From 538236772ce6edde390079c32533c1e70cf8bdca Mon Sep 17 00:00:00 2001
From: Adrian Kummerlaender
Date: Fri, 6 Oct 2017 14:26:44 +0200
Subject: Implement BASE64 decoding of locations given by dictionary index

---
 CMakeLists.txt     |  7 ++++++-
 example.cc         | 11 ++++++++---
 src/util/base64.cc | 29 +++++++++++++++++++++++++++++
 src/util/base64.h  | 10 ++++++++++
 4 files changed, 53 insertions(+), 4 deletions(-)
 create mode 100644 src/util/base64.cc
 create mode 100644 src/util/base64.h

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 2514635..da53cc3 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -13,10 +13,16 @@ include_directories(
 add_library(
 	DictzipQuery
 	SHARED
+		src/util/base64.cc
 		src/istream/stream.cc
 		src/istream/buffer.cc
 )
 
+target_link_libraries(
+	DictzipQuery
+		z
+)
+
 add_executable(
 	example
 		example.cc
@@ -25,7 +31,6 @@ add_executable(
 target_link_libraries(
 	example
 		DictzipQuery
-		z
 )
 
 install(
diff --git a/example.cc b/example.cc
index 26bf404..5af1975 100644
--- a/example.cc
+++ b/example.cc
@@ -1,4 +1,5 @@
 #include "istream/stream.h"
+#include "util/base64.h"
 
 #include <string>
 #include <iostream>
@@ -6,12 +7,16 @@
 int main() {
 	dictzip::Istream stream("gcide.dict.dz");
 
+	// Decode location of _Accession_
+	const std::size_t offset = dictzip::base64_decode("8Aw");
+	const std::size_t length = dictzip::base64_decode("Wt");
+
 	// Print the GCIDE definition of _Accession_
 	std::string data;
-	data.reserve(1453);
+	data.reserve(length);
 
-	stream.seekg(245808);
-	stream.read(const_cast<char*>(data.data()), 1453);
+	stream.seekg(offset);
+	stream.read(const_cast<char*>(data.data()), length);
 
 	std::cout << data.c_str() << std::endl;
 }
diff --git a/src/util/base64.cc b/src/util/base64.cc
new file mode 100644
index 0000000..4e8d0a4
--- /dev/null
+++ b/src/util/base64.cc
@@ -0,0 +1,29 @@
+#include "base64.h"
+
+#include <vector>
+#include <stdexcept>
+
+namespace dictzip {
+
+std::size_t base64_decode(const std::string& encoded) {
+	std::vector<std::int8_t> map(256,-1);
+
+    for ( int i = 0; i < 64; ++i ) {
+		map["ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"[i]] = i;
+	}
+
+	std::size_t value = 0;
+
+    for ( std::uint8_t c : encoded ) {
+        if ( map[c] == -1 ) {
+			throw std::invalid_argument("Invalid character in BASE64 string.");
+		} else {
+			value *= 64;
+			value += map[c];
+		}
+    }
+
+    return value;
+}
+
+}
diff --git a/src/util/base64.h b/src/util/base64.h
new file mode 100644
index 0000000..448609d
--- /dev/null
+++ b/src/util/base64.h
@@ -0,0 +1,10 @@
+#pragma once
+
+#include <cstdint>
+#include <string>
+
+namespace dictzip {
+
+std::size_t base64_decode(const std::string& encoded);
+
+}
-- 
cgit v1.2.3