Feat/cran submission

.Rbuildignore

@@ -6,3 +6,11 @@
 ^\.DS_Store$
 ^README\.md$
 ^roadmap\.md$
+^Benchmark$
+^gedi\.Rcheck$
+^cran-comments\.md$
+^\.claude$
+^Logo\.svg$
+^check_output\.txt$
+^check_run\.log$
+^nohup\.out$

.gitignore

@@ -58,3 +58,5 @@ rsconnect/
 
 # Temp directory for development
 temp/
+Benchmark/
+.claude/

DESCRIPTION

@@ -1,33 +1,28 @@
 Package: gedi
 Type: Package
 Title: Gene Expression Decomposition and Integration
-Version: 2.3.0
-Date: 2026-01-08
+Version: 2.3.1
+Date: 2026-02-08
 Authors@R: c(
     person("Arsham", "Mikaeili Namini", email = "arsham.mikaeilinamini@mail.mcgill.ca", role = c("aut", "cre")),
     person("Hamed", "S.Najafabadi", email = "hamed.najafabadi@mcgill.ca", role = c("aut"))
     )
-Maintainer: Arsham Mikaeili Namini <arsham.mikaeilinamini@mail.mcgill.ca>
 Description: A memory-efficient implementation for integrating gene expression data
-    from single-cell RNA sequencing experiments. GEDI v2 uses a C++ backend with
+    from single-cell RNA sequencing experiments. Uses a C++ backend with
     thin R wrappers to enable analysis of large-scale single-cell datasets. The
     package supports multiple data modalities including count matrices, paired
-    data (Splicing, Velocyto, CITE-seq), and binary indicators. It implements a latent variable model
-    with block coordinate descent optimization for dimensionality reduction and
-    batch effect correction.
+    data (splicing, RNA velocity, CITE-seq), and binary indicators. It implements
+    a latent variable model with block coordinate descent optimization for
+    dimensionality reduction and batch effect correction.
 License: MIT + file LICENSE
-URL: https://github.com/Arshammik/gedi
-BugReports: https://github.com/Arshammik/gedi/issues
+URL: https://github.com/csglab/gedi2
+BugReports: https://github.com/csglab/gedi2/issues
 Depends: R (>= 4.0.0)
-Imports: Rcpp (>= 1.0.0), R6 (>= 2.5.0), Matrix (>= 1.3.0), hdf5r,
-        methods, stats, utils
+Imports: Rcpp (>= 1.0.0), R6 (>= 2.5.0), Matrix (>= 1.3.0),
+        ggplot2, scales, methods, stats, utils
 LinkingTo: Rcpp, RcppEigen
-Suggests: Seurat, SingleCellExperiment
-SystemRequirements: C++14, GNU make, Eigen3 (>= 3.3.0)
+Suggests: hdf5r, uwot, digest, glmnet, Seurat, SingleCellExperiment
+SystemRequirements: C++14, GNU make
 Encoding: UTF-8
 RoxygenNote: 7.3.3
 Roxygen: list(markdown = TRUE)
-NeedsCompilation: yes
-Packaged: 2025-10-22 18:58:46 UTC; arshammikaeili
-Author: Arsham Mikaeili Namini [aut, cre],
-  Hamed S.Najafabadi [aut]

LICENSE

@@ -1,21 +1,2 @@
-MIT License
-
-Copyright (c) 2025 Arsham Mikaeili Namini
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+YEAR: 2025
+COPYRIGHT HOLDER: Arsham Mikaeili Namini

NAMESPACE

@@ -1,5 +1,9 @@
 # Generated by roxygen2: do not edit by hand
 
+S3method(print,gedi_dynamics)
+S3method(print,gedi_dynamics_svd)
+S3method(print,gedi_imputation)
+S3method(print,gedi_pathway_associations)
 export(CreateGEDIObject)
 export(check_optional_dependencies)
 export(gedi_to_seurat)
@@ -11,10 +15,6 @@ export(plot_embedding)
 export(plot_feature_ratio)
 export(plot_features)
 export(plot_vector_field)
-export(print.gedi_dynamics)
-export(print.gedi_dynamics_svd)
-export(print.gedi_imputation)
-export(print.gedi_pathway_associations)
 export(read_h5)
 export(read_h5ad)
 export(seurat_to_gedi)
@@ -25,14 +25,15 @@ importFrom(Matrix,Matrix)
 importFrom(Matrix,sparseMatrix)
 importFrom(Matrix,t)
 importFrom(Rcpp,sourceCpp)
-importFrom(hdf5r,"h5attr<-")
-importFrom(hdf5r,H5File)
-importFrom(hdf5r,existsGroup)
-importFrom(hdf5r,h5attr)
 importFrom(methods,as)
 importFrom(methods,is)
+importFrom(stats,coef)
+importFrom(stats,median)
 importFrom(stats,rnorm)
 importFrom(stats,runif)
-importFrom(utils,installed.packages)
+importFrom(stats,var)
+importFrom(utils,object.size)
 importFrom(utils,packageVersion)
+importFrom(utils,setTxtProgressBar)
+importFrom(utils,txtProgressBar)
 useDynLib(gedi, .registration = TRUE)

NEWS.md

@@ -0,0 +1,29 @@
+# gedi 2.3.1
+
+## CRAN Compliance
+
+* Replace all `cat()` / `print()` calls with `message()` or `warning()` for
+  suppressible console output.
+* Add `verbose` parameters to `seurat_to_gedi()`, `gedi_to_seurat()`,
+  `check_optional_dependencies()`, and `install_optional_dependencies()`.
+* Replace `cat()`-based progress bars with `txtProgressBar()` across imputation
+  and training routines.
+* Replace `installed.packages()` with `requireNamespace()` for dependency
+  checking.
+* Move `hdf5r` from Imports to Suggests (optional dependency for H5AD I/O).
+* Add `ggplot2` and `scales` to Imports; add `uwot` and `digest` to Suggests.
+* Use CRAN-required two-line LICENSE format.
+* Add `@return` documentation tags to all exported and documented functions.
+* Replace non-ASCII characters in C++ source files with ASCII equivalents.
+* Remove redundant Maintainer field from DESCRIPTION.
+
+# gedi 2.3.0
+
+* Initial public release with C++ backend and R6 interface.
+* Support for multiple data modalities (count matrices, paired data, binary
+  indicators).
+* Latent variable model with block coordinate descent optimization.
+* Dimensionality reduction, batch correction, and imputation.
+* Differential expression and pathway association analysis.
+* H5AD file I/O for Python interoperability.
+* Seurat and SingleCellExperiment integration.

R/RcppExports.R

@@ -216,7 +216,7 @@ run_factorized_svd_cpp <- function(Z, projDB, verbose = 0L) {
 #' @return Dense matrix (J x Ni) - residual Yi after removing sample effects
 #'
 #' @details
-#' Computes: Yi - QiDBi - (si ⊗ 1ᵀ) - (o + oi) ⊗ 1ᵀ
+#' Computes: Yi - QiDBi - (si (x) 1^T) - (o + oi) (x) 1^T
 #' This leaves only the ZDBi component (shared biological signal).
 #'
 #' @keywords internal
@@ -238,7 +238,7 @@ Yi_resZ <- function(Yi, QiDBi, si, o, oi) {
 #' @return Dense matrix (J x Ni) - predicted Yi
 #'
 #' @details
-#' Computes: Ŷi = ZDBi + QiDBi + (si ⊗ 1ᵀ) + (o + oi) ⊗ 1ᵀ
+#' Computes: Y_hati = ZDBi + QiDBi + (si (x) 1^T) + (o + oi) (x) 1^T
 #' This is the full model prediction for log-expression.
 #'
 #' @keywords internal
@@ -262,7 +262,7 @@ predict_Yhat <- function(ZDBi, QiDBi, si, o, oi) {
 #'
 #' This comes from the Poisson-lognormal model where:
 #' - Mi ~ Poisson(exp(Yi))
-#' - Yi ~ N(Ŷi, sigma2)
+#' - Yi ~ N(Y_hati, sigma2)
 #'
 #' The posterior variance decreases where:
 #' - Counts are high (exp(Yi) large)
@@ -292,7 +292,7 @@ Yi_var_M <- function(Yi, sigma2) {
 #'
 #' This comes from the binomial-logistic-normal model where:
 #' - M1i ~ Binomial(M1i + M2i, p)
-#' - logit(p) = Yi ~ N(Ŷi, sigma2)
+#' - logit(p) = Yi ~ N(Y_hati, sigma2)
 #'
 #' The variance depends on:
 #' - Total counts M (more counts = less variance)
@@ -324,7 +324,7 @@ Yi_var_M_paired <- function(Yi, M1i, M2i, sigma2) {
 #' for downstream dispersion analysis (binning and aggregation done in R).
 #'
 #' Memory optimization: Works only on sparse nonzero positions (typically 5-10% of matrix).
-#' For 30K genes × 10K cells with 5% nonzero: samples from ~15M positions instead of 300M.
+#' For 30K genes x 10K cells with 5% nonzero: samples from ~15M positions instead of 300M.
 #'
 #' @keywords internal
 #' @noRd
@@ -362,7 +362,7 @@ Yi_SSE_M_paired <- function(Yi, M1i, M2i, ZDBi, QiDBi, si, o, oi, sigma2) {
 #'
 #' @param feature_weights Vector of length K (factor loadings for this feature)
 #' @param D Scaling vector of length K
-#' @param Bi_list List of sample-specific cell projection matrices (K × Ni each)
+#' @param Bi_list List of sample-specific cell projection matrices (K x Ni each)
 #' @param verbose Integer verbosity level
 #'
 #' @return Vector of length N (total cells) with projected values
@@ -378,12 +378,12 @@ compute_feature_projection <- function(feature_weights, D, Bi_list, verbose = 0L
 #' Projects multiple features through the GEDI model simultaneously.
 #' Computes: (feature_weights * D) %*% B for F features.
 #'
-#' @param feature_weights Matrix K × F (factor loadings for F features)
+#' @param feature_weights Matrix K x F (factor loadings for F features)
 #' @param D Scaling vector of length K
-#' @param Bi_list List of sample-specific cell projection matrices (K × Ni each)
+#' @param Bi_list List of sample-specific cell projection matrices (K x Ni each)
 #' @param verbose Integer verbosity level
 #'
-#' @return Matrix N × F with projected values for each feature
+#' @return Matrix N x F with projected values for each feature
 #'
 #' @keywords internal
 #' @noRd
@@ -420,18 +420,18 @@ aggregate_vectors <- function(Dim1, Dim2, To1, To2, color, alpha, n_bins, min_pe
 #' the concatenation of all sample-specific Bi matrices. This is the main
 #' integrated representation of cells in the GEDI latent space.
 #'
-#' @param Z Shared metagene matrix (J × K), where J = genes, K = latent factors
+#' @param Z Shared metagene matrix (J x K), where J = genes, K = latent factors
 #' @param D Scaling vector (length K) representing the importance of each factor
 #' @param Bi_list List of sample-specific cell projection matrices, where each
-#'   Bi is a K × Ni matrix (Ni = number of cells in sample i)
+#'   Bi is a K x Ni matrix (Ni = number of cells in sample i)
 #' @param verbose Integer verbosity level:
 #'   \itemize{
 #'     \item 0: Silent (no output)
 #'     \item 1: Progress bar and summary statistics
 #'     \item 2: Detailed per-sample information
 #'   }
 #'
-#' @return Dense matrix ZDB of dimensions J × N, where N = sum(Ni) is the total
+#' @return Dense matrix ZDB of dimensions J x N, where N = sum(Ni) is the total
 #'   number of cells across all samples. Each column represents a cell in the
 #'   integrated latent space.
 #'
@@ -441,17 +441,17 @@ aggregate_vectors <- function(Dim1, Dim2, To1, To2, color, alpha, n_bins, min_pe
 #' 
 #' Computational strategy:
 #' \enumerate{
-#'   \item Pre-compute ZD = Z * diag(D) once (saves K×J×numSamples operations)
+#'   \item Pre-compute ZD = Z * diag(D) once (saves KxJxnumSamples operations)
 #'   \item For each sample i: compute ZD * Bi and concatenate
 #'   \item Use Eigen block operations for efficient memory access
 #' }
 #'
-#' Memory: Allocates one J × N dense matrix. For large datasets (e.g., 20k genes,
+#' Memory: Allocates one J x N dense matrix. For large datasets (e.g., 20k genes,
 #' 50k cells), this requires ~8 GB RAM. Consider computing projections on demand
 #' or working with subsets if memory is limited.
 #'
 #' Performance: OpenMP parallelization available if enabled during compilation.
-#' Typical speed: ~100-200ms for 20k × 5k dataset on modern CPU.
+#' Typical speed: ~100-200ms for 20k x 5k dataset on modern CPU.
 #'
 #' @keywords internal
 #' @noRd
@@ -467,10 +467,10 @@ compute_ZDB_cpp <- function(Z, D, Bi_list, verbose = 0L) {
 #'
 #' @param D Scaling vector (length K) representing factor importance
 #' @param Bi_list List of sample-specific cell projection matrices, where each
-#'   Bi is a K × Ni matrix
+#'   Bi is a K x Ni matrix
 #' @param verbose Integer verbosity level (0 = silent, 1 = progress, 2 = detailed)
 #'
-#' @return Dense matrix DB of dimensions K × N, where N = sum(Ni). Each column
+#' @return Dense matrix DB of dimensions K x N, where N = sum(Ni). Each column
 #'   represents a cell's coordinates in the latent factor space.
 #'
 #' @details
@@ -488,7 +488,7 @@ compute_ZDB_cpp <- function(Z, D, Bi_list, verbose = 0L) {
 #'   \item Apply diagonal scaling: diag(D) * B
 #' }
 #'
-#' Memory: Much smaller than ZDB (K × N vs. J × N). For K=10 and N=50k cells,
+#' Memory: Much smaller than ZDB (K x N vs. J x N). For K=10 and N=50k cells,
 #' requires only ~4 MB vs. ~8 GB for ZDB when J=20k.
 #'
 #' Performance: Very fast (~10-50ms) since K << J typically.

R/gedi-package.R

@@ -55,7 +55,6 @@
 #' @seealso
 #' \itemize{
 #'   \item \code{\link{CreateGEDIObject}}: Create a GEDI model
-#'   \item \code{\link{GEDI}}: R6 class documentation
 #' }
 #'
 #' @examples
@@ -86,5 +85,6 @@
 #' @import R6
 #' @import Matrix
 #' @importFrom methods as is
-#' @importFrom stats rnorm runif
+#' @importFrom stats coef median rnorm runif var
+#' @importFrom utils object.size setTxtProgressBar txtProgressBar
 NULL