T, \u2026) vary across the dataset.\n\nIncludes an X/Y size scatter plot and per-dimension strip plots.\n\nIf a dimension has **no variance**, it is summarized in a table instead of a plot.\n\nUse this to identify **unexpected dimension sizes** and **mismatched shapes** within/between groupings.", "inputs": [ "size_*" ] }, { "id": "file-stats", "group": "File Stats", "scope": "file", "label": "File Statistics", "shortLabel": "File Metadata", "info": "High-level **file statistics** for the dataset.\n\nIf a property has **no variance** (e.g. all files share the same extension), it is summarized in the table instead of a chart.", "inputs": [ "modification_date" ], "required_inputs": [ "file_extension", "size_bytes" ] }, { "id": "histogram", "group": "Dataset Stats", "scope": "image", "label": "Pixel Value Histograms", "shortLabel": "Histograms", "info": "Histograms are computed per image (256 bins) and averaged per group.\nEach image is area-normalized before averaging, so image size does not affect the result.\nA **NaN** bucket left of the real data shows the fraction of pixels excluded as NaN, if any.\n\nIf the dataset mixes unsigned integer, signed integer, and/or floating-point images, each kind gets its own plot.\n\n**Display mode** (integer panels only)\n- **Actual values** \u2013 x-axis shows real pixel values.\n- **Dtype range** \u2013 x-axis normalized to the representable range of the pixel type (unsigned integers: 0 to 1, signed integers: -1 to 1). Useful for comparing images across bit depths.\n\nFloat images always show actual values.\n\n**Spread** \u2013 shaded +/-1 std band around the group mean.\nA wider band means more variation between images in that group.\n\n**Show individual images** \u2013 visible when one group is selected with 50 or fewer images.\nReplaces the group mean with one line per image.", "inputs": [ "histogram_min", "histogram_max", "histogram_nan_count", "dtype", "name" ], "required_inputs": [ "histogram_counts" ] }, { "id": "image-table", "group": "Summary", "scope": "image", "label": "Image Table", "shortLabel": "Table", "info": "Full-image statistics \u2014 one row per image, no per-slice or per-channel rows. Click a column header to sort. Search (Enter): substring match across all columns for small datasets, path and child_id only for large ones (\u226510k images).", "inputs": [ "*", "!thumbnail", "!histogram_counts", "!channel_names", "!obs_level", "!histogram_min", "!histogram_max", "!histogram_nan_count", "!file_row_number" ] }, { "id": "map-points", "group": "Geospatial extension", "scope": "image", "label": "Locations" }, { "id": "metadata", "group": "Metadata", "scope": "image", "label": "Metadata", "shortLabel": "Image Metadata", "info": "Shows the distribution of **pixel data types** and **dimension ordering** across groupings.\n\nIf a property has **no variance** (e.g. every image has the same pixel dtype), it is summarized in a table instead of a chart.", "inputs": [ "dim_order", "dtype", "ndim", "pixel_size_X", "pixel_size_Y", "pixel_size_Z" ] }, { "id": "mosaic", "group": "Visualization", "scope": "image", "label": "Image Thumbnails", "shortLabel": "Thumbnails", "info": "Displays a **thumbnail mosaic**, one thumbnail per image.\n\n- Thumbnails are generated from the central slice in all non-XY, non-RGB dimensions.\n- Sorting by a numeric column can reveal visual trends.\n- Border colors indicate the group of each image.\n- **Hover** over an image to see its filename.", "inputs": [ "thumbnail_norm_min", "thumbnail_norm_max", "thumbnail_dtype", "any metric column", "name" ], "required_inputs": [ "thumbnail" ] }, { "id": "nodata-value-count", "group": "no-data", "scope": "image", "label": "NoData", "info": "NoData value frequency, nodata-pixel percentage, and non-finite pixels (float images)." }, { "id": "stats-across-dims-basic", "group": "Dataset Stats", "scope": "slice", "label": "Basic Statistics Across Dimensions", "shortLabel": "Intensity across Slices", "info": "Shows how image statistics (e.g., mean, std, min, max) change **across different dimension slices**.\n\nUseful for identifying drift, artifacts, or unexpected variation.\n\nThe dashed line shows the **% of images that still have a slice at this position**, per condition - useful for spotting datasets where images have different numbers of slices.", "inputs": [ "mean_intensity", "std_intensity", "min_intensity", "max_intensity" ] }, { "id": "stats-across-dims-quality", "group": "Dataset Stats", "scope": "slice", "label": "Quality Metrics Across Dimensions", "shortLabel": "Quality across Slices", "info": "Shows how **image quality metrics** change across different dimension slices.\n\nUseful to detect e.g. drift in focus or noise over time (T) and depth (Z) as well as channel (C) specific artifacts.\n\nThe dashed line shows the **% of images that still have a slice at this position**, per condition.", "inputs": [ "laplacian_variance", "spectral_slope", "dark_clipping_fraction", "bright_clipping_fraction", "nan_fraction" ] }, { "id": "summary", "group": "Summary", "label": "File Data Summary", "info": "High-level overview of the dataset: total number of files and images, total size, file extensions, and (when grouped) a per-group breakdown.", "inputs": [ "path" ], "required_inputs": [ "size_bytes", "file_extension" ] }, { "id": "sunburst", "group": "Summary", "scope": "file", "label": "File Structure Sunburst", "shortLabel": "Folder Structure", "info": "Sunburst view of the **file and folder hierarchy**.\n\nClick a slice to zoom in; click the center to zoom out.", "inputs": [], "required_inputs": [ "path", "size_bytes" ] }, { "id": "violin-basic", "group": "Dataset Stats", "scope": "image", "label": "Pixel Value Statistics", "shortLabel": "Intensity", "info": "Shows **per-image intensity statistics** across groups.\n\nPlots with 5,000 or fewer datapoints show the actual distribution shape (a violin); larger plots switch to a summary box plot (quartiles, min/max, mean).\n\nUse the **Slice by** toggles to control whether each datapoint is a whole-image aggregate (**per image**) or one (image \u00d7 dimension) combination (**per slice**) - shown by the badge in the card header. Pinning a dimension in the sidebar does the same, for that one slice. More toggles mean more datapoints, so check the sample size in the plot subtitle before trusting significance.\n\n**Statistical Comparisons**\n\nIf selected (side menu), pairwise group comparisons use the Mann\u2013Whitney U test (a non-parametric test that makes no assumptions about the data distribution) with Bonferroni correction for multiple comparisons.\n\nSignificance levels: `ns` not significant (p \u2265 0.05), `*` p < 0.05, `**` p < 0.01, `***` p < 0.001.", "inputs": [ "mean_intensity", "std_intensity", "min_intensity", "max_intensity" ] }, { "id": "violin-quality", "group": "Dataset Stats", "scope": "image", "label": "Image Quality Metrics", "shortLabel": "Image Quality", "info": "Visualizes **image quality metrics** across groups.\n\nThese metrics are **whole-image aggregates** \u2014 they detect global quality issues such as blur, noise, or exposure problems that affect the image broadly. They are not sensitive to localized defects or anomalies that cover only a small fraction of pixels.\n\nPlots with 5,000 or fewer datapoints show the actual distribution shape (a violin); larger plots switch to a summary box plot (quartiles, min/max, mean).\n\nUse the **Slice by** toggles to control whether each datapoint is a whole-image aggregate (**per image**) or one (image \u00d7 dimension) combination (**per slice**) - shown by the badge in the card header. Pinning a dimension in the sidebar does the same, for that one slice. More toggles mean more datapoints, so check the sample size in the plot subtitle before trusting significance.\n\n**Statistical Comparisons**\n\nIf selected (side menu), pairwise group comparisons use the Mann\u2013Whitney U test (a non-parametric test that makes no assumptions about the data distribution) with Bonferroni correction for multiple comparisons.\n\nSignificance levels: `ns` not significant (p \u2265 0.05), `*` p < 0.05, `**` p < 0.01, `***` p < 0.001.", "inputs": [ "laplacian_variance", "spectral_slope", "dark_clipping_fraction", "bright_clipping_fraction", "nan_fraction" ] } ], "connections": [ { "source": "loader:bioio", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:bioio", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:bioio", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:bioio", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:bioio", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:bioio", "target": "derived", "via": "image data" }, { "source": "loader:dicom", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:dicom", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:dicom", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:dicom", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:dicom", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:dicom", "target": "derived", "via": "image data" }, { "source": "loader:geospatial", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:geospatial", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:geospatial", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:geospatial", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:geospatial", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:geospatial", "target": "derived", "via": "image data" }, { "source": "loader:h5", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:h5", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:h5", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:h5", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:h5", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:h5", "target": "derived", "via": "image data" }, { "source": "loader:nifti", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:nifti", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:nifti", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:nifti", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:nifti", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:nifti", "target": "derived", "via": "image data" }, { "source": "loader:tifffile", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:tifffile", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:tifffile", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:tifffile", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:tifffile", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:tifffile", "target": "derived", "via": "image data" }, { "source": "loader:zarr", "target": "processor:nodata-statistics", "via": "raster image" }, { "source": "loader:zarr", "target": "processor:raster-basic", "via": "raster image" }, { "source": "loader:zarr", "target": "processor:raster-histogram", "via": "raster image" }, { "source": "loader:zarr", "target": "processor:raster-quality", "via": "raster image" }, { "source": "loader:zarr", "target": "processor:thumbnail", "via": "raster image" }, { "source": "loader:zarr", "target": "derived", "via": "image data" }, { "source": "processor:nodata-statistics", "target": "widget:custom-plot", "via": "metrics" }, { "source": "processor:raster-basic", "target": "widget:custom-plot", "via": "metrics" }, { "source": "processor:raster-quality", "target": "widget:custom-plot", "via": "metrics" }, { "source": "derived", "target": "widget:dim-size", "via": "size_*" }, { "source": "file", "target": "widget:file-stats", "via": "modification_date" }, { "source": "processor:raster-histogram", "target": "widget:histogram", "via": "histogram_counts" }, { "source": "loader:h5", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:tifffile", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:dicom", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:bioio", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:zarr", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:nifti", "target": "widget:histogram", "via": "dtype" }, { "source": "loader:geospatial", "target": "widget:histogram", "via": "dtype" }, { "source": "file", "target": "widget:histogram", "via": "name" }, { "source": "processor:raster-basic", "target": "widget:image-table", "via": "min_intensity" }, { "source": "loader:geospatial", "target": "widget:image-table", "via": "crs_str" }, { "source": "file", "target": "widget:image-table", "via": "size_bytes" }, { "source": "loader:nifti", "target": "widget:image-table", "via": "nifti_intent" }, { "source": "loader:h5", "target": "widget:image-table", "via": "dim_order" }, { "source": "loader:tifffile", "target": "widget:image-table", "via": "dim_order" }, { "source": "loader:dicom", "target": "widget:image-table", "via": "dim_order" }, { "source": "loader:bioio", "target": "widget:image-table", "via": "dim_order" }, { "source": "loader:zarr", "target": "widget:image-table", "via": "dim_order" }, { "source": "derived", "target": "widget:image-table", "via": "num_pixels" }, { "source": "processor:raster-quality", "target": "widget:image-table", "via": "bright_clipping_fraction" }, { "source": "agg", "target": "widget:image-table", "via": "dim_" }, { "source": "processor:nodata-statistics", "target": "widget:image-table", "via": "nodata_count" }, { "source": "processor:thumbnail", "target": "widget:image-table", "via": "thumbnail_norm_max" }, { "source": "loader:h5", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:tifffile", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:dicom", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:bioio", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:zarr", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:nifti", "target": "widget:metadata", "via": "dim_order" }, { "source": "loader:geospatial", "target": "widget:metadata", "via": "dim_order" }, { "source": "derived", "target": "widget:metadata", "via": "ndim" }, { "source": "processor:thumbnail", "target": "widget:mosaic", "via": "thumbnail_norm_min" }, { "source": "file", "target": "widget:mosaic", "via": "name" }, { "source": "processor:nodata-statistics", "target": "widget:mosaic", "via": "metrics" }, { "source": "processor:raster-basic", "target": "widget:mosaic", "via": "metrics" }, { "source": "processor:raster-quality", "target": "widget:mosaic", "via": "metrics" }, { "source": "processor:raster-basic", "target": "widget:stats-across-dims-basic", "via": "min_intensity" }, { "source": "processor:raster-basic", "target": "widget:stats-across-dims-quality", "via": "nan_fraction" }, { "source": "processor:raster-quality", "target": "widget:stats-across-dims-quality", "via": "bright_clipping_fraction" }, { "source": "file", "target": "widget:summary", "via": "path" }, { "source": "file", "target": "widget:sunburst", "via": "path" }, { "source": "processor:raster-basic", "target": "widget:violin-basic", "via": "min_intensity" }, { "source": "processor:raster-basic", "target": "widget:violin-quality", "via": "nan_fraction" }, { "source": "processor:raster-quality", "target": "widget:violin-quality", "via": "bright_clipping_fraction" } ] }