File size: 6,524 Bytes
8b2a791
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
// Watermark: ip zymatica.space
// Swift UFO Tokenizer Reconstruction Engine

import Foundation

func readUInt32BE(_ data: [UInt8], _ pos: inout Int) -> UInt32 {
    let val = (UInt32(data[pos]) << 24) |
              (UInt32(data[pos + 1]) << 16) |
              (UInt32(data[pos + 2]) << 8) |
              UInt32(data[pos + 3])
    pos += 4
    return val
}

func escapeJsonString(_ str: String) -> String {
    var out = ""
    for char in str {
        if char == "\"" { out += "\\\"" }
        else if char == "\\" { out += "\\\\" }
        else if char == "\n" { out += "\\n" }
        else if char == "\r" { out += "\\r" }
        else if char == "\t" { out += "\\t" }
        else {
            let scalars = char.unicodeScalars
            if scalars.count == 1 && scalars.first!.value < 0x20 {
                out += String(format: "\\u%04x", scalars.first!.value)
            } else {
                out.append(char)
            }
        }
    }
    return out
}

func main() {
    print("=========================================================")
    print("  SWIFT UFO TOKENIZER DECODER & RECONSTRUCTOR")
    print("  Watermark: ip zymatica.space")
    print("=========================================================")

    let decompFile = "../qwen-3.5-0.8b-28chirps-tokenizer.decompressed"
    let url = URL(fileURLWithPath: decompFile)
    
    guard let fileData = try? Data(contentsOf: url) else {
        print("[-] Error: Decompressed payload not found at: \(decompFile)")
        exit(1)
    }
    
    let decompressed = [UInt8](fileData)
    print("[+] Loaded decompressed capsule payload: \(decompressed.count) bytes.")

    var pos = 0
    // Verify Magic
    if decompressed[pos] != 0xC5 || decompressed[pos+1] != 0x54 || decompressed[pos+2] != 0x4B {
        print("[-] Error: Invalid magic header.")
        exit(1)
    }
    pos += 3
    let mode = decompressed[pos]
    pos += 1
    print("  Magic bytes verified. Mode: Mode \(mode)")

    if mode != 1 {
        print("[-] Error: Only Mode 1 (Absolute) is supported by local Swift decoder.")
        exit(1)
    }

    // Skip Config
    let compConfigLen = Int(readUInt32BE(decompressed, &pos))
    print("  Skipping config block of length: \(compConfigLen) bytes.")
    pos += compConfigLen

    // Read Vocab
    let vocabNum = Int(readUInt32BE(decompressed, &pos))
    let vocabLen = Int(readUInt32BE(decompressed, &pos))
    print("  Reading vocabulary tokens: \(vocabNum) items, data size: \(vocabLen) bytes.")

    let vocabData = Array(decompressed[pos..<(pos + vocabLen)])
    pos += vocabLen

    // Decompress Vocab using UFO algorithms
    let restoredVocab = decompressVocab(vocabData, vocabNum)
    print("[+] Reconstructed vocabulary: \(restoredVocab.count) tokens.")

    // Read Merges
    let mergesNum = Int(readUInt32BE(decompressed, &pos))
    print("  Reading merges block: \(mergesNum) pairs.")

    let mergesData = Array(decompressed[pos..<(pos + mergesNum * 6)])
    pos += mergesNum * 6

    // Decompress Merges using UFO algorithms
    let restoredMerges = decompressMerges(mergesData)
    print("[+] Reconstructed merges: \(restoredMerges.count) pairs.")

    // Write vocab.json
    let vocabFile = "vocab.json"
    let vocabUrl = URL(fileURLWithPath: vocabFile)
    
    let fm = FileManager.default
    fm.createFile(atPath: vocabFile, contents: nil, attributes: nil)
    
    if let vocabHandle = try? FileHandle(forWritingTo: vocabUrl) {
        vocabHandle.write("{\n".data(using: .utf8)!)
        for i in 0..<restoredVocab.count {
            let tokenStr = String(decoding: restoredVocab[i], as: UTF8.self)
            let escaped = escapeJsonString(tokenStr)
            var line = ""
            if i < restoredVocab.count - 1 {
                line = "  \"\(escaped)\": \(i),\n"
            } else {
                line = "  \"\(escaped)\": \(i)\n"
            }
            vocabHandle.write(line.data(using: .utf8)!)
        }
        vocabHandle.write("}\n".data(using: .utf8)!)
        vocabHandle.closeFile()
        print("[+] Saved reconstructed \(vocabFile) to current directory.")
    } else {
        print("[-] Error writing vocab.json")
    }

    // Write merges.txt
    let mergesFile = "merges.txt"
    let mergesUrl = URL(fileURLWithPath: mergesFile)
    fm.createFile(atPath: mergesFile, contents: nil, attributes: nil)
    
    if let mergesHandle = try? FileHandle(forWritingTo: mergesUrl) {
        for pair in restoredMerges {
            let t0 = String(decoding: restoredVocab[Int(pair.0)], as: UTF8.self)
            let t1 = String(decoding: restoredVocab[Int(pair.1)], as: UTF8.self)
            let line = "\(t0) \(t1)\n"
            mergesHandle.write(line.data(using: .utf8)!)
        }
        mergesHandle.closeFile()
        print("[+] Saved reconstructed \(mergesFile) to current directory.")
    } else {
        print("[-] Error writing merges.txt")
    }

    // Copy config files from local models directory
    print("  Copying tokenizer configuration files...")
    let srcConfigPath = "j:/Language-U/Language-U-V2/qwen-3.5-0.8b-local/tokenizer_config.json"
    let fallbackConfigPath = "/mnt/j/Language-U/Language-U-V2/qwen-3.5-0.8b-local/tokenizer_config.json"
    var targetConfig = srcConfigPath
    if !fm.fileExists(atPath: targetConfig) && fm.fileExists(atPath: fallbackConfigPath) {
        targetConfig = fallbackConfigPath
    }
    
    if fm.fileExists(atPath: targetConfig) {
        try? fm.removeItem(atPath: "tokenizer_config.json")
        try? fm.copyItem(atPath: targetConfig, toPath: "tokenizer_config.json")
        print("[+] Copied tokenizer_config.json to current directory.")
    }

    let srcTokenizerPath = "j:/Language-U/Language-U-V2/qwen-3.5-0.8b-local/tokenizer.json"
    let fallbackTokenizerPath = "/mnt/j/Language-U/Language-U-V2/qwen-3.5-0.8b-local/tokenizer.json"
    var targetTokenizer = srcTokenizerPath
    if !fm.fileExists(atPath: targetTokenizer) && fm.fileExists(atPath: fallbackTokenizerPath) {
        targetTokenizer = fallbackTokenizerPath
    }
    
    if fm.fileExists(atPath: targetTokenizer) {
        try? fm.removeItem(atPath: "tokenizer.json")
        try? fm.copyItem(atPath: targetTokenizer, toPath: "tokenizer.json")
        print("[+] Reconstructed tokenizer.json copied to current directory.")
    }

    print("=========================================================")
    print("  SWIFT DECODER SUCCESSFUL!")
    print("=========================================================")
}

main()