Merge pull request #307071 from Kupac/fix_immunotation

rPackages.immunotation: cache external URLs

2 files changed, 84 insertions, 0 deletions

diff --git a/pkgs/development/r-modules/default.nix b/pkgs/development/r-modules/default.nix
index 8d4004d7580f..ce5f581be1bc 100644
--- a/pkgs/development/r-modules/default.nix
+++ b/pkgs/development/r-modules/default.nix
@@ -1498,6 +1498,43 @@ let
       buildInputs = [ cacert ] ++ attrs.buildInputs;
     });
 
+
+    immunotation = let
+      MHC41alleleList = fetchurl {
+        url = "https://services.healthtech.dtu.dk/services/NetMHCpan-4.1/allele.list";
+        hash = "sha256-CRZ+0uHzcq5zK5eONucAChXIXO8tnq5sSEAS80Z7jhg=";
+      };
+
+      MHCII40alleleList = fetchurl {
+        url = "https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.0/alleles_name.list";
+        hash = "sha256-K4Ic2NUs3P4IkvOODwZ0c4Yh8caex5Ih0uO5jXRHp40=";
+      };
+
+      # List of valid countries, regions and ethnic groups
+      # The original page is changing a bit every day, but the relevant
+      # content does not. Use archive.org to get a stable snapshot.
+      # It can be updated from time to time, or when the package becomes
+      # deficient. This may be difficult to know.
+      # Update the snapshot date, and add id_ after it, as described here:
+      # https://web.archive.org/web/20130806040521/http://faq.web.archive.org/page-without-wayback-code/
+      validGeographics = fetchurl {
+        url = "https://web.archive.org/web/20240418194005id_/http://www.allelefrequencies.net/hla6006a.asp";
+        hash = "sha256-m7Wkmh/cPxeqn94LwoznIh+fcFXskmSGErUYj6kTqak=";
+      };
+    in old.immunotation.overrideAttrs (attrs: {
+      patches = [ ./patches/immunotation.patch ];
+      postPatch = ''
+        substituteInPlace "R/external_resources_input.R" --replace-fail \
+          "nix-NetMHCpan-4.1-allele-list" ${MHC41alleleList}
+
+        substituteInPlace "R/external_resources_input.R" --replace-fail \
+          "nix-NETMHCIIpan-4.0-alleles-name-list" ${MHCII40alleleList}
+
+        substituteInPlace "R/AFND_interface.R" --replace-fail \
+          "nix-valid-geographics" ${validGeographics}
+      '';
+    });
+
     rstan = old.rstan.overrideAttrs (attrs: {
       env = (attrs.env or { }) // {
         NIX_CFLAGS_COMPILE = attrs.env.NIX_CFLAGS_COMPILE + " -DBOOST_PHOENIX_NO_VARIADIC_EXPRESSION";
diff --git a/pkgs/development/r-modules/patches/immunotation.patch b/pkgs/development/r-modules/patches/immunotation.patch
new file mode 100644
index 000000000000..2019e8e50a02
--- /dev/null
+++ b/pkgs/development/r-modules/patches/immunotation.patch
@@ -0,0 +1,47 @@
+diff --git a/R/AFND_interface.R b/R/AFND_interface.R
+index b62e8e0..0f22d85 100644
+--- a/R/AFND_interface.R
++++ b/R/AFND_interface.R
+@@ -244,9 +244,9 @@ check_population <- function(hla_population) {
+ #' @return list of valid countries, regions and ethnic origin
+ #' @keywords internal
+ get_valid_geographics <- function() {
+-    url <- "http://www.allelefrequencies.net/hla6006a.asp?"
+-    html_input <- getURL(url, read_method = "html")
+-    
++    # http://www.allelefrequencies.net/hla6006a.asp?
++    html_input <- xml2::read_html("nix-valid-geographics")
++      
+     rvest_tables <- rvest::html_table(html_input, fill = TRUE)
+     
+     # country
+diff --git a/R/external_resources_input.R b/R/external_resources_input.R
+index c4b1dc1..8fc5881 100644
+--- a/R/external_resources_input.R
++++ b/R/external_resources_input.R
+@@ -74,16 +74,17 @@ getURL <- function(URL, N.TRIES=2L,
+ # MHC I
+ # netmhcI_input_template is an internal variable containing list of valid 
+ # NetMHCpan input alleles
+-netmhcI_input_template <- getURL(
+-    URL="https://services.healthtech.dtu.dk/services/NetMHCpan-4.1/allele.list",
+-    read_method = "delim", delim = "\t",
+-    col_names = c("netmhc_input", "hla_chain_name", "HLA_gene"))
++netmhcI_input_template <- readr::read_delim(
++        # https://services.healthtech.dtu.dk/services/NetMHCpan-4.1/allele.list
++	"nix-NetMHCpan-4.1-allele-list",
++	delim = "\t",
++	skip = 0,
++        col_names = c("netmhc_input", "hla_chain_name", "HLA_gene")
++    )
+ 
+ # MHC II
+-lines <- getURL(
+-    URL = paste0("https://services.healthtech.dtu.dk/services/",
+-    "NetMHCIIpan-4.0/alleles_name.list"),
+-    read_method = "lines")
++# https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.0/alleles_name.list
++lines <- readr::read_lines("nix-NETMHCIIpan-4.0-alleles-name-list")
+ lines_rep <- stringr::str_replace_all(lines, "\t+|\\s\\s+", "\t")
+ netmhcII_input_template <- suppressWarnings(
+     suppressMessages(read.delim(textConnection(lines_rep), sep = "\t")))