From 8896def814571d549ecb6c412182e27686926702 Mon Sep 17 00:00:00 2001
From: Dnomd343 <i@343.re>
Date: Sun, 15 Jan 2023 11:48:00 +0800
Subject: [PATCH] perf: short code fast decode and encode

---
 src/main.cc                   | 25 -------------------------
 src/short_code/data_loader.cc | 26 +++++++-------------------
 2 files changed, 7 insertions(+), 44 deletions(-)

diff --git a/src/main.cc b/src/main.cc
index 3a0ec5c..86f286e 100644
--- a/src/main.cc
+++ b/src/main.cc
@@ -262,32 +262,7 @@ int main() {
 //        }
 //    }
 
-
-    const uint32_t ALL_CASES_OFFSET[16] = {
-    ///        0         1         2         3
-               0,  2942906,  5203298,  8146204,
-    ///        4         5         6         7
-         8146204, 10468254, 12345199, 14667249,
-    ///        8         9        10        11
-        14667249, 16989299, 18866244, 21188294,
-    ///       12        13        14        15
-        21188294, 24131200, 26391592, 29334498,
-    };
-
-    /// 2942905 -> 0
-    /// 2942906 -> 1
-    /// 8146203 -> 2
-    /// 8146204 -> 4
-    /// 8146205 -> 4
-    /// 29334497 -> 14
-
-//    std::cout << std::upper_bound(ALL_CASES_OFFSET, ALL_CASES_OFFSET + 16, 8146203) - ALL_CASES_OFFSET - 1 << std::endl;
-
-
     for (uint32_t i = 0; i < 29334498; ++i) {
-//        std::cout << ShortCode(i).to_common_code().to_string() << std::endl;
-//        if (CommonCode::unsafe_create(ShortCode::fast_decode(i)).to_short_code().unwrap() != i) {
-//        if (ShortCode(i).to_common_code().to_short_code().unwrap() != i) {
         if (ShortCode::fast_encode(ShortCode::fast_decode(i)) != i) {
             std::cout << "error" << std::endl;
         }
diff --git a/src/short_code/data_loader.cc b/src/short_code/data_loader.cc
index f367288..41b598e 100644
--- a/src/short_code/data_loader.cc
+++ b/src/short_code/data_loader.cc
@@ -52,29 +52,17 @@ void ShortCode::build_mappings() { // build fast search mappings
     map_building.unlock();
 }
 
-#include <iostream>
 #include <algorithm>
 
 uint64_t ShortCode::fast_decode(uint32_t short_code) {
-//    std::cout << "short code: " << short_code << std::endl;
-
-    uint32_t head = std::upper_bound(ALL_CASES_OFFSET, ALL_CASES_OFFSET + 16, short_code) - ALL_CASES_OFFSET - 1;
-//    std::cout << head << std::endl;
-
-//    printf("%08X\n", AllCases::fetch()[head][ short_code - ALL_CASES_OFFSET[head] ]);
-
-    return ((uint64_t)head << 32) | AllCases::fetch()[head][short_code - ALL_CASES_OFFSET[head]];
+    auto offset = std::upper_bound(ALL_CASES_OFFSET, ALL_CASES_OFFSET + 16, short_code) - 1; // binary search
+    uint64_t head = offset - ALL_CASES_OFFSET; // head index
+    return (head << 32) | AllCases::fetch()[head][short_code - *offset]; // release common code
 }
 
 uint32_t ShortCode::fast_encode(uint64_t common_code) {
-//    printf("common code: %09lX\n", common_code);
-
-    uint32_t head = common_code >> 32;
-
-//    std::cout << "head = " << head << std::endl;
-
-    const auto &a = AllCases::fetch();
-
-    return std::lower_bound(a[head].begin(), a[head].end(), (uint32_t)common_code) - a[head].begin() + ALL_CASES_OFFSET[head];
-
+    auto head = common_code >> 32; // head index
+    const auto &ranges = AllCases::fetch()[head]; // available ranges
+    auto offset = std::lower_bound(ranges.begin(), ranges.end(), (uint32_t)common_code) - ranges.begin();
+    return ALL_CASES_OFFSET[head] + offset; // release short code
 }