From 78a3f0eda630c31595b49a5432093c12a3651494 Mon Sep 17 00:00:00 2001
From: Franck Dary <franck.dary@lis-lab.fr>
Date: Sat, 21 Mar 2020 17:14:56 +0100
Subject: [PATCH] replacing all newline and tab in rawInput with spaces
---
common/include/utf8string.hpp | 1 +
common/src/utf8string.cpp | 5 +++++
reading_machine/src/BaseConfig.cpp | 2 ++
3 files changed, 8 insertions(+)
diff --git a/common/include/utf8string.hpp b/common/include/utf8string.hpp
index 42fe5d4..2a7400e 100644
--- a/common/include/utf8string.hpp
+++ b/common/include/utf8string.hpp
@@ -29,6 +29,7 @@ class utf8string : public std::vector<utf8char>
utf8string & operator=(const std::string & other);
utf8string & operator=(const char * const other);
bool operator==(const std::string & other);
+ void replace(utf8char from, utf8char to);
};
};
diff --git a/common/src/utf8string.cpp b/common/src/utf8string.cpp
index d430197..42755db 100644
--- a/common/src/utf8string.cpp
+++ b/common/src/utf8string.cpp
@@ -70,3 +70,8 @@ bool util::utf8string::operator==(const std::string & other)
return true;
}
+void util::utf8string::replace(utf8char from, utf8char to)
+{
+ std::replace(begin(), end(), from, to);
+}
+
diff --git a/reading_machine/src/BaseConfig.cpp b/reading_machine/src/BaseConfig.cpp
index 696d76c..f3a1ed9 100644
--- a/reading_machine/src/BaseConfig.cpp
+++ b/reading_machine/src/BaseConfig.cpp
@@ -49,6 +49,8 @@ void BaseConfig::readRawInput(std::string_view rawFilename)
std::fclose(file);
rawInputUtf8 = util::splitAsUtf8(rawInputTemp);
+ rawInputUtf8.replace(util::utf8char("\n"), util::utf8char(" "));
+ rawInputUtf8.replace(util::utf8char("\t"), util::utf8char(" "));
}
void BaseConfig::readTSVInput(std::string_view tsvFilename)
--
GitLab