Skip to content

Latest commit

 

History

History
166 lines (116 loc) · 6.12 KB

HDF5.rst

File metadata and controls

166 lines (116 loc) · 6.12 KB

NeuroRD HDF5 output structure

The file contains one model top-level group, and one or more trialXXX groups, each one corresponding to a different execution. Each table has metadata fields which describe the table.

Contents

The root group /

The attributes on this group contain information about NeuroRD version that was used to create this output file.

Example

/ (RootGroup) ''
  /._v_attrs (AttributeSet), 2 attributes:
   [build-time := b'2016-09-13 11:43:54',
    git-version := b'v3.1.4-6-g5308d22815+']

/model

This group contains the input model, the grid structure created from the morphology description, and the calculated dependency graph.

/model/species

A simple table with the names of all species, ordered the same as <Specie> elements under <ReactionScheme>.

/model/neighbors and /model/couplings

A pair of tables listing neighbours of each voxel (indices, -1 is used fill empty positions), and the coupling coefficient for each neighbour.

/model/regions

Names of regions.

/model/grid

This table describes the voxels.

/model/grid (Table(1,)) 'voxels'
  description := {
  "x0": Float64Col(shape=(), dflt=0.0, pos=0),
  "y0": Float64Col(shape=(), dflt=0.0, pos=1),
  "z0": Float64Col(shape=(), dflt=0.0, pos=2),
  "x1": Float64Col(shape=(), dflt=0.0, pos=3),
  "y1": Float64Col(shape=(), dflt=0.0, pos=4),
  "z1": Float64Col(shape=(), dflt=0.0, pos=5),
  "x2": Float64Col(shape=(), dflt=0.0, pos=6),
  "y2": Float64Col(shape=(), dflt=0.0, pos=7),
  "z2": Float64Col(shape=(), dflt=0.0, pos=8),
  "x3": Float64Col(shape=(), dflt=0.0, pos=9),
  "y3": Float64Col(shape=(), dflt=0.0, pos=10),
  "z3": Float64Col(shape=(), dflt=0.0, pos=11),
  "volume": Float64Col(shape=(), dflt=0.0, pos=12),
  "deltaZ": Float64Col(shape=(), dflt=0.0, pos=13),
  "label": StringCol(itemsize=100, shape=(), dflt='', pos=14),
  "region": Int32Col(shape=(), dflt=0, pos=15),
  "type": StringCol(itemsize=100, shape=(), dflt='', pos=16),
  "group": StringCol(itemsize=100, shape=(), dflt='', pos=17)}

(x0, y0, z0), (x1, y1, z1), (x2, y2, z2) are in one plane, and (x3, y3, z3) is the opposite corner in the other plane.

region is an index into the regions table described above.

/model/serialized_config

This "table" contains one string element with the input XML configuration.

/model/reactions

This group describes all reactions (not diffusion or stimulation):

  • reactants and reactant_stoichiometry contain the indices of consumed species and their stoichiometry;
  • products and product_stoichiometry contain the indices of produced species and their stoichiometry;
  • rates contain the rate constants of the reactions;
  • reversible_pairs contains the indices of reverse reactions (-1 is used to fill empty positions).

/model/stimulation

This group desribes the stimulation channels:

  • target_names contains the names of stimulation targets (this corresponds to the injectionSite= attribute),
  • targets contains the indices of stimulated voxels.

/model/dependencies

Calculated dependency graph.

  • descriptions contains the label of each reactions channel used in logging;
  • elements specifies in which voxels where reactions take place, where diffusion starts, or where molecules are injected;
  • types describes the type of each reaction
  • dependent constains the indices of dependent reaction channels

/model/output

This group contains a number of subgroups, each describing an a set of tables containing particle populations. One set (__main__) is configured through <outputInterval> in the main configuration file and contains all species for all voxels. The other sets are configured through <OutputScheme>. Each corresponds to a different <OutputSet>, and the name is taken from the filename attribute.

For each set, elements contains the indices of output elements, and species contains the names of output species.

/trialXXX/output groups contain corresponding tables with the simulation results and are described below.

Example

<outputInterval>   5   </outputInterval>

<OutputScheme>
  <OutputSet filename = "all"  outputInterval="100.0">
    <OutputSpecie name="A"/>
    <OutputSpecie name="B"/>
    <OutputSpecie name="C"/>
    <OutputSpecie name="D"/>
  </OutputSet>

  <OutputSet filename = "some"  outputInterval="50">
    <OutputSpecie name="C"/>
    <OutputSpecie name="D"/>
  </OutputSet>
</OutputScheme>
/model/output (Group) 'output species'
/model/output/__main__ (Group) ''
/model/output/__main__/elements (Array(1,)) ''
/model/output/__main__/species (Array(4,)) 'names of output species'
/model/output/all (Group) ''
/model/output/all/elements (Array(1,)) 'indices of output elements'
/model/output/all/species (Array(4,)) 'names of output species'
/model/output/some (Group) ''
/model/output/some/elements (Array(1,)) 'indices of output elements'
/model/output/some/species (Array(2,)) 'names of output species'

/trialXXX

The number of trials is specified with -Dneurord.trials=N. Each trial results in one top-level group with names trial0, trial, ..., trial<N-1>.

Attribute simulation_seed contains the simulation seed that was used for this trial. Even if the input XML file species a simulation seed, it is ignored for trials with numbers above 0.

/trialXXX/output

This group mirrors the /model/output group described above. Each output set contains two tables:

  • times is a uni-dimensional array containg the times at which the state of the system was reported,
  • population is an array of size T×V×S, where T is the size of the times table, V is the number of voxel in this output set, and S is the number of species in this output set.