|
| 1 | +use std::env; |
| 2 | +use std::fs; |
| 3 | +use std::path::Path; |
| 4 | +use walkdir::WalkDir; |
| 5 | +fn decompress_blosc(compressed_data: &[u8]) -> Result<Vec<u8>, String> { |
| 6 | + unsafe { |
| 7 | + // Get decompressed size first |
| 8 | + let mut nbytes = 0usize; |
| 9 | + let mut cbytes = 0usize; |
| 10 | + let mut blocksize = 0usize; |
| 11 | + |
| 12 | + blosc_sys::blosc_cbuffer_sizes( |
| 13 | + compressed_data.as_ptr() as *const std::ffi::c_void, |
| 14 | + &mut nbytes as *mut usize, |
| 15 | + &mut cbytes as *mut usize, |
| 16 | + &mut blocksize as *mut usize, |
| 17 | + ); |
| 18 | + |
| 19 | + if nbytes == 0 { |
| 20 | + return Err("Invalid compressed data".to_string()); |
| 21 | + } |
| 22 | + |
| 23 | + // Allocate output buffer |
| 24 | + let mut decompressed = vec![0u8; nbytes]; |
| 25 | + |
| 26 | + // Decompress |
| 27 | + let result = blosc_sys::blosc_decompress( |
| 28 | + compressed_data.as_ptr() as *const std::ffi::c_void, |
| 29 | + decompressed.as_mut_ptr() as *mut std::ffi::c_void, |
| 30 | + nbytes, |
| 31 | + ); |
| 32 | + |
| 33 | + if result < 0 { |
| 34 | + return Err(format!("Blosc decompression failed with code: {}", result)); |
| 35 | + } |
| 36 | + |
| 37 | + decompressed.truncate(result as usize); |
| 38 | + Ok(decompressed) |
| 39 | + } |
| 40 | +} |
| 41 | + |
| 42 | +fn print_hexdump(data: &[u8], offset: usize, chunk_name: &str) { |
| 43 | + println!("=== {} ===", chunk_name); |
| 44 | + for (i, chunk) in data.chunks(16).enumerate() { |
| 45 | + let addr = offset + i * 16; |
| 46 | + |
| 47 | + // Print address |
| 48 | + print!("{:08x} ", addr); |
| 49 | + |
| 50 | + // Print hex bytes |
| 51 | + for (j, &byte) in chunk.iter().enumerate() { |
| 52 | + if j == 8 { |
| 53 | + print!(" "); // Extra space in the middle |
| 54 | + } |
| 55 | + print!("{:02x} ", byte); |
| 56 | + } |
| 57 | + |
| 58 | + // Pad if chunk is less than 16 bytes |
| 59 | + if chunk.len() < 16 { |
| 60 | + for j in chunk.len()..16 { |
| 61 | + if j == 8 { |
| 62 | + print!(" "); |
| 63 | + } |
| 64 | + print!(" "); |
| 65 | + } |
| 66 | + } |
| 67 | + |
| 68 | + // Print ASCII representation |
| 69 | + print!(" |"); |
| 70 | + for &byte in chunk { |
| 71 | + if byte >= 32 && byte <= 126 { |
| 72 | + print!("{}", byte as char); |
| 73 | + } else { |
| 74 | + print!("."); |
| 75 | + } |
| 76 | + } |
| 77 | + println!("|"); |
| 78 | + } |
| 79 | + println!(); |
| 80 | +} |
| 81 | + |
| 82 | +fn main() -> Result<(), Box<dyn std::error::Error>> { |
| 83 | + let args: Vec<String> = env::args().collect(); |
| 84 | + |
| 85 | + if args.len() != 2 { |
| 86 | + eprintln!("Usage: {} <zarr_array_path>", args[0]); |
| 87 | + eprintln!("Example: {} /path/to/zarr/array", args[0]); |
| 88 | + std::process::exit(1); |
| 89 | + } |
| 90 | + |
| 91 | + let zarr_path = Path::new(&args[1]); |
| 92 | + |
| 93 | + // Verify the path exists |
| 94 | + if !zarr_path.exists() { |
| 95 | + eprintln!("Error: Path '{}' does not exist", zarr_path.display()); |
| 96 | + std::process::exit(1); |
| 97 | + } |
| 98 | + |
| 99 | + println!("Reading Zarr array from: {}", zarr_path.display()); |
| 100 | + println!("========================================"); |
| 101 | + |
| 102 | + // Read zarr.json metadata |
| 103 | + let zarr_json_path = zarr_path.join("zarr.json"); |
| 104 | + if !zarr_json_path.exists() { |
| 105 | + eprintln!("Error: zarr.json not found in {}", zarr_path.display()); |
| 106 | + std::process::exit(1); |
| 107 | + } |
| 108 | + |
| 109 | + let metadata_content = fs::read_to_string(&zarr_json_path)?; |
| 110 | + let metadata: serde_json::Value = serde_json::from_str(&metadata_content)?; |
| 111 | + |
| 112 | + // Extract information from metadata |
| 113 | + let shape = metadata["shape"].as_array().unwrap(); |
| 114 | + let chunk_shape = metadata["chunk_grid"]["configuration"]["chunk_shape"].as_array().unwrap(); |
| 115 | + |
| 116 | + println!("Array shape: {:?}", shape); |
| 117 | + println!("Chunk shape: {:?}", chunk_shape); |
| 118 | + println!("Data type: {}", metadata["data_type"]["name"]); |
| 119 | + if let Some(config) = metadata["data_type"]["configuration"].as_object() { |
| 120 | + if let Some(length_bytes) = config.get("length_bytes") { |
| 121 | + println!("Length bytes: {}", length_bytes); |
| 122 | + } |
| 123 | + } |
| 124 | + println!(); |
| 125 | + |
| 126 | + // Calculate expected chunks based on the metadata we know: |
| 127 | + // Shape: [345, 188], Chunk shape: [128, 128] |
| 128 | + // This means we have ceil(345/128) = 3 chunks in dimension 0 |
| 129 | + // and ceil(188/128) = 2 chunks in dimension 1 |
| 130 | + // So we expect chunks: c/0/0, c/0/1, c/1/0, c/1/1, c/2/0, c/2/1 |
| 131 | + |
| 132 | + let mut chunk_files = Vec::new(); |
| 133 | + |
| 134 | + // Find all chunk files by walking the directory |
| 135 | + for entry in WalkDir::new(zarr_path) { |
| 136 | + let entry = entry?; |
| 137 | + let path = entry.path(); |
| 138 | + |
| 139 | + // Look for chunk files (they start with 'c/' in Zarr v3) |
| 140 | + if path.is_file() { |
| 141 | + let relative_path = path.strip_prefix(zarr_path)?; |
| 142 | + let path_str = relative_path.to_string_lossy(); |
| 143 | + |
| 144 | + if path_str.starts_with("c/") { |
| 145 | + chunk_files.push((path.to_path_buf(), path_str.to_string())); |
| 146 | + } |
| 147 | + } |
| 148 | + } |
| 149 | + |
| 150 | + // Sort chunk files for consistent ordering |
| 151 | + chunk_files.sort_by(|a, b| a.1.cmp(&b.1)); |
| 152 | + |
| 153 | + println!("Found {} chunk files:", chunk_files.len()); |
| 154 | + for (_, chunk_name) in &chunk_files { |
| 155 | + println!(" {}", chunk_name); |
| 156 | + } |
| 157 | + println!(); |
| 158 | + |
| 159 | + let mut total_offset = 0; |
| 160 | + |
| 161 | + // Read, decompress, and hexdump each chunk file |
| 162 | + for (chunk_path, chunk_name) in chunk_files { |
| 163 | + match fs::read(&chunk_path) { |
| 164 | + Ok(compressed_data) => { |
| 165 | + if compressed_data.is_empty() { |
| 166 | + println!("=== {} ===", chunk_name); |
| 167 | + println!("(empty chunk)"); |
| 168 | + println!(); |
| 169 | + } else { |
| 170 | + println!("Compressed size: {} bytes", compressed_data.len()); |
| 171 | + |
| 172 | + // Decompress the Blosc-compressed data using blosc-sys directly |
| 173 | + match decompress_blosc(&compressed_data) { |
| 174 | + Ok(decompressed_data) => { |
| 175 | + println!("Decompressed size: {} bytes", decompressed_data.len()); |
| 176 | + print_hexdump(&decompressed_data, total_offset, &chunk_name); |
| 177 | + total_offset += decompressed_data.len(); |
| 178 | + } |
| 179 | + Err(e) => { |
| 180 | + eprintln!("Error decompressing chunk {}: {}", chunk_name, e); |
| 181 | + println!("Showing raw compressed data instead:"); |
| 182 | + print_hexdump(&compressed_data, total_offset, &format!("{} (compressed)", chunk_name)); |
| 183 | + total_offset += compressed_data.len(); |
| 184 | + } |
| 185 | + } |
| 186 | + } |
| 187 | + } |
| 188 | + Err(e) => { |
| 189 | + eprintln!("Error reading chunk {}: {}", chunk_name, e); |
| 190 | + } |
| 191 | + } |
| 192 | + } |
| 193 | + |
| 194 | + println!("Total decompressed bytes processed: {}", total_offset); |
| 195 | + println!(); |
| 196 | + println!("Note: This shows the decompressed array data as it would appear in memory."); |
| 197 | + println!("Each element is 240 bytes (raw_bytes with length_bytes: 240)."); |
| 198 | + |
| 199 | + Ok(()) |
| 200 | +} |
0 commit comments