From 961d2ec2f9075f30e9264ed2db6c394cfa5967f9 Mon Sep 17 00:00:00 2001 From: Jose Colon Rodriguez Date: Mon, 25 Mar 2024 20:05:36 -0400 Subject: CaseData --- build.zig | 64 ++++++++++++++- codegen/case_prop.zig | 136 ++++++++++++++++++++++++++++++ codegen/lower.zig | 58 +++++++++++++ codegen/numeric.zig | 45 +++++----- codegen/title.zig | 58 +++++++++++++ codegen/upper.zig | 58 +++++++++++++ src/CaseData.zig | 223 ++++++++++++++++++++++++++++++++++++++++++++++++++ 7 files changed, 613 insertions(+), 29 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 create mode 100644 src/CaseData.zig diff --git a/build.zig b/build.zig index 316292e..7272336 100644 --- a/build.zig +++ b/build.zig @@ -107,6 +107,46 @@ pub fn build(b: *std.Build) void { const run_num_gen_exe = b.addRunArtifact(num_gen_exe); const num_gen_out = run_num_gen_exe.addOutputFileArg("numeric.bin.z"); + // Letter case properties + const case_prop_gen_exe = b.addExecutable(.{ + .name = "case_prop", + .root_source_file = .{ .path = "codegen/case_prop.zig" }, + .target = b.host, + .optimize = .Debug, + }); + const run_case_prop_gen_exe = b.addRunArtifact(case_prop_gen_exe); + const case_prop_gen_out = run_case_prop_gen_exe.addOutputFileArg("case_prop.bin.z"); + + // Uppercase mappings + const upper_gen_exe = b.addExecutable(.{ + .name = "upper", + .root_source_file = .{ .path = "codegen/upper.zig" }, + .target = b.host, + .optimize = .Debug, + }); + const run_upper_gen_exe = b.addRunArtifact(upper_gen_exe); + const upper_gen_out = run_upper_gen_exe.addOutputFileArg("upper.bin.z"); + + // Lowercase mappings + const lower_gen_exe = b.addExecutable(.{ + .name = "lower", + .root_source_file = .{ .path = "codegen/lower.zig" }, + .target = b.host, + .optimize = .Debug, + }); + const run_lower_gen_exe = b.addRunArtifact(lower_gen_exe); + const lower_gen_out = run_lower_gen_exe.addOutputFileArg("lower.bin.z"); + + // Titlecase mappings + const title_gen_exe = b.addExecutable(.{ + .name = "title", + .root_source_file = .{ .path = "codegen/title.zig" }, + .target = b.host, + .optimize = .Debug, + }); + const run_title_gen_exe = b.addRunArtifact(title_gen_exe); + const title_gen_out = run_title_gen_exe.addOutputFileArg("title.bin.z"); + // Modules we provide // Code points const code_point = b.addModule("code_point", .{ @@ -221,7 +261,7 @@ pub fn build(b: *std.Build) void { }); gencat_data.addAnonymousImport("gencat", .{ .root_source_file = gencat_gen_out }); - // Case + // Case folding const fold_data = b.createModule(.{ .root_source_file = .{ .path = "src/FoldData.zig" }, .target = target, @@ -246,6 +286,18 @@ pub fn build(b: *std.Build) void { }); num_data.addAnonymousImport("numeric", .{ .root_source_file = num_gen_out }); + // Letter case + const case_data = b.createModule(.{ + .root_source_file = .{ .path = "src/CaseData.zig" }, + .target = target, + .optimize = optimize, + }); + case_data.addImport("code_point", code_point); + case_data.addAnonymousImport("case_prop", .{ .root_source_file = case_prop_gen_out }); + case_data.addAnonymousImport("upper", .{ .root_source_file = upper_gen_out }); + case_data.addAnonymousImport("lower", .{ .root_source_file = lower_gen_out }); + case_data.addAnonymousImport("title", .{ .root_source_file = title_gen_out }); + // Benchmark rig const exe = b.addExecutable(.{ .name = "zg", @@ -274,12 +326,12 @@ pub fn build(b: *std.Build) void { // Tests const exe_unit_tests = b.addTest(.{ - .root_source_file = .{ .path = "src/NumericData.zig" }, + .root_source_file = .{ .path = "src/CaseData.zig" }, .target = target, .optimize = optimize, }); // exe_unit_tests.root_module.addImport("ascii", ascii); - // exe_unit_tests.root_module.addImport("code_point", code_point); + exe_unit_tests.root_module.addImport("code_point", code_point); // exe_unit_tests.root_module.addImport("GraphemeData", grapheme_data); // exe_unit_tests.root_module.addImport("grapheme", grapheme); // exe_unit_tests.root_module.addImport("ziglyph", ziglyph.module("ziglyph")); @@ -288,7 +340,11 @@ pub fn build(b: *std.Build) void { // exe_unit_tests.root_module.addImport("NormData", norm_data); // exe_unit_tests.root_module.addImport("Normalize", norm); // exe_unit_tests.root_module.addImport("FoldData", fold_data); - exe_unit_tests.root_module.addAnonymousImport("numeric", .{ .root_source_file = num_gen_out }); + // exe_unit_tests.root_module.addAnonymousImport("numeric", .{ .root_source_file = num_gen_out }); + exe_unit_tests.root_module.addAnonymousImport("case_prop", .{ .root_source_file = case_prop_gen_out }); + exe_unit_tests.root_module.addAnonymousImport("upper", .{ .root_source_file = upper_gen_out }); + exe_unit_tests.root_module.addAnonymousImport("lower", .{ .root_source_file = lower_gen_out }); + exe_unit_tests.root_module.addAnonymousImport("title", .{ .root_source_file = title_gen_out }); // exe_unit_tests.filter = "nfd !ASCII"; const run_exe_unit_tests = b.addRunArtifact(exe_unit_tests); 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(); +} diff --git a/src/CaseData.zig b/src/CaseData.zig new file mode 100644 index 0000000..38830e3 --- /dev/null +++ b/src/CaseData.zig @@ -0,0 +1,223 @@ +const std = @import("std"); +const builtin = @import("builtin"); +const compress = std.compress; +const mem = std.mem; +const testing = std.testing; +const unicode = std.unicode; + +const CodePointIterator = @import("code_point").Iterator; + +allocator: mem.Allocator, +case_map: [][3]u21, +prop_s1: []u16 = undefined, +prop_s2: []u8 = undefined, + +const Self = @This(); + +pub fn init(allocator: mem.Allocator) !Self { + const decompressor = compress.deflate.decompressor; + const endian = builtin.cpu.arch.endian(); + + var self = Self{ + .allocator = allocator, + .case_map = try allocator.alloc([3]u21, 0x110000), + }; + errdefer allocator.free(self.case_map); + + for (0..0x110000) |i| { + const cp: u21 = @intCast(i); + self.case_map[cp] = .{ cp, cp, cp }; + } + + // Uppercase + const upper_bytes = @embedFile("upper"); + var upper_fbs = std.io.fixedBufferStream(upper_bytes); + var upper_decomp = try decompressor(allocator, upper_fbs.reader(), null); + defer upper_decomp.deinit(); + var upper_reader = upper_decomp.reader(); + + while (true) { + const cp = try upper_reader.readInt(u24, endian); + if (cp == 0) break; + self.case_map[cp][0] = @intCast(try upper_reader.readInt(u24, endian)); + } + + // Lowercase + const lower_bytes = @embedFile("lower"); + var lower_fbs = std.io.fixedBufferStream(lower_bytes); + var lower_decomp = try decompressor(allocator, lower_fbs.reader(), null); + defer lower_decomp.deinit(); + var lower_reader = lower_decomp.reader(); + + while (true) { + const cp = try lower_reader.readInt(u24, endian); + if (cp == 0) break; + self.case_map[cp][1] = @intCast(try lower_reader.readInt(u24, endian)); + } + + // Titlercase + const title_bytes = @embedFile("title"); + var title_fbs = std.io.fixedBufferStream(title_bytes); + var title_decomp = try decompressor(allocator, title_fbs.reader(), null); + defer title_decomp.deinit(); + var title_reader = title_decomp.reader(); + + while (true) { + const cp = try title_reader.readInt(u24, endian); + if (cp == 0) break; + self.case_map[cp][2] = @intCast(try title_reader.readInt(u24, endian)); + } + + // Case properties + const cp_bytes = @embedFile("case_prop"); + var cp_fbs = std.io.fixedBufferStream(cp_bytes); + var cp_decomp = try decompressor(allocator, cp_fbs.reader(), null); + defer cp_decomp.deinit(); + var cp_reader = cp_decomp.reader(); + + const stage_1_len: u16 = try cp_reader.readInt(u16, endian); + self.prop_s1 = try allocator.alloc(u16, stage_1_len); + errdefer allocator.free(self.prop_s1); + for (0..stage_1_len) |i| self.prop_s1[i] = try cp_reader.readInt(u16, endian); + + const stage_2_len: u16 = try cp_reader.readInt(u16, endian); + self.prop_s2 = try allocator.alloc(u8, stage_2_len); + errdefer allocator.free(self.prop_s2); + _ = try cp_reader.readAll(self.prop_s2); + + return self; +} + +pub fn deinit(self: *Self) void { + self.allocator.free(self.case_map); + self.allocator.free(self.prop_s1); + self.allocator.free(self.prop_s2); +} + +// Returns true if `cp` is either upper, lower, or title case. +pub inline fn isCased(self: Self, cp: u21) bool { + return self.prop_s2[self.prop_s1[cp >> 8] + (cp & 0xff)] & 4 == 4; +} + +// Returns true if `cp` is uppercase. +pub fn isUpper(self: Self, cp: u21) bool { + if (!self.isCased(cp)) return true; + return self.prop_s2[self.prop_s1[cp >> 8] + (cp & 0xff)] & 2 == 2; +} + +/// Returns true if `str` is all uppercase. +pub fn isUpperStr(self: Self, str: []const u8) bool { + var iter = CodePointIterator{ .bytes = str }; + + return while (iter.next()) |cp| { + if (!self.isUpper(cp.code)) break false; + } else true; +} + +test "isUpperStr" { + var cd = try init(testing.allocator); + defer cd.deinit(); + + try testing.expect(cd.isUpperStr("HELLO, WORLD 2112!")); + try testing.expect(!cd.isUpperStr("hello, world 2112!")); + try testing.expect(!cd.isUpperStr("Hello, World 2112!")); +} + +/// Returns a new string with all letters in uppercase. +/// Caller must free returned bytes with `allocator`. +pub fn toUpperStr( + self: Self, + allocator: mem.Allocator, + str: []const u8, +) ![]u8 { + var bytes = std.ArrayList(u8).init(allocator); + defer bytes.deinit(); + + var iter = CodePointIterator{ .bytes = str }; + var buf: [4]u8 = undefined; + + while (iter.next()) |cp| { + const len = try unicode.utf8Encode(self.toUpper(cp.code), &buf); + try bytes.appendSlice(buf[0..len]); + } + + return try bytes.toOwnedSlice(); +} + +test "toUpperStr" { + var cd = try init(testing.allocator); + defer cd.deinit(); + + const uppered = try cd.toUpperStr(testing.allocator, "Hello, World 2112!"); + defer testing.allocator.free(uppered); + try testing.expectEqualStrings("HELLO, WORLD 2112!", uppered); +} + +/// Returns uppercase mapping for `cp`. +pub inline fn toUpper(self: Self, cp: u21) u21 { + return self.case_map[cp][0]; +} + +// Returns true if `cp` is lowercase. +pub fn isLower(self: Self, cp: u21) bool { + if (!self.isCased(cp)) return true; + return self.prop_s2[self.prop_s1[cp >> 8] + (cp & 0xff)] & 1 == 1; +} + +/// Returns lowercase mapping for `cp`. +pub inline fn toLower(self: Self, cp: u21) u21 { + return self.case_map[cp][1]; +} + +/// Returns true if `str` is all lowercase. +pub fn isLowerStr(self: Self, str: []const u8) bool { + var iter = CodePointIterator{ .bytes = str }; + + return while (iter.next()) |cp| { + if (!self.isLower(cp.code)) break false; + } else true; +} + +test "isLowerStr" { + var cd = try init(testing.allocator); + defer cd.deinit(); + + try testing.expect(cd.isLowerStr("hello, world 2112!")); + try testing.expect(!cd.isLowerStr("HELLO, WORLD 2112!")); + try testing.expect(!cd.isLowerStr("Hello, World 2112!")); +} + +/// Returns a new string with all letters in lowercase. +/// Caller must free returned bytes with `allocator`. +pub fn toLowerStr( + self: Self, + allocator: mem.Allocator, + str: []const u8, +) ![]u8 { + var bytes = std.ArrayList(u8).init(allocator); + defer bytes.deinit(); + + var iter = CodePointIterator{ .bytes = str }; + var buf: [4]u8 = undefined; + + while (iter.next()) |cp| { + const len = try unicode.utf8Encode(self.toLower(cp.code), &buf); + try bytes.appendSlice(buf[0..len]); + } + + return try bytes.toOwnedSlice(); +} + +test "toLowerStr" { + var cd = try init(testing.allocator); + defer cd.deinit(); + + const lowered = try cd.toLowerStr(testing.allocator, "Hello, World 2112!"); + defer testing.allocator.free(lowered); + try testing.expectEqualStrings("hello, world 2112!", lowered); +} + +/// Returns titlecase mapping for `cp`. +pub inline fn toTitle(self: Self, cp: u21) u21 { + return self.case_map[cp][2]; +} -- cgit v1.2.3