Engage!

1 files changed, 64 insertions, 0 deletions

diff --git a/llama.cpp/common/unicode.cpp b/llama.cpp/common/unicode.cpp
new file mode 100644
index 0000000..56ab0f4
--- /dev/null
+++ b/llama.cpp/common/unicode.cpp
@@ -0,0 +1,64 @@
+#include "unicode.h"
+
+// implementation adopted from src/unicode.cpp
+
+size_t utf8_sequence_length(unsigned char first_byte) {
+    const size_t lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 4 };
+    uint8_t highbits = static_cast<uint8_t>(first_byte) >> 4;
+    return lookup[highbits];
+}
+
+utf8_parse_result parse_utf8_codepoint(std::string_view input, size_t offset) {
+    if (offset >= input.size()) {
+        return utf8_parse_result(utf8_parse_result::INCOMPLETE);
+    }
+
+    // ASCII fast path
+    if (!(input[offset] & 0x80)) {
+        return utf8_parse_result(utf8_parse_result::SUCCESS, input[offset], 1);
+    }
+
+    // Invalid: continuation byte as first byte
+    if (!(input[offset] & 0x40)) {
+        return utf8_parse_result(utf8_parse_result::INVALID);
+    }
+
+    // 2-byte sequence
+    if (!(input[offset] & 0x20)) {
+        if (offset + 1 >= input.size()) {
+            return utf8_parse_result(utf8_parse_result::INCOMPLETE);
+        }
+        if ((input[offset + 1] & 0xc0) != 0x80) {
+            return utf8_parse_result(utf8_parse_result::INVALID);
+        }
+        auto result = ((input[offset] & 0x1f) << 6) | (input[offset + 1] & 0x3f);
+        return utf8_parse_result(utf8_parse_result::SUCCESS, result, 2);
+    }
+
+    // 3-byte sequence
+    if (!(input[offset] & 0x10)) {
+        if (offset + 2 >= input.size()) {
+            return utf8_parse_result(utf8_parse_result::INCOMPLETE);
+        }
+        if ((input[offset + 1] & 0xc0) != 0x80 || (input[offset + 2] & 0xc0) != 0x80) {
+            return utf8_parse_result(utf8_parse_result::INVALID);
+        }
+        auto result = ((input[offset] & 0x0f) << 12) | ((input[offset + 1] & 0x3f) << 6) | (input[offset + 2] & 0x3f);
+        return utf8_parse_result(utf8_parse_result::SUCCESS, result, 3);
+    }
+
+    // 4-byte sequence
+    if (!(input[offset] & 0x08)) {
+        if (offset + 3 >= input.size()) {
+            return utf8_parse_result(utf8_parse_result::INCOMPLETE);
+        }
+        if ((input[offset + 1] & 0xc0) != 0x80 || (input[offset + 2] & 0xc0) != 0x80 || (input[offset + 3] & 0xc0) != 0x80) {
+            return utf8_parse_result(utf8_parse_result::INVALID);
+        }
+        auto result = ((input[offset] & 0x07) << 18) | ((input[offset + 1] & 0x3f) << 12) | ((input[offset + 2] & 0x3f) << 6) | (input[offset + 3] & 0x3f);
+        return utf8_parse_result(utf8_parse_result::SUCCESS, result, 4);
+    }
+
+    // Invalid first byte
+    return utf8_parse_result(utf8_parse_result::INVALID);
+}