more work on the huffman

This commit is contained in:
Andre Henriques 2023-06-13 20:40:13 +01:00
parent 3e9ddba3a3
commit 99d8ce3484

View File

@ -1,4 +1,5 @@
const std = @import("std"); const std = @import("std");
const mem = std.mem;
const stdout = std.io.getStdOut().writer(); const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer(); const stderr = std.io.getStdErr().writer();
@ -20,6 +21,36 @@ fn usage() void {
exit("", .{}, 1); exit("", .{}, 1);
} }
fn BitWalkerUint(T: anytype) type {
var typeInfo = @typeInfo(T);
if (typeInfo == std.builtin.Type.Int) {
@compileError("This needs to be a int");
}
if (typeInfo.Int.is_signed) {
@compileError("The integer needs to be unsigned");
}
var size = typeInfo.Int.bits;
return struct {
t: T,
// TODO this is probably wrong
in_byte_position: u8,
const Self = @This();
fn init(value: T) Self {
return Self {
.t = value,
.in_byte_position = 0,
};
}
};
}
const BitWalker = struct { const BitWalker = struct {
const Self = @This(); const Self = @This();
@ -62,10 +93,10 @@ const BitWalker = struct {
var result = @shrExact(byte & old_mask, self.in_byte_position) + @shlExact(next_byte & new_mask, @intCast(u3, 8 - @as(u4, self.in_byte_position))); var result = @shrExact(byte & old_mask, self.in_byte_position) + @shlExact(next_byte & new_mask, @intCast(u3, 8 - @as(u4, self.in_byte_position)));
print("mask: {b}, new_mask: {b}", .{ old_mask, new_mask }); //print("mask: {b}, new_mask: {b}", .{ old_mask, new_mask });
print("here {b} {b}", .{ byte, old_mask }); //print("here {b} {b}", .{ byte, old_mask });
print("here_new {b} {b}", .{ next_byte, new_mask }); //print("here_new {b} {b}", .{ next_byte, new_mask });
print("result {}", .{result}); //print("result {}", .{result});
self.position += 1; self.position += 1;
self.in_byte_position = new_byte_pos; self.in_byte_position = new_byte_pos;
@ -100,11 +131,46 @@ fn HuffmanGraph(valueType: type) type {
const Node = struct { const Node = struct {
left: ?Node, left: ?Node,
right: ?Node, right: ?Node,
value: valueType, value: ?valueType,
fn init() Node {}
allocator: mem.Allocator,
fn init(allocator: mem.Allocator) !*Node {
var node = try allocator.create(Node);
node.left = null;
node.right = null;
node.value = null;
node.allocator = allocator;
return node;
}
fn deinit(self: *Node) void {
if (self.left) |left| {
left.deinit();
}
if (self.right) |right| {
right.deinit();
}
self.allocator.destory(self);
}
}; };
head: Node, root: Node,
const Self = @This();
fn init(allocator: mem.Allocator) !Self {
var root = try Node.init(allocator);
return Self {
root,
};
}
fn deinit(self: *Self) void {
self.root.deinit();
}
}; };
} }
@ -114,25 +180,49 @@ const DynamicDecoder = struct {
len_to_read: usize, len_to_read: usize,
codes: [19]u3, codes: [19]u3,
walker: *BitWalker, walker: *BitWalker,
allocator: mem.Allocator,
fn init(walker: *BitWalker, len_to_read: usize) !Self { fn init(walker: *BitWalker, len_to_read: usize, allocator: mem.Allocator) !Self {
const list: [19]u8 = .{ 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 }; const list: [19]u8 = .{ 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 };
var codes: [19]u3 = .{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; var lenList: [19]u3 = .{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
var bl_count: [7]u8 = .{0, 0, 0, 0, 0, 0, 0};
for (0..(len_to_read + 4)) |i| { for (0..(len_to_read + 4)) |i| {
print("{} {}", .{ walker.position, walker.in_byte_position });
var data: u3 = @intCast(u3, try walker.walk(3)); var data: u3 = @intCast(u3, try walker.walk(3));
lenList[i] = data;
if (data == 0) { if (data == 0) {
continue; continue;
} }
print("{} - {}", .{ list[i], data }); bl_count[data - 1] += 1;
codes[i] = data; }
var bl_count: []u8 = allocator.allocator(u8, max - 1);
var next_codes: [19]u3 = .{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
var code = 0;
bl_count[0] = 0;
for (1..max) |bits| {
code = (code + bl_count[bits-1]) << 1;
next_code[bits] = code;
}
var codes: [19]u3 = .{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
for (0..19) |n| {
var len = lenList[n];
if (len != 0) {
codes = next_code[len];
next_code[len] += 1;
}
} }
return .{ return .{
.len_to_read = len_to_read, .len_to_read = len_to_read,
.codes = codes, .codes = codes,
.walker = walker, .walker = walker,
.allocator = allocator,
}; };
} }
}; };
@ -230,7 +320,7 @@ const ZipFileHeader = struct {
var number_of_literal_codes = try bitw.walk(5); var number_of_literal_codes = try bitw.walk(5);
var number_of_dist_codes = try bitw.walk(5); var number_of_dist_codes = try bitw.walk(5);
var number_of_length_codes = try bitw.walk(5); var number_of_length_codes = try bitw.walk(4);
print("number of literal codes: {}", .{number_of_literal_codes}); print("number of literal codes: {}", .{number_of_literal_codes});
print("number of dist codes: {}", .{number_of_dist_codes}); print("number of dist codes: {}", .{number_of_dist_codes});