amrex::ParticleHeader Struct Reference

Metadata prefix of an AMReX particle "Header" file. More...

#include <AMReX_ParticleHeader.H>

Classes
struct	GridEntry
	One entry of the per-level grid table. More...

Public Member Functions
void	parse (std::istream &is)
	Parse the metadata prefix from an input stream.
void	parse_grid_table (std::istream &is)
	Parse the per-level grid table into grids.

Static Public Member Functions
static ParticleHeader	read (const std::string &dir, const std::string &file)
	Read and parse the "Header" file of a particle plotfile/checkpoint.

Public Attributes
std::string	version
	raw version string, e.g. "Version_Two_Dot_One_double"
std::string	how
	precision the particle data was written in: "single" or "double"
bool	convert_ids = false
	whether particle ids need to be converted (Version_Two_Dot_One and later)
int	dim = 0
	AMREX_SPACEDIM the file was written with.
int	num_real = 0
	number of real components (pure SoA: excludes the AMREX_SPACEDIM positions)
Vector< std::string >	real_comp_names
	names of the real components, size() == num_real
int	num_int = 0
	number of integer components
Vector< std::string >	int_comp_names
	names of the integer components, size() == num_int
bool	is_checkpoint = false
	true if the file is a checkpoint (full restart), false for a plotfile
Long	num_particles = 0
	total number of particles in the file
Long	next_id = 0
	the next particle id to hand out (maxnextid)
int	finest_level = 0
	finest level present in the file
Vector< Vector< GridEntry > >	grids
	Per level and grid: where each grid's particle data is stored.

Detailed Description

Metadata prefix of an AMReX particle "Header" file.

This is the single source of truth for parsing the leading metadata of a particle plotfile/checkpoint "Header" file, as written by WriteBinaryParticleData. It is used both by ParticleContainer::Restart (to read a file into a container) and standalone (e.g. from language bindings) to discover the on-disk layout - number and names of the real/int components, the precision, and whether the file is a checkpoint - before constructing or configuring a container to read it.

The component counts follow the same convention as the on-disk format: for pure Struct-of-Arrays particles the AMREX_SPACEDIM position components are implicit and num_real counts only the remaining (runtime) real components.

After the metadata prefix, the "Header" file contains a per-level grid table that locates each grid's binary particle data on disk. It can be parsed with parse_grid_table(); read() returns a fully parsed header including the table.

Member Function Documentation

parse()

void amrex::ParticleHeader::parse

(

std::istream &

is

)

Parse the metadata prefix from an input stream.

Consumes exactly the leading metadata block (version through finest level), leaving the stream positioned right after the finest-level entry and before the per-level grid table. The table can then be parsed from the same stream with parse_grid_table().

Parameters

is	input stream positioned at the start of a particle Header

parse_grid_table()

void amrex::ParticleHeader::parse_grid_table

(

std::istream &

is

)

Parse the per-level grid table into grids.

Must be called on a stream positioned right after the metadata prefix, i.e. directly following a call to parse().

Parameters

is	input stream positioned at the start of the grid table

read()

ParticleHeader amrex::ParticleHeader::read ( const std::string &  dir, const std::string &  file  )

static

Read and parse the "Header" file of a particle plotfile/checkpoint.

Parses the complete header: the metadata prefix and the per-level grid table. Collective: the file is read on the I/O process and broadcast.

Parameters

dir	plotfile/checkpoint directory
file	particle sub-directory name (e.g. "particle0")

Returns

the parsed header metadata

Member Data Documentation

convert_ids

bool amrex::ParticleHeader::convert_ids = false

whether particle ids need to be converted (Version_Two_Dot_One and later)

dim

int amrex::ParticleHeader::dim = 0

AMREX_SPACEDIM the file was written with.

finest_level

int amrex::ParticleHeader::finest_level = 0

finest level present in the file

grids

Vector<Vector<GridEntry> > amrex::ParticleHeader::grids

Per level and grid: where each grid's particle data is stored.

Filled by parse_grid_table(); grids.size() == finest_level + 1 and grids[lev].size() is the number of grids of level lev. Grids without particles have a zero count.

how

std::string amrex::ParticleHeader::how

precision the particle data was written in: "single" or "double"

int_comp_names

Vector<std::string> amrex::ParticleHeader::int_comp_names

names of the integer components, size() == num_int

is_checkpoint

bool amrex::ParticleHeader::is_checkpoint = false

true if the file is a checkpoint (full restart), false for a plotfile

next_id

Long amrex::ParticleHeader::next_id = 0

the next particle id to hand out (maxnextid)

num_int

int amrex::ParticleHeader::num_int = 0

number of integer components

num_particles

Long amrex::ParticleHeader::num_particles = 0

total number of particles in the file

num_real

int amrex::ParticleHeader::num_real = 0

number of real components (pure SoA: excludes the AMREX_SPACEDIM positions)

real_comp_names

Vector<std::string> amrex::ParticleHeader::real_comp_names

names of the real components, size() == num_real

version

std::string amrex::ParticleHeader::version

raw version string, e.g. "Version_Two_Dot_One_double"

---

The documentation for this struct was generated from the following files:

• /home/runner/work/amrex/amrex/Src/Particle/AMReX_ParticleHeader.H
• /home/runner/work/amrex/amrex/Src/Particle/AMReX_ParticleHeader.cpp