From d985a44f26c32535c9d23ea094d1923ca9093278 Mon Sep 17 00:00:00 2001
From: David Cosgrove <davidacosgroveaz@gmail.com>
Date: Mon, 9 Dec 2024 16:29:17 +0000
Subject: [PATCH] Handle DOS files in SynthonSpaceSearch (#8075)

* Handle DOS files.

* Smaller test file.
Add DOS file to .gitattributes.

* Update Code/GraphMol/SynthonSpaceSearch/substructure_search_catch_tests.cpp

---------

Co-authored-by: David Cosgrove <david@cozchemix.co.uk>
Co-authored-by: Greg Landrum <greg.landrum@gmail.com>
---
 .gitattributes                                           | 2 +-
 Code/GraphMol/SynthonSpaceSearch/SynthonSpace.cpp        | 4 +++-
 .../GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt | 8 ++++++++
 .../substructure_search_catch_tests.cpp                  | 9 +++++++++
 4 files changed, 21 insertions(+), 2 deletions(-)
 create mode 100644 Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt

diff --git a/.gitattributes b/.gitattributes
index ec7598870..0f1249eff 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -16,6 +16,7 @@ CMakeLists.txt text
 
 # Declare files that will always have CRLF line endings on checkout.
 #*.sln text eol=crlf
+Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt eol=crlf
 
 # Declare files that will always have LF line endings on checkout.
 #*.svg text eol=lf
@@ -26,4 +27,3 @@ CMakeLists.txt text
 *.pkl binary
 *.fpb binary
 
-
diff --git a/Code/GraphMol/SynthonSpaceSearch/SynthonSpace.cpp b/Code/GraphMol/SynthonSpaceSearch/SynthonSpace.cpp
index dfc97fe2d..37889ef0e 100644
--- a/Code/GraphMol/SynthonSpaceSearch/SynthonSpace.cpp
+++ b/Code/GraphMol/SynthonSpaceSearch/SynthonSpace.cpp
@@ -26,6 +26,7 @@
 #include <GraphMol/SynthonSpaceSearch/SynthonSpaceSubstructureSearcher.h>
 #include <GraphMol/SynthonSpaceSearch/SynthonSet.h>
 #include <GraphMol/SmilesParse/SmilesParse.h>
+#include <RDGeneral/StreamOps.h>
 
 namespace RDKit::SynthonSpaceSearch {
 
@@ -123,7 +124,8 @@ void SynthonSpace::readTextFile(const std::string &inFilename) {
   std::string nextLine;
   int lineNum = 1;
   std::vector<std::string> nextSynthon;
-  while (getline(ifs, nextLine)) {
+  while (!ifs.eof()) {
+    nextLine = getLine(ifs);
     ++lineNum;
     if (nextLine.empty() || nextLine[0] == '#') {
       continue;
diff --git a/Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt b/Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt
new file mode 100644
index 000000000..d97c3d35d
--- /dev/null
+++ b/Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt
@@ -0,0 +1,8 @@
+SMILES	synton_id	synton#	reaction_id
+c1ccccc1C(=O)[1*]	1-1	0	amide-1
+C1CCCCC1C(=O)[1*]	1-2	0	amide-1
+Clc1ccccc1C(=O)[1*]	1-3	0	amide-1
+CCCCC(=O)[1*]		1-4	0	amide-1
+[1*]N1CCCC1		2-1	1	amide-1
+[1*]NCCCCC		2-2	1	amide-1
+[1*]NCC1CC1		2-3	1	amide-1
diff --git a/Code/GraphMol/SynthonSpaceSearch/substructure_search_catch_tests.cpp b/Code/GraphMol/SynthonSpaceSearch/substructure_search_catch_tests.cpp
index 4270063bb..06fd4ba5a 100644
--- a/Code/GraphMol/SynthonSpaceSearch/substructure_search_catch_tests.cpp
+++ b/Code/GraphMol/SynthonSpaceSearch/substructure_search_catch_tests.cpp
@@ -550,4 +550,13 @@ TEST_CASE("Greg Space Failure") {
 
   auto results = synthonspace.substructureSearch(*queryMol, params);
   CHECK(results.getHitMolecules().size() == 1);
+}
+
+TEST_CASE("DOS File") {
+  REQUIRE(rdbase);
+  std::string fName(rdbase);
+  std::string libName = fName + "/Code/GraphMol/SynthonSpaceSearch/data/amide_space_dos.txt";
+  SynthonSpace synthonspace;
+  synthonspace.readTextFile(libName);
+  CHECK(synthonspace.getNumProducts() == 12);
 }
\ No newline at end of file