Add frt

DESCRIPTION

@@ -18,9 +18,11 @@ Imports:
     Rcpp,
     magrittr,
     MASS,
-    stats
+    stats,
+    tidyr,
+    coop
 LinkingTo: 
     Rcpp,
     RcppEigen
 Depends: 
-    R (>= 3.50)
+    R (>= 3.5.0)

NAMESPACE

@@ -3,5 +3,6 @@
 export(staggered)
 export(staggered_cs)
 export(staggered_sa)
+import(Rcpp)
 importFrom(magrittr,"%>%")
 useDynLib(staggered)

R/RcppExports.R

@@ -5,3 +5,11 @@ eigenMapMatMult <- function(A, B) {
     .Call('_staggered_eigenMapMatMult', PACKAGE = 'staggered', A, B)
 }
 
+solve_least_squares_svd <- function(A, B) {
+    .Call('_staggered_solve_least_squares_svd', PACKAGE = 'staggered', A, B)
+}
+
+solve_least_squares_normal <- function(A, B) {
+    .Call('_staggered_solve_least_squares_normal', PACKAGE = 'staggered', A, B)
+}
+

README.Rmd

@@ -146,9 +146,11 @@ We also provide a Stata implementation (staggered_stata) via the RCall package,
 To install the staggered_stata package, the user first needs to install the github and RCall packages. This can be done with the following commands
 
     > net install github, from("https://haghish.github.io/github/")
-    > github install haghish/rcall, stable
+    > github install haghish/rcall, version("2.5.0")
 
-Note that the user must have R installed before installing the rcall package. The latest version of R can be downloaded [here](https://cloud.r-project.org/). The staggered_stata package can then be installed with
+**Important note:** the staggered_stata package was built under Rcall version 2.5.0, and the recent release of RCall version 3.0 has created some compatibility issues. We will try our best to fix these issues, but in the meantime it is best to install version 2.5.0, as in the command above.
+
+Note that the user must have R installed before installing the RCall package. The latest version of R can be downloaded [here](https://cloud.r-project.org/). The staggered_stata package can then be installed with
 
     > github install jonathandroth/staggered_stata
 
@@ -164,4 +166,4 @@ The syntax for staggered_stata is very similar to that for the staggered R packa
 
 ![Stata examples.](man/figures/Stata_screenshot.png)
 
-The staggered, staggered_cs, and staggered_as commands all return a vector of coefficients and covariance matrix in ereturn list, and thus can be used with any post-estimation command in Stata (e.g. coefplot).
+The staggered, staggered_cs, and staggered_as commands all return a vector of coefficients and covariance matrix in ereturn list, and thus can be used with any post-estimation command in Stata (e.g. coefplot). 

README.md

@@ -4,6 +4,7 @@
 # staggered
 
 <!-- badges: start -->
+
 <!-- badges: end -->
 
 The staggered R package computes the efficient estimator for settings
@@ -45,7 +46,7 @@ for modifying and plotting the results.
 ``` r
 library(staggered) #load the staggered package
 library(dplyr) #load dplyr for data manipulation
-#> Warning: package 'dplyr' was built under R version 4.0.2
+#> Warning: package 'dplyr' was built under R version 4.0.5
 #> 
 #> Attaching package: 'dplyr'
 #> The following objects are masked from 'package:stats':
@@ -55,7 +56,7 @@ library(dplyr) #load dplyr for data manipulation
 #> 
 #>     intersect, setdiff, setequal, union
 library(ggplot2) #load ggplot2 for plotting the results
-#> Warning: package 'ggplot2' was built under R version 4.0.2
+#> Warning: package 'ggplot2' was built under R version 4.0.5
 library(purrr)
 
 df <- staggered::pj_officer_level_balanced #load the officer data
@@ -204,9 +205,15 @@ the github and RCall packages. This can be done with the following
 commands
 
     > net install github, from("https://haghish.github.io/github/")
-    > github install haghish/rcall, stable
+    > github install haghish/rcall, version("2.5.0")
+
+**Important note:** the staggered\_stata package was built under Rcall
+version 2.5.0, and the recent release of RCall version 3.0 has created
+some compatibility issues. We will try our best to fix these issues, but
+in the meantime it is best to install version 2.5.0, as in the command
+above.
 
-Note that the user must have R installed before installing the rcall
+Note that the user must have R installed before installing the RCall
 package. The latest version of R can be downloaded
 [here](https://cloud.r-project.org/). The staggered\_stata package can
 then be installed with

src/RcppExports.cpp

@@ -18,9 +18,35 @@ BEGIN_RCPP
     return rcpp_result_gen;
 END_RCPP
 }
+// solve_least_squares_svd
+Eigen::MatrixXd solve_least_squares_svd(Eigen::MatrixXd A, Eigen::MatrixXd B);
+RcppExport SEXP _staggered_solve_least_squares_svd(SEXP ASEXP, SEXP BSEXP) {
+BEGIN_RCPP
+    Rcpp::RObject rcpp_result_gen;
+    Rcpp::RNGScope rcpp_rngScope_gen;
+    Rcpp::traits::input_parameter< Eigen::MatrixXd >::type A(ASEXP);
+    Rcpp::traits::input_parameter< Eigen::MatrixXd >::type B(BSEXP);
+    rcpp_result_gen = Rcpp::wrap(solve_least_squares_svd(A, B));
+    return rcpp_result_gen;
+END_RCPP
+}
+// solve_least_squares_normal
+Eigen::MatrixXd solve_least_squares_normal(Eigen::MatrixXd A, Eigen::MatrixXd B);
+RcppExport SEXP _staggered_solve_least_squares_normal(SEXP ASEXP, SEXP BSEXP) {
+BEGIN_RCPP
+    Rcpp::RObject rcpp_result_gen;
+    Rcpp::RNGScope rcpp_rngScope_gen;
+    Rcpp::traits::input_parameter< Eigen::MatrixXd >::type A(ASEXP);
+    Rcpp::traits::input_parameter< Eigen::MatrixXd >::type B(BSEXP);
+    rcpp_result_gen = Rcpp::wrap(solve_least_squares_normal(A, B));
+    return rcpp_result_gen;
+END_RCPP
+}
 
 static const R_CallMethodDef CallEntries[] = {
     {"_staggered_eigenMapMatMult", (DL_FUNC) &_staggered_eigenMapMatMult, 2},
+    {"_staggered_solve_least_squares_svd", (DL_FUNC) &_staggered_solve_least_squares_svd, 2},
+    {"_staggered_solve_least_squares_normal", (DL_FUNC) &_staggered_solve_least_squares_normal, 2},
     {NULL, NULL, 0}
 };
 

src/code.cpp

@@ -10,3 +10,24 @@ SEXP eigenMapMatMult(const Eigen::Map<Eigen::MatrixXd> A, Eigen::Map<Eigen::Matr
 
   return Rcpp::wrap(C);
 }
+
+
+// [[Rcpp::export]]
+Eigen::MatrixXd solve_least_squares_svd(
+    Eigen::MatrixXd A,
+    Eigen::MatrixXd B
+) {
+  Eigen::MatrixXd x = A.bdcSvd(Eigen::ComputeThinU | Eigen::ComputeThinV).solve(B);
+  return(x);
+}
+
+
+// [[Rcpp::export]]
+Eigen::MatrixXd solve_least_squares_normal(
+    Eigen::MatrixXd A,
+    Eigen::MatrixXd B
+) {
+  Eigen::MatrixXd x = (A.transpose() * A).ldlt().solve(A.transpose() * B);
+  return(x);
+}
+