rename package -> 'survinng'

nkoenen · nkoenen · commit 370e4315434e · 2025-05-28T22:53:47.000+02:00
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,4 +1,4 @@
-Package: Survinng
+Package: survinng
 Title: Gradient-Based Feature Attribution for Survival Neural Networks
 Version: 0.1.0
 Authors@R: c(
@@ -42,5 +42,5 @@ Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.2.3
 Config/Needs/website: rmarkdown
-URL: https://bips-hb.github.io/Survinng/
+URL: https://bips-hb.github.io/survinng/
 VignetteBuilder: knitr
diff --git a/NEWS.md b/NEWS.md
@@ -1,4 +1,4 @@
-# Survinng 0.1.0
+# survinng 0.1.0
 
 * Initial release as part of the ICML'25 paper *Gradient-based 
   Explanations for Deep Learning Survival Models*
diff --git a/R/BatchLoader.R b/R/BatchLoader.R
@@ -18,7 +18,7 @@ Surv_BatchLoader <- function(inputs, inputs_ref, method_pre_fun, model_pre_fun,
                              scale_fun, batch_size, n_timepoints, n) {
   batch_size <- min(batch_size, dim(inputs[[1]])[1] * n_timepoints)
   if (batch_size < n_timepoints) {
-    message("[Survinng] Batch size is smaller than the number of timepoints. Setting ",
+    message("[survinng] Batch size is smaller than the number of timepoints. Setting ",
             "batch size (", batch_size, ") to the number of timepoints (",
             n_timepoints, ").")
     batch_size <- n_timepoints
@@ -51,7 +51,7 @@ setMethod("update", "Surv_BatchLoader", function(object) {
 
   # Update the current index
   if (object@current_index + length(idx) > object@max_index) {
-    warning("[Survinng] You'are trying to get more batches than available. ",
+    warning("[survinng] You'are trying to get more batches than available. ",
             "Resetting the current index to 0.")
     object@current_index <- 0
   } else {
diff --git a/R/base_method.R b/R/base_method.R
@@ -28,7 +28,7 @@ base_method <- function(exp, instance, n = 1, model_class, inputs_ref = NULL,
     final_batch_size <- n_timepoints * (final_batch_size %/% n_timepoints)
   }
   num_dims <- sum(sapply(inputs, function(x) final_batch_size * prod(dim(x)[-1])))
-  if (num_dims > 20000) {
+  if (num_dims > 2e6) {
     warning("The resulting tensor has ", num_dims, " entries in total per batch. ",
             "This may lead to runtime or memory issues. Consider using a ",
             "smaller batch size (currently ", batch_size, ") ",
diff --git a/README.md b/README.md
@@ -1,17 +1,17 @@
 
-# `Survinng`: Gradient-based Explanations for Deep Learning Survival Models
+# `survinng`: Gradient-based Explanations for Deep Learning Survival Models
 
 <!-- badges: start -->
-<a href="https://bips-hb.github.io/Survinng/"><img src="man/figures/logo.jpeg" align="right" height="120" alt="Survinng website" /></a>
+<a href="https://bips-hb.github.io/survinng/"><img src="man/figures/logo.jpeg" align="right" height="120" alt="survinng website" /></a>
 [![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)
 [![arXiv](https://img.shields.io/badge/arXiv-2502.04970-b31b1b.svg)](https://arxiv.org/abs/2502.04970)
-[![Website](https://img.shields.io/badge/docs-📘%20Survinng%20Site-blue)](https://bips-hb.github.io/Survinng/)
+[![Website](https://img.shields.io/badge/docs-📘%20survinng%20Site-blue)](https://bips-hb.github.io/survinng/)
 [![Lifecycle: experimental](https://img.shields.io/badge/lifecycle-experimental-orange.svg)](https://lifecycle.r-lib.org/articles/stages.html#experimental)
-[![R-CMD-check](https://github.com/bips-hb/Survinng/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/bips-hb/Survinng/actions/workflows/R-CMD-check.yaml)
+[![R-CMD-check](https://github.com/bips-hb/survinng/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/bips-hb/survinng/actions/workflows/R-CMD-check.yaml)
 <!-- badges: end -->
 
 
-The package `Survinng` provides gradient-based feature attribution methods 
+The package `survinng` provides gradient-based feature attribution methods 
 for deep learning survival models. It implements novel adaptions of model-specific
 XAI methods for the survival analysis domain, like **Grad(t)**, **SG(t)**, **GxI(t)**,
 **IntGrad(t)**, and **GradSHAP(t)**. The whole package is designed to be compatible with the
@@ -21,9 +21,9 @@ However, it can also be used with models from `pycox` directly and other
 survival models loaded in `torch`. Currently, the package supports the models
 types `DeepSurv`/`CoxPH`, `DeepHit`, and `CoxTime`.
 
-### 🚀 Why `Survinng`?
+### 🚀 Why `survinng`?
 
-With `Survinng`, you get:
+With `survinng`, you get:
 
 * ⏱ **Model-specific** and **time-resolved** feature attributions for individuals  
 * ⚡️ **Fast and scalable** explanations using model gradients, especially for SHAP-like explanations  
@@ -45,25 +45,25 @@ To install the latest development version directly from GitHub:
 
 ```r
 # install.packages("devtools")
-devtools::install_github("bips-hb/Survinng")
+devtools::install_github("bips-hb/survinng")
 ```
 
 ## 📖 Usage
 
 You have a trained survival neural network model from `survivalmodels` or
 `pycox` and your model input data data. Now you want to interpret individual 
-data points by using the methods from the package `Survinng`, then stick to the 
+data points by using the methods from the package `survinng`, then stick to the 
 following pseudo code:
 
 ```r
-library(Survinng)
+library(survinng)
 
 # Load a survival model and corresponding data
 model <- ... (e.g., from survivalmodels or pycox)
 data <- ... (e.g., the test set of the model)
 
 # Create explainer object
-explainer <- Survinng::explain(model, data)
+explainer <- survinng::explain(model, data)
 
 # Compute feature attributions
 idx <- 1 # index of the instance to explain
@@ -77,7 +77,7 @@ shap <- surv_gradSHAP(explainer, instance = idx) # GradSHAP(t)
 plot(shap)
 ```
 👉 For full documentation and advanced use cases, visit the 
-[📘 package website](https://bips-hb.github.io/Survinng/).
+[📘 package website](https://bips-hb.github.io/survinng/).
 
 ## 💻 Quick Example
 
@@ -102,7 +102,7 @@ ext <- callr::r(function(train) {
                   early_stopping = TRUE)
 
   # Extract the model
-  Survinng::extract_model(fit)
+  survinng::extract_model(fit)
 }, args = list(train = train))
 
 
@@ -118,10 +118,10 @@ plot(shap)
 
 ## 🖥 Other Examples and Articles
 
-- How to load my model? [→ article](https://bips-hb.github.io/Survinng/articles/how_to_load_model.html)
-- Simulation: Time-independent effects (`survivalmodels`) [→ article](https://bips-hb.github.io/Survinng/articles/Sim_time_independent.html)
-- Simulation: Time-dependent effects (`survivalmodels`) [→ article](https://bips-hb.github.io/Survinng/articles/Sim_time_dependent.html)
-- Example on real multi-modal medical data [→ article](https://bips-hb.github.io/Survinng/articles/multimodal.html)
+- How to load my model? [→ article](https://bips-hb.github.io/survinng/articles/how_to_load_model.html)
+- Simulation: Time-independent effects (`survivalmodels`) [→ article](https://bips-hb.github.io/survinng/articles/Sim_time_independent.html)
+- Simulation: Time-dependent effects (`survivalmodels`) [→ article](https://bips-hb.github.io/survinng/articles/Sim_time_dependent.html)
+- Example on real multi-modal medical data [→ article](https://bips-hb.github.io/survinng/articles/multimodal.html)
 
 
 ## 📚 Citation
diff --git a/_pkgdown.yml b/_pkgdown.yml
@@ -1,4 +1,4 @@
-url: https://bips-hb.github.io/Survinng/
+url: https://bips-hb.github.io/survinng/
 template:
   bootstrap: 5
 
diff --git a/man/survinng-package.Rd b/man/survinng-package.Rd
diff --git a/tests/testthat.R b/tests/testthat.R
@@ -7,13 +7,13 @@
 # * https://testthat.r-lib.org/articles/special-files.html
 
 library(testthat)
-library(Survinng)
+library(survinng)
 
 if (Sys.getenv("TORCH_TEST", unset = 0) == 1) {
   set.seed(42)
   torch::torch_manual_seed(42)
 
   library(torch)
 
-  test_check("Survinng")
+  test_check("survinng")
 }
diff --git a/vignettes/articles/Sim_time_dependent.Rmd b/vignettes/articles/Sim_time_dependent.Rmd
@@ -17,7 +17,7 @@ knitr::opts_chunk$set(
 Load the necessary libraries and source the utility functions.
 
 ```{r setup, message=FALSE, warning=FALSE}
-library(Survinng)
+library(survinng)
 library(ggplot2)
 library(cowplot)
 library(dplyr)
@@ -60,7 +60,7 @@ We consider a simulated dataset with the following characteristics:
 
 ## Load Models and Data
 
-The models used in this vignette are the same as those used in the main paper. The models were trained using the `survivalmodels` package, and the training process is not shown here, but can be found in the `vignettes/articles/Sim_time_dependent` directory or on [GitHub](https://github.com/bips-hb/Survinng/tree/master/vignettes/articles/Sim_time_dependent).
+The models used in this vignette are the same as those used in the main paper. The models were trained using the `survivalmodels` package, and the training process is not shown here, but can be found in the `vignettes/articles/Sim_time_dependent` directory or on [GitHub](https://github.com/bips-hb/survinng/tree/master/vignettes/articles/Sim_time_dependent).
 
 ```{r td fit models, fig.width=10, fig.height=15}
 # Load data
@@ -97,9 +97,9 @@ knitr::include_graphics(here('vignettes/articles/Sim_time_dependent/sim_td_brier
 The `explain()` function creates an explainer object for the survival models. The `data` argument specifies the dataset used for explanation, and the `model` argument specifies the model to be explained. The `target` argument indicates the type of prediction to be explained (e.g., "survival", "risk", "cumulative hazard").
 
 ```{r td explainer}
-exp_deephit <- Survinng::explain(ext_deephit[[1]], data = test)
-exp_coxtime <- Survinng::explain(ext_coxtime[[1]], data = test)
-exp_deepsurv <- Survinng::explain(ext_deepsurv[[1]], data = test)
+exp_deephit <- survinng::explain(ext_deephit[[1]], data = test)
+exp_coxtime <- survinng::explain(ext_coxtime[[1]], data = test)
+exp_deepsurv <- survinng::explain(ext_deepsurv[[1]], data = test)
 ```
 
 
diff --git a/vignettes/articles/Sim_time_independent.Rmd b/vignettes/articles/Sim_time_independent.Rmd
@@ -16,7 +16,7 @@ knitr::opts_chunk$set(
 Load the necessary libraries and source the utility functions.
 
 ```{r setup, message=FALSE, warning=FALSE}
-library(Survinng)
+library(survinng)
 library(ggplot2)
 library(cowplot)
 library(dplyr)
@@ -55,7 +55,7 @@ We consider a simulated dataset with the following characteristics:
 
 ## Load Models and Data
 
-The models used in this vignette are the same as those used in the main paper. The models were trained using the `survivalmodels` package, and the training process is not shown here, but can be found in the `vignettes/articles/Sim_time_independent` directory or on [GitHub](https://github.com/bips-hb/Survinng/tree/master/vignettes/articles/Sim_time_independent).
+The models used in this vignette are the same as those used in the main paper. The models were trained using the `survivalmodels` package, and the training process is not shown here, but can be found in the `vignettes/articles/Sim_time_independent` directory or on [GitHub](https://github.com/bips-hb/survinng/tree/master/vignettes/articles/Sim_time_independent).
 
 ```{r td fit models, fig.width=10, fig.height=15}
 # Load data
@@ -93,9 +93,9 @@ The `explain()` function creates an explainer object for the survival models. Th
 
 
 ```{r tid explainer}
-exp_deephit <- Survinng::explain(ext_deephit[[1]], data = test)
-exp_coxtime <- Survinng::explain(ext_coxtime[[1]], data = test)
-exp_deepsurv <- Survinng::explain(ext_deepsurv[[1]], data = test)
+exp_deephit <- survinng::explain(ext_deephit[[1]], data = test)
+exp_coxtime <- survinng::explain(ext_coxtime[[1]], data = test)
+exp_deepsurv <- survinng::explain(ext_deepsurv[[1]], data = test)
 ```
 
 ## Survival Prediction
diff --git a/vignettes/articles/how_to_load_model.Rmd b/vignettes/articles/how_to_load_model.Rmd
@@ -15,11 +15,11 @@ torch::torch_manual_seed(42)
 ```
 
 ```{r setup, warning=FALSE, message=FALSE}
-library(Survinng)
+library(survinng)
 library(here)
 ```
 
-The `Survinng` package is designed in a flexible way to allow users to load 
+The `survinng` package is designed in a flexible way to allow users to load 
 models trained in R or Python using `survivalmodels` or `pycox`, respectively.
 Under the hood, the package uses the `torch` package and any base neural network
 can be loaded and will be coverted to the corresponding survival model, e.g.,
@@ -78,7 +78,7 @@ print(output$shape)
 ```
 
 Now, we have our (untrained) base model and corresponding data. The next step
-is to create a `Survinng` explainer object. This is done using the `explain()`
+is to create a `survinng` explainer object. This is done using the `explain()`
 function. The function takes the model, data, and time bins (required for
 a DeepHit model) as input and internally combines the base model with the
 DeepHit architecture. The `model_type` argument is set to `"deephit"` to 
@@ -88,7 +88,7 @@ feature attribution methods in the survival context.
 
 ```{r, fig.width= 15, fig.height= 7}
 # Create explainer object
-exp_deephit <- Survinng::explain(model_deephit, data = data, model_type = "deephit",
+exp_deephit <- survinng::explain(model_deephit, data = data, model_type = "deephit",
                                  time_bins = time_bins)
 print(exp_deephit)
 
@@ -137,7 +137,7 @@ print(output$shape)
 ```
 
 Now, we have our (untrained) base model and corresponding data, including the
-baseline hazard values The next step is to create a `Survinng` explainer
+baseline hazard values The next step is to create a `survinng` explainer
 object. This is done using the `explain()` function. The function takes the
 model, data, and baseline hazard (required for a DeepSurv model) as input
 and internally combines the base model with the DeepSurv architecture. The
@@ -146,7 +146,7 @@ DeepSurv model.
 
 ```{r, fig.width= 15, fig.height= 7}
 # Create explainer object
-exp_deepsurv <- Survinng::explain(model_deepsurv, data = data, 
+exp_deepsurv <- survinng::explain(model_deepsurv, data = data, 
                                   model_type = "deepsurv",
                                   baseline_hazard = base_hazard)
 print(exp_deepsurv)
@@ -234,7 +234,7 @@ print(output$shape) # the instances are replicated t times
 ```
 
 Now, we have our (untrained) base model and corresponding data, including the
-baseline hazard values. The next step is to create a `Survinng` explainer
+baseline hazard values. The next step is to create a `survinng` explainer
 object. This is done using the `explain()` function. The function takes the
 model, data, and baseline hazard (required for a CoxTime model) as input
 and internally combines the base model with the CoxTime architecture. The
@@ -244,7 +244,7 @@ earlier to add the time points to the input features.
 
 ```{r, fig.width= 15, fig.height= 7}
 # Create explainer object
-exp_coxtime <- Survinng::explain(model_coxtime, data = data, 
+exp_coxtime <- survinng::explain(model_coxtime, data = data, 
                                  model_type = "coxtime",
                                  baseline_hazard = base_hazard)
 print(exp_coxtime)
@@ -261,7 +261,7 @@ plot(ig, type = "contr")
 
 The `survivalmodels` package is a R package for survival analysis using deep
 learning. Especially for the deep learning models, it uses the `pycox` package 
-in Python under the hood via `reticulate`. The `Survinng` package is designed
+in Python under the hood via `reticulate`. The `survinng` package is designed
 to load models trained in `survivalmodels` directly. However, there are some
 issues when `torch` in R (i.e., `LibTorch`) and `Pytorch` in Python are 
 loaded in the same session (see [issue #815](https://github.com/mlverse/torch/issues/815)).
@@ -304,10 +304,10 @@ models <- callr::r(function(data) {
   fit_coxtime <- coxtime(Surv(time, status) ~ ., data = data, epochs = 10,
                          num_nodes = 8L)
 
-  # Extract the models using `Survinng`
-  ext_deepsurv <- Survinng::extract_model(fit_deepsurv)
-  ext_deephit <- Survinng::extract_model(fit_deephit)
-  ext_coxtime <- Survinng::extract_model(fit_coxtime)
+  # Extract the models using `survinng`
+  ext_deepsurv <- survinng::extract_model(fit_deepsurv)
+  ext_deephit <- survinng::extract_model(fit_deephit)
+  ext_coxtime <- survinng::extract_model(fit_coxtime)
   
   # Return the extracted models
   list(deepsurv = ext_deepsurv, deephit = ext_deephit, coxtime = ext_coxtime)
@@ -316,15 +316,15 @@ models <- callr::r(function(data) {
 
 
 Now, we have our trained models in the `models` list. The models are
-extracted using the `extract_model()` function from the `Survinng` package.
+extracted using the `extract_model()` function from the `survinng` package.
 These objects and the data to be explained can be passed to the `explain()` 
-function to create the `Survinng` explainer objects. 
+function to create the `survinng` explainer objects. 
 
 ```{r}
 # Create explainer objects
-exp_deepsurv <- Survinng::explain(models$deepsurv, data = data)
-exp_deephit <- Survinng::explain(models$deephit, data = data)
-exp_coxtime <- Survinng::explain(models$coxtime, data = data)
+exp_deepsurv <- survinng::explain(models$deepsurv, data = data)
+exp_deephit <- survinng::explain(models$deephit, data = data)
+exp_coxtime <- survinng::explain(models$coxtime, data = data)
 ```
 
 The `explain()` function will automatically detect the model type and
@@ -466,15 +466,15 @@ labtrans <- list(
 )
 ```
 
-Now, we can create the `Survinng` explainer object using the `explain()` function
+Now, we can create the `survinng` explainer object using the `explain()` function
 and apply all implemented gradient-based feature attribution methods on this 
 model. The `model_type` argument is set to `"coxtime"` to indicate that the
 model is a CoxTime model and the `labtrans` argument is set to the defined 
 transformation.
 
 ```{r, fig.width= 10, fig.height= 15}
 # Create explainer
-explainer <- Survinng::explain(model, data = x_val, model_type = "coxtime",
+explainer <- survinng::explain(model, data = x_val, model_type = "coxtime",
                                baseline_hazard = base_hazard,
                                labtrans = labtrans)
 
diff --git a/vignettes/articles/multimodal.Rmd b/vignettes/articles/multimodal.Rmd

Original file line number	Diff line number	Diff line change
`@@ -28,7 +28,7 @@ base_method <- function(exp, instance, n = 1, model_class, inputs_ref = NULL,`
`28`	`28`	`final_batch_size <- n_timepoints * (final_batch_size %/% n_timepoints)`
`29`	`29`	`}`
`30`	`30`	`num_dims <- sum(sapply(inputs, function(x) final_batch_size * prod(dim(x)[-1])))`
`31`		`- if (num_dims > 20000) {`
	`31`	`+ if (num_dims > 2e6) {`
`32`	`32`	`warning("The resulting tensor has ", num_dims, " entries in total per batch. ",`
`33`	`33`	`"This may lead to runtime or memory issues. Consider using a ",`
`34`	`34`	`"smaller batch size (currently ", batch_size, ") ",`
Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-url: https://bips-hb.github.io/Survinng/`
	`1`	`+url: https://bips-hb.github.io/survinng/`
`2`	`2`	`template:`
`3`	`3`	`bootstrap: 5`
`4`	`4`