summaryrefslogtreecommitdiff
path: root/llama.cpp/src/unicode-data.h
diff options
context:
space:
mode:
authorMitja Felicijan <mitja.felicijan@gmail.com>2026-02-12 20:57:17 +0100
committerMitja Felicijan <mitja.felicijan@gmail.com>2026-02-12 20:57:17 +0100
commitb333b06772c89d96aacb5490d6a219fba7c09cc6 (patch)
tree211df60083a5946baa2ed61d33d8121b7e251b06 /llama.cpp/src/unicode-data.h
downloadllmnpc-b333b06772c89d96aacb5490d6a219fba7c09cc6.tar.gz
Engage!
Diffstat (limited to 'llama.cpp/src/unicode-data.h')
-rw-r--r--llama.cpp/src/unicode-data.h20
1 files changed, 20 insertions, 0 deletions
diff --git a/llama.cpp/src/unicode-data.h b/llama.cpp/src/unicode-data.h
new file mode 100644
index 0000000..f6973eb
--- /dev/null
+++ b/llama.cpp/src/unicode-data.h
@@ -0,0 +1,20 @@
+#pragma once
+
+#include <cstdint>
+#include <vector>
+#include <unordered_map>
+#include <unordered_set>
+
+struct range_nfd {
+ uint32_t first;
+ uint32_t last;
+ uint32_t nfd;
+};
+
+static const uint32_t MAX_CODEPOINTS = 0x110000;
+
+extern const std::initializer_list<std::pair<uint32_t, uint16_t>> unicode_ranges_flags;
+extern const std::unordered_set<uint32_t> unicode_set_whitespace;
+extern const std::initializer_list<std::pair<uint32_t, uint32_t>> unicode_map_lowercase;
+extern const std::initializer_list<std::pair<uint32_t, uint32_t>> unicode_map_uppercase;
+extern const std::initializer_list<range_nfd> unicode_ranges_nfd;