From c6634976c88980649866a65f6baa976c9ffa7e8d Mon Sep 17 00:00:00 2001
From: cutecutecat <starkind1997@gmail.com>
Date: Wed, 7 Jul 2021 11:54:02 -0500
Subject: [PATCH] Fix the forbidden symbols when using unpackdb (#467)

Fix unpackdb trying to produce files with forbidden filenames #466

<>:"/ \| ?* is not allowed in file path of Windows(all) and Linux(only /) when created separated database files with sequence names. Therefore, these symbols should be substituted with others.

Co-authored-by: Milot Mirdita <milot@mirdita.de>
---
 src/commons/FileUtil.cpp | 18 ++++++++++++++++++
 src/commons/FileUtil.h   |  3 +++
 src/util/unpackdb.cpp    |  3 ++-
 3 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/src/commons/FileUtil.cpp b/src/commons/FileUtil.cpp
index f772abda8..c50421abc 100644
--- a/src/commons/FileUtil.cpp
+++ b/src/commons/FileUtil.cpp
@@ -465,3 +465,21 @@ void FileUtil::fixRlimitNoFile() {
         }
     }
 }
+
+std::string FileUtil::sanitizeFilename(std::string name){
+    static const std::vector<std::pair<char, char>> symbolTable =
+            {{'\\', '@'},
+             {'/', '@'},
+             {':', '@'},
+             {'*', '@'},
+             {'?', '@'},
+             {'<', '@'},
+             {'>', '@'},
+             {'|', '!'}};
+
+    std::vector<std::pair<char, char>>::const_iterator it;
+    for (it = symbolTable.begin(); it != symbolTable.end(); ++it) {
+        std::replace(name.begin(), name.end(), it->first, it->second);
+    }
+    return name;
+}
diff --git a/src/commons/FileUtil.h b/src/commons/FileUtil.h
index 53d088422..019783bcc 100644
--- a/src/commons/FileUtil.h
+++ b/src/commons/FileUtil.h
@@ -64,6 +64,9 @@ class FileUtil {
     static std::string createTemporaryDirectory(const std::string& basePath, const std::string& subDirectory);
 
     static void fixRlimitNoFile();
+
+    // remove forbidden symbols in from filenames
+    static std::string sanitizeFilename(std::string name);
 };
 
 
diff --git a/src/util/unpackdb.cpp b/src/util/unpackdb.cpp
index e7f54d264..b576145e6 100644
--- a/src/util/unpackdb.cpp
+++ b/src/util/unpackdb.cpp
@@ -27,6 +27,7 @@ int unpackdb(int argc, const char **argv, const Command& command) {
 
     size_t entries = reader.getSize();
     Debug::Progress progress(entries);
+
 #pragma omp parallel
     {
         unsigned int thread_idx = 0;
@@ -44,7 +45,7 @@ int unpackdb(int argc, const char **argv, const Command& command) {
             }
             if (par.unpackNameMode == Parameters::UNPACK_NAME_ACCESSION) {
                 size_t lookupId = reader.getLookupIdByKey(key);
-                name.append(reader.getLookupEntryName(lookupId));
+                name.append(FileUtil::sanitizeFilename(reader.getLookupEntryName(lookupId)));
             } else {
                 name.append(SSTR(key));
             }