From 961d2ec2f9075f30e9264ed2db6c394cfa5967f9 Mon Sep 17 00:00:00 2001
From: Jose Colon Rodriguez
Date: Mon, 25 Mar 2024 20:05:36 -0400
Subject: CaseData

---
 codegen/case_prop.zig | 136 ++++++++++++++++++++++++++++++++++++++++++++++++++
 codegen/lower.zig     |  58 +++++++++++++++++++++
 codegen/numeric.zig   |  45 ++++++++---------
 codegen/title.zig     |  58 +++++++++++++++++++++
 codegen/upper.zig     |  58 +++++++++++++++++++++
 5 files changed, 330 insertions(+), 25 deletions(-)
 create mode 100644 codegen/case_prop.zig
 create mode 100644 codegen/lower.zig
 create mode 100644 codegen/title.zig
 create mode 100644 codegen/upper.zig

(limited to 'codegen')

diff --git a/codegen/case_prop.zig b/codegen/case_prop.zig
new file mode 100644
index 0000000..ce7ee0d
--- /dev/null
+++ b/codegen/case_prop.zig
@@ -0,0 +1,136 @@
+const std = @import("std");
+const builtin = @import("builtin");
+const mem = std.mem;
+
+const block_size = 256;
+const Block = [block_size]u8;
+
+const BlockMap = std.HashMap(
+    Block,
+    u16,
+    struct {
+        pub fn hash(_: @This(), k: Block) u64 {
+            var hasher = std.hash.Wyhash.init(0);
+            std.hash.autoHashStrat(&hasher, k, .DeepRecursive);
+            return hasher.final();
+        }
+
+        pub fn eql(_: @This(), a: Block, b: Block) bool {
+            return mem.eql(u8, &a, &b);
+        }
+    },
+    std.hash_map.default_max_load_percentage,
+);
+
+pub fn main() !void {
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    var flat_map = std.AutoHashMap(u21, u8).init(allocator);
+    defer flat_map.deinit();
+
+    var line_buf: [4096]u8 = undefined;
+
+    // Process DerivedCoreProperties.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/DerivedCoreProperties.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
+
+    while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+        if (line.len == 0 or line[0] == '#') continue;
+        const no_comment = if (mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line;
+
+        var field_iter = mem.tokenizeAny(u8, no_comment, "; ");
+        var current_code: [2]u21 = undefined;
+
+        var i: usize = 0;
+        while (field_iter.next()) |field| : (i += 1) {
+            switch (i) {
+                0 => {
+                    // Code point(s)
+                    if (mem.indexOf(u8, field, "..")) |dots| {
+                        current_code = .{
+                            try std.fmt.parseInt(u21, field[0..dots], 16),
+                            try std.fmt.parseInt(u21, field[dots + 2 ..], 16),
+                        };
+                    } else {
+                        const code = try std.fmt.parseInt(u21, field, 16);
+                        current_code = .{ code, code };
+                    }
+                },
+                1 => {
+                    // Props
+                    var bit: u8 = 0;
+
+                    if (mem.eql(u8, field, "Lowercase")) bit = 1;
+                    if (mem.eql(u8, field, "Uppercase")) bit = 2;
+                    if (mem.eql(u8, field, "Cased")) bit = 4;
+
+                    if (bit != 0) {
+                        for (current_code[0]..current_code[1] + 1) |cp| {
+                            const gop = try flat_map.getOrPut(@intCast(cp));
+                            if (!gop.found_existing) gop.value_ptr.* = 0;
+                            gop.value_ptr.* |= bit;
+                        }
+                    }
+                },
+                else => {},
+            }
+        }
+    }
+
+    var blocks_map = BlockMap.init(allocator);
+    defer blocks_map.deinit();
+
+    var stage1 = std.ArrayList(u16).init(allocator);
+    defer stage1.deinit();
+
+    var stage2 = std.ArrayList(u8).init(allocator);
+    defer stage2.deinit();
+
+    var block: Block = [_]u8{0} ** block_size;
+    var block_len: u16 = 0;
+
+    for (0..0x110000) |i| {
+        const cp: u21 = @intCast(i);
+        const prop = flat_map.get(cp) orelse 0;
+
+        // Process block
+        block[block_len] = prop;
+        block_len += 1;
+
+        if (block_len < block_size and cp != 0x10ffff) continue;
+
+        const gop = try blocks_map.getOrPut(block);
+        if (!gop.found_existing) {
+            gop.value_ptr.* = @intCast(stage2.items.len);
+            try stage2.appendSlice(&block);
+        }
+
+        try stage1.append(gop.value_ptr.*);
+        block_len = 0;
+    }
+
+    var args_iter = try std.process.argsWithAllocator(allocator);
+    defer args_iter.deinit();
+    _ = args_iter.skip();
+    const output_path = args_iter.next() orelse @panic("No output file arg!");
+
+    const compressor = std.compress.deflate.compressor;
+    var out_file = try std.fs.cwd().createFile(output_path, .{});
+    defer out_file.close();
+    var out_comp = try compressor(allocator, out_file.writer(), .{ .level = .best_compression });
+    defer out_comp.deinit();
+    const writer = out_comp.writer();
+
+    const endian = builtin.cpu.arch.endian();
+    try writer.writeInt(u16, @intCast(stage1.items.len), endian);
+    for (stage1.items) |i| try writer.writeInt(u16, i, endian);
+
+    try writer.writeInt(u16, @intCast(stage2.items.len), endian);
+    try writer.writeAll(stage2.items);
+
+    try out_comp.flush();
+}
diff --git a/codegen/lower.zig b/codegen/lower.zig
new file mode 100644
index 0000000..5a1f1b3
--- /dev/null
+++ b/codegen/lower.zig
@@ -0,0 +1,58 @@
+const std = @import("std");
+const builtin = @import("builtin");
+
+pub fn main() !void {
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    // Process UnicodeData.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/UnicodeData.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
+
+    var args_iter = try std.process.argsWithAllocator(allocator);
+    defer args_iter.deinit();
+    _ = args_iter.skip();
+    const output_path = args_iter.next() orelse @panic("No output file arg!");
+
+    const compressor = std.compress.deflate.compressor;
+    var out_file = try std.fs.cwd().createFile(output_path, .{});
+    defer out_file.close();
+    var out_comp = try compressor(allocator, out_file.writer(), .{ .level = .best_compression });
+    defer out_comp.deinit();
+    const writer = out_comp.writer();
+
+    const endian = builtin.cpu.arch.endian();
+    var line_buf: [4096]u8 = undefined;
+
+    lines: while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+        if (line.len == 0) continue;
+
+        var field_iter = std.mem.splitScalar(u8, line, ';');
+        var cps: [2]u24 = undefined;
+
+        var i: usize = 0;
+        while (field_iter.next()) |field| : (i += 1) {
+            switch (i) {
+                0 => cps[0] = try std.fmt.parseInt(u24, field, 16),
+
+                13 => {
+                    // Simple lowercase mapping
+                    if (field.len == 0) continue :lines;
+                    cps[1] = try std.fmt.parseInt(u24, field, 16);
+                },
+
+                2 => if (line[0] == '<') continue :lines,
+
+                else => {},
+            }
+        }
+
+        for (&cps) |cp| try writer.writeInt(u24, cp, endian);
+    }
+
+    try writer.writeInt(u24, 0, endian);
+    try out_comp.flush();
+}
diff --git a/codegen/numeric.zig b/codegen/numeric.zig
index 80fc7c2..ad8490c 100644
--- a/codegen/numeric.zig
+++ b/codegen/numeric.zig
@@ -1,5 +1,6 @@
 const std = @import("std");
 const builtin = @import("builtin");
+const mem = std.mem;
 
 const block_size = 256;
 const Block = [block_size]u8;
@@ -15,7 +16,7 @@ const BlockMap = std.HashMap(
         }
 
         pub fn eql(_: @This(), a: Block, b: Block) bool {
-            return std.mem.eql(u8, &a, &b);
+            return mem.eql(u8, &a, &b);
         }
     },
     std.hash_map.default_max_load_percentage,
@@ -31,17 +32,17 @@ pub fn main() !void {
 
     var line_buf: [4096]u8 = undefined;
 
-    // Process DerivedCombiningClass.txt
-    var num_file = try std.fs.cwd().openFile("data/unicode/extracted/DerivedNumericType.txt", .{});
-    defer num_file.close();
-    var num_buf = std.io.bufferedReader(num_file.reader());
-    const num_reader = num_buf.reader();
+    // Process DerivedNumericType.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/extracted/DerivedNumericType.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
 
-    while (try num_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+    while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
         if (line.len == 0 or line[0] == '#') continue;
-        const no_comment = if (std.mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line;
+        const no_comment = if (mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line;
 
-        var field_iter = std.mem.tokenizeAny(u8, no_comment, "; ");
+        var field_iter = mem.tokenizeAny(u8, no_comment, "; ");
         var current_code: [2]u21 = undefined;
 
         var i: usize = 0;
@@ -49,7 +50,7 @@ pub fn main() !void {
             switch (i) {
                 0 => {
                     // Code point(s)
-                    if (std.mem.indexOf(u8, field, "..")) |dots| {
+                    if (mem.indexOf(u8, field, "..")) |dots| {
                         current_code = .{
                             try std.fmt.parseInt(u21, field[0..dots], 16),
                             try std.fmt.parseInt(u21, field[dots + 2 ..], 16),
@@ -61,25 +62,19 @@ pub fn main() !void {
                 },
                 1 => {
                     // Numeric type
-                    if (std.mem.eql(u8, field, "Numeric")) {
-                        for (current_code[0]..current_code[1] + 1) |cp| {
-                            const gop = try flat_map.getOrPut(@intCast(cp));
-                            if (!gop.found_existing) gop.value_ptr.* = 0;
-                            gop.value_ptr.* |= 1;
-                        }
-                    } else if (std.mem.eql(u8, field, "Digit")) {
-                        for (current_code[0]..current_code[1] + 1) |cp| {
-                            const gop = try flat_map.getOrPut(@intCast(cp));
-                            if (!gop.found_existing) gop.value_ptr.* = 0;
-                            gop.value_ptr.* |= 2;
-                        }
-                    } else if (std.mem.eql(u8, field, "Decimal")) {
+                    var bit: u8 = 0;
+
+                    if (mem.eql(u8, field, "Numeric")) bit = 1;
+                    if (mem.eql(u8, field, "Digit")) bit = 2;
+                    if (mem.eql(u8, field, "Decimal")) bit = 4;
+
+                    if (bit != 0) {
                         for (current_code[0]..current_code[1] + 1) |cp| {
                             const gop = try flat_map.getOrPut(@intCast(cp));
                             if (!gop.found_existing) gop.value_ptr.* = 0;
-                            gop.value_ptr.* |= 4;
+                            gop.value_ptr.* |= bit;
                         }
-                    } else continue;
+                    }
                 },
                 else => {},
             }
diff --git a/codegen/title.zig b/codegen/title.zig
new file mode 100644
index 0000000..653b812
--- /dev/null
+++ b/codegen/title.zig
@@ -0,0 +1,58 @@
+const std = @import("std");
+const builtin = @import("builtin");
+
+pub fn main() !void {
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    // Process UnicodeData.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/UnicodeData.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
+
+    var args_iter = try std.process.argsWithAllocator(allocator);
+    defer args_iter.deinit();
+    _ = args_iter.skip();
+    const output_path = args_iter.next() orelse @panic("No output file arg!");
+
+    const compressor = std.compress.deflate.compressor;
+    var out_file = try std.fs.cwd().createFile(output_path, .{});
+    defer out_file.close();
+    var out_comp = try compressor(allocator, out_file.writer(), .{ .level = .best_compression });
+    defer out_comp.deinit();
+    const writer = out_comp.writer();
+
+    const endian = builtin.cpu.arch.endian();
+    var line_buf: [4096]u8 = undefined;
+
+    lines: while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+        if (line.len == 0) continue;
+
+        var field_iter = std.mem.splitScalar(u8, line, ';');
+        var cps: [2]u24 = undefined;
+
+        var i: usize = 0;
+        while (field_iter.next()) |field| : (i += 1) {
+            switch (i) {
+                0 => cps[0] = try std.fmt.parseInt(u24, field, 16),
+
+                14 => {
+                    // Simple titlecase mapping
+                    if (field.len == 0) continue :lines;
+                    cps[1] = try std.fmt.parseInt(u24, field, 16);
+                },
+
+                2 => if (line[0] == '<') continue :lines,
+
+                else => {},
+            }
+        }
+
+        for (&cps) |cp| try writer.writeInt(u24, cp, endian);
+    }
+
+    try writer.writeInt(u24, 0, endian);
+    try out_comp.flush();
+}
diff --git a/codegen/upper.zig b/codegen/upper.zig
new file mode 100644
index 0000000..6fea608
--- /dev/null
+++ b/codegen/upper.zig
@@ -0,0 +1,58 @@
+const std = @import("std");
+const builtin = @import("builtin");
+
+pub fn main() !void {
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    // Process UnicodeData.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/UnicodeData.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
+
+    var args_iter = try std.process.argsWithAllocator(allocator);
+    defer args_iter.deinit();
+    _ = args_iter.skip();
+    const output_path = args_iter.next() orelse @panic("No output file arg!");
+
+    const compressor = std.compress.deflate.compressor;
+    var out_file = try std.fs.cwd().createFile(output_path, .{});
+    defer out_file.close();
+    var out_comp = try compressor(allocator, out_file.writer(), .{ .level = .best_compression });
+    defer out_comp.deinit();
+    const writer = out_comp.writer();
+
+    const endian = builtin.cpu.arch.endian();
+    var line_buf: [4096]u8 = undefined;
+
+    lines: while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+        if (line.len == 0) continue;
+
+        var field_iter = std.mem.splitScalar(u8, line, ';');
+        var cps: [2]u24 = undefined;
+
+        var i: usize = 0;
+        while (field_iter.next()) |field| : (i += 1) {
+            switch (i) {
+                0 => cps[0] = try std.fmt.parseInt(u24, field, 16),
+
+                12 => {
+                    // Simple uppercase mapping
+                    if (field.len == 0) continue :lines;
+                    cps[1] = try std.fmt.parseInt(u24, field, 16);
+                },
+
+                2 => if (line[0] == '<') continue :lines,
+
+                else => {},
+            }
+        }
+
+        for (&cps) |cp| try writer.writeInt(u24, cp, endian);
+    }
+
+    try writer.writeInt(u24, 0, endian);
+    try out_comp.flush();
+}
-- 
cgit v1.2.3