Parse file formats used by Heroes of Might and Magic III in node.js or modern browsers.
If there is any interest in the modding community (to say, edit files in browser), I can easily make DEFs/LODs repackable.
If anyone is in need of parsing some other file type (H3M, H3C, SND, VID…) feel free to open an issue and I'll see what I can do.
Node.js
npm install homm3-unpacker
const { unpackLOD, unpackDEF, unpackPCX } = require("homm3-unpacker")Browser
<script src="dist/homm3-unpacker.js"></script>const { unpackLOD, unpackDEF, unpackPCX } = window["homm3-unpacker"]Parse .LOD archive files.
buffer » binary data of a file - ArrayBuffer/node.js Buffer.
processor(buffer, filename, skip) » function to process extracted binary data for each contained file or object with processors for specific file formats; calling skip discards the result - function/object - defaults to (buffer) => buffer.
unpackLOD(file)
// {
// type: "lod (expansion)",
// files: {
// "ab.h3c": ArrayBuffer,
// "AdvEvent.txt": ArrayBuffer,
// "Ar_Bg.pcx": ArrayBuffer,
// …
// }
// }unpackLOD(file, (buffer, filename, skip) => (filename === "BoArt084.pcx" ? buffer : skip()))
// {
// type: "lod (expansion)",
// files: {
// "BoArt084.pcx": ArrayBuffer
// }
// }unpackLOD(file, {
"h3c": (buffer, filename, skip) => skip(),
"txt": (buffer) => buffer,
"pcx": (buffer) => buffer
})
// {
// type: "lod (expansion)",
// files: {
// "AdvEvent.txt": ArrayBuffer,
// "Ar_Bg.pcx": ArrayBuffer,
// …
// }
// }Parse .DEF animation files.
buffer » binary data of a file - ArrayBuffer/node.js Buffer.
options.format » format to encode image data to - "png"/"bitmap" - defaults to "png".
options.padding » include padding around images - boolean - defaults to false.
options.palette(colours) » function to process the file's colour palette - function - defaults to (colours) => colours.
unpackDEF(file, { format: "png", padding: true })
// {
// type: "def (creature)",
// fullWidth: 450,
// fullHeight: 400,
// palette: [ // 256 rgba colours.
// { r: 0, g: 0, b: 0, a: 0 },
// { r: 0, g: 0, b: 0, a: 64 },
// { r: 255, g: 100, b: 255, a: 0 },
// …
// ],
// groups: [
// "moving": ["CAbehe10.pcx", "CAbehe11.pcx", "CAbehe12.pcx", "CAbehe13.pcx", "CAbehe14.pcx", …],
// "mouse over": ["CAbehe01.pcx", "CAbehe05.pcx", "CAbehe06.pcx", "CAbehe07.pcx", "CAbehe08.pcx", …]
// …
// ],
// images: {
// "CAbehe10.pcx": {
// width: 116,
// height: 110,
// x: 173,
// y: 158,
// data: ArrayBuffer // PNG file.
// },
// "CAbehe01.pcx": {
// width: 88,
// height: 106,
// x: 174,
// y: 165,
// data: ArrayBuffer,
// selection: ArrayBuffer // Selection extracted into separate PNG.
// },
// …
// }
// }Parse .PCX images.
buffer » binary data of a file - ArrayBuffer/node.js Buffer.
options.format » format to encode image data to - "png"/"bitmap" - defaults to "png".
options.transparency » array of colours to make transparent when they are the first colour in palette - defaults to [{r:0,g:255,b:255}, {r:255,g:0,b:255}, {r:0,g:0,b:0}].
unpackPCX(file, { format: "png" })
// {
// type: "pcx (indexed)",
// width: 58,
// height: 64,
// data: ArrayBuffer // PNG file.
// }Note: not to be confused with PCX images in DEFs.
Reading/writing to files with the help of these two.
const fs = require("graceful-fs") // So we don't care about the number of open files.
function read(path) {
return new Promise((resolve, reject) => {
fs.readFile(path, (error, data) => (error ? reject(error) : resolve(data)))
})
}
function write(path, buffer) {
return new Promise((resolve, reject) => {
fs.writeFile(path, Buffer.from(buffer), (error) => (error ? reject(error) : resolve()))
})
}Extract all files from a LOD and keep them in memory.
Without a second parameter, the extracted files are stored as binary data in the .files dictionary object, under their respective filenames.
async function example1(path) {
const file = await read(path)
return unpackLOD(file)
}Previous example is equivalent to the following when a processor function is supplied as the second parameter.
The callback is invoked with (buffer, filename) for every file and its return value stored in .files dictionary.
async function example2(path) {
const file = await read(path)
return unpackLOD(file, (buffer) => buffer)
}When passing an object as the second parameter, its keys are matched against the file extension (case insensitively), and the callback's return value is stored instead of raw binary data.
To parse all DEF files with unpackDEF and store any other file as ArrayBuffer:
async function example3(path) {
const file = await read(path)
return unpackLOD(file, {
def: (buffer) => unpackDEF(buffer)
})
}
// Equivalent to this.
async function example4(path) {
const file = await read(path)
return unpackLOD(file, (buffer, filename) => (/\.def$/i.test(filename) ? unpackDEF(buffer) : buffer))
}Unpacking images from large LODs can take several minutes, it'd be nice to get some feedback while running. Let's log the completion of each parsed DEF:
async function example5(path) {
const file = await read(path)
return unpackLOD(file, {
def: (buffer, filename) => {
const result = unpackDEF(buffer)
console.log(`Unpacked "${filename}".`)
return result
}
})
}To extract files archived in a LOD into separate files.
// Extract while parsing.
async function example6(path) {
const file = await read(path)
unpackLOD(file, (buffer, filename) => { write(filename, buffer) })
}
// Extract after parsing.
async function example7(path) {
const file = await read(path)
const lod = unpackLOD(file)
for (const [filename, buffer] of Object.entries(lod.files))
write(filename, buffer)
}The following would extract every image in a LOD (animation frames and standalone images) as a PNG file.
async function example8(path) {
const file = await read(path)
const lod = unpackLOD(file, {
def: (buffer, filename) {
const def = unpackDEF(buffer, { format: "png", padding: true })
for (const [filename, image] of Object.entries(def.images)) {
write(filename, image.data)
if (image.selection)
write(filename, image.selection)
}
},
pcx: (buffer, filename) {
const image = unpackPCX(buffer, { format: "png" })
write(filename, image.data)
}
})
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent