PolyGun/docs/MAP_FILE_FORMAT.MD

Map file format

Overview

Format is little endian. It has following structure:

Section
Header
Chunks data
Chunk headers

Information about data types used in format:

Name	Description
char string	Unterminated non-prefixed static size array of bytes
uint16	Unsigned 16-bit integer
uint32	Unsigned 32-bit integer
uint64	Unsigned 64-bit integer

Header

File header consists of following fields:

Name	Size (bytes)	Type	Offset	Remarks
File magic	7	char string	0	It always has value polymap
Format version number	2	uint16	7	For development version 0xFFFF number should be used
Chunk count	4	uint32	9
Chunk headers array offset	8	uint64	13	Offset from beginning of file

Single chunk data entry

Immediately after header chunk data array is located. Each entry has following structure:

Name	Size (bytes)	Type	Offset	Remarks
Data	varying	uint8 array	0	Compressed chunk data of size specified earlier

Single chunk header

Name	Size (bytes)	Type	Offset	Remarks
Chunk X position	4	int32	0
Chunk Y position	4	int32	4
Chunk Z position	4	int32	8
Compression mode	1	uint8	12	See Compression modes for possible values
Data size	2	uint16	13	Size of compressed chunk data
Data offset	8	uint64	15	Offset of chunk data in Chunk data file section

Compression modes

• 0: No compression. Every node is stored using it's 16-bit ID
• 1: Efficient Huffman coding. Node count is stored first then node ID, however following encoding is used to decrease size: node count is stored in first 7 bits in byte. Last bit is used to denote if count spans to next byte, if this is a case then that bit is high (1), else it's low (0). Because there are only 7 bits available maximum value that can be stored is 127. Situation repeats for next bytes if last bit was 1. Same applies to node ID.
• 2: Fill. If chunk is filled with just one node type then the only thing that is stored is node ID as regular uint16.

Compression mode will be selected depending on which mode gives best result in particular case.