This page was created by the IDL library routine
mk_html_help2.
Last modified: Wed Feb 26 03:16:53 2025.
Convert a byte array to an array of given type
:Private:
:Params:
aPktData : in, required, type=byte array
The binary data for an entire das2 packet
:Keywords:
debug : in, optional, hidden, type=bool
If !true print debugging info to standard error
:Returns:
an array of given type, time values are converted to TT2000
:History:
Jul. 2018, D. Pisa : original
May 2019, C. Piker : updates to auto-convert times to TT2000
(See external/das2pro/src/das2decoder__define.pro)
Initialize a das2 stream data value decoder. Handles coversion of
time values to TT2000.
:Private:
TODO: This should be in das2_parsestream.pro since it uses packet
headers.
:Param:
hPlane : in, required, type=hash
A hash as return by xml_parse for the <x>, <y>, <z> or <yscan> plane
of interest.
(See external/das2pro/src/das2decoder__define.pro)
Return the total number of bytes converted to data values in each invocation of das2decoder::decode. :Private:
(See external/das2pro/src/das2decoder__define.pro)
Data plane decoder object, plain object, does not derive from IDL_Object, can't have auto properties :Private:
(See external/das2pro/src/das2decoder__define.pro)
Das2 dimension object, provides arrays and properties for a single physical dimension, such as Time or Spectral Density
(See external/das2pro/src/das2dim__define.pro)
Answers the question, how big are data packets for this dataset.
Datasets can be parsed from das2 streams. A stream consists of
header and data packets. Each data packet must have a sufficent
number of data values to increment the highest index of the internal
data arrays by 1.
:Returns:
The number of total bytes in each data packet for this dataset.
(See external/das2pro/src/das2ds__define.pro)
Inspect all owned variables and get the index range of the overall dataset. Map sizes to array dimensions using the index maps. :Private:
(See external/das2pro/src/das2ds__define.pro)
Provide the key names for all physical dimensions in the dataset
:Keywords:
D : in, type=boolean, optional
Return just the string ids for the data dimensions
C : in, type=boolean, optional
Return just the string ids for the coordinate dimensions
:Returns:
list - A list of strings containing the hask keys for the requested
physical dimensions.
:Author:
Chris Piker
(See external/das2pro/src/das2ds__define.pro)
Dataset objects (das2ds) explicitly denote, and separate, array index
dimensions and physical dimensions.
For each physical space, such as Time or Spectral Density, there is a das2dim
(das2 dimension) member object of a dataset. Each dimension may have more
than one set of applicable values. For example in the Time dimension there
may be a set of values that represent when a measurement was taken in UTC
and, the time uncertianty in seconds. Each set of measurements and thier
units are contained within a das2var (das2 variable) object.
To make plots and analyze data we have to know which values "go together".
Considering an energy spectra dataset for a moment, we would need to know
what array index numbers should be used to match a time, with an energy level
with a count rate, with a look direction. This is a bookkeeping problem.
We could solve this problem in a manner similar to the ISTP standard by
insisting that each dimension of an array correspond with a single physical
dimension. Doing so quickly results in display tools confusing index space
for physical space, a problem that we want to avoid. Instead, to provide this
information, an overall dataset index space is defined. Each variable within
the dataset provides a mapping between it's array indices and the overall
dataset index space. This mapping can be used to correlate values from
various arrays without constraining the nature of the physical measurements
contained within the dataset.
:Author:
Chris Piker
(See external/das2pro/src/das2ds__define.pro)
Das2 property has a type and value
(See external/das2pro/src/das2prop__define.pro)
Provide index ranges for this variable in terms of the top level dataset
index space.
Cordinate variables often have lower dimensional arrays than the data
variables. For example, think of an insturment that collects energetic
partical hit-counts in 45 energy bands once per minute. The coordinates of
this dataset would be energy and time, with the data being the count rate.
An hour's worth of these measurements could be stored in the following
arrays:
```
aTime = long64arr(60)
aEnergy = fltarr(45)
aCounts = intarr(45, 60)
```
Looking at the index ranges for this simple dataeset it's apparent that the
first index of array aTime must correspond to the second index of array
aCounts. To help visualize this mapping, especially when datasets become
more complex, we could "line-up" all the index ranges to make the mapping
more explicit:
```
Values Extents
------ -------
Time [ - , 60]
Energy [ 45, - ]
Counts [ 45, 60]
```
This is what the idxmap function outputs, the mapping of index space of a
single array to the overall dataset index space. Assume now that these
arrays are actually the `.values` member of three different das2var objects.
Calling `.dshape()` would yield the following (without comments of course):
```
print, vTime.dshape()
-1 0 ; var is degenerate in first dataset index (0 values)
0 60 ; second dataset index maps to first var index (60 values)
print, vEnergy.dshape()
0 45 ; first dataset index maps to first var index (45 values)
-1 0 ; var is degenerate in the second dataset index (0 values)
print, vCounts.dshape()
0 45 ; first dataset index maps to first var index (45 values)
1 60 ; second dataset index maps to second var index (60 values)
```
:Returns:
A 2 by N array, where N is the number of independent indexes required to
correlate all values in a dataset.
:Author: Chris Piker
(See external/das2pro/src/das2var__define.pro)
Provide access to the underlying value using standard array indexing. Due to limitations of the IDL language (i.e. array._overloadBracketsRightSide doesn't exist) ranges are only supported up to 3 index dimensions. To get a block range for arrays larger than rank 3 use the .array() method to get the array variable directly. To avoid suprise return values when gathing a slice of data, the indexmap is ignore. If you are getting single values (not recommended in IDL) and wish to work in dataset index space use the .at() method.
(See external/das2pro/src/das2var__define.pro)
Das2 Variable, an array, it's units and it's index map.
(See external/das2pro/src/das2var__define.pro)
Description: Convert a double represeting a binary das2 time value to tt2000
tt2000 is the number of nano-seconds since 2000-01-01T11:58:55.816
:Params:
timeunits: in, required
This is one of the strings 'us2000', 'mj1958', 't2000' or 't1970'
value: a DOUBLE value
To read UTC time strings in a variety of formats use das2_text_to_tt2000
instead of this function.
To avoid leap seconds, epoch time is first converted to a calendar
representation and the broken down calendar time is converted to TT2000
using the built-in CDF_TT2000 proceedure
(See external/das2pro/src/das2_double_to_tt2000.pro)
Parse the contents of a byte buffer into a list of das2 dataset (das2ds)
objects. This is an all-at-once parser that can't handle datasets larger
than around 1/2 to 1/3 of the host machine's ram.
:Returns:
list - a list of das2 dataset (das2ds) objects.
(See external/das2pro/src/das2_parsestream.pro)
Try to determine a decent name for a dimension given a units string, return !null if nothing works :Private:
(See external/das2pro/src/das2_parsestream.pro)
Inspect the properties sub item of a stream header object hash and
pull out the requested property value, if present.
:Private:
:Params:
hObj : in, required, type=hash
A hash, as returned by xml_parse() of the <stream>, <x>, <y>, <z>,
<yscan> element in question.
sProp: in, required, type=string
The property to find
:Returns:
!null if the property is not present or a type object
(int, String, datum, datum range, etc.)
(See external/das2pro/src/das2_parsestream.pro)
Make a new variable given a plane header :Private:
(See external/das2pro/src/das2_parsestream.pro)
Given a stream header, a plane header and a plane type,
make a new physical dimension structure
:Private:
:Returns:
The new dimension
(See external/das2pro/src/das2_parsestream.pro)
When multiple planes of the same type are present in the stream it's possible that they are related
(See external/das2pro/src/das2_parsestream.pro)
Add dimenions and variables from a single type of plane to a dataset
:Private:
:Params:
hStream : in, required, type=hash
hPkt : in, required, type=hash
sPlaneType : in, required, type=string
The plane tag, one of x,y,z,yscan depending on which type of
planes to find
sDatAxis : in, required, type=string
The axis for data from this plane, which is not necessarily
the plane tag. For examply yscan data can be either y values
(waveform) or z values (spectra)
idxmap : in, required, type=intarr
How values from this dat array map to overall dataset indices
since IDL is backwards from everyone else these are easy to
mix up.
iOffest : in, required, type=int
The offset into the packet header for data that has
already been claimed by other dimensions. Depends on all
<x><x>... coming before all <y><y>... before all <z><z>...
before all <yscan><yscan>... in the packet definition.
No way to check this currently since we are using xml_parse()
from IDL that does not preserve this information.
dataset : in, required, type=das2ds
The dataset object to receive these new dimensions
:Keywords:
FIRST : out, optional, type=das2dim
Used to save off the first dimension of a type.
:Returns:
The new offset into the records for the start of data from other
dimension sets.
unction _das2_addDimsFromHdr, $
hStream, hPkt, sPlaneType, sDatAxis, idxmap, iOffset, sKind, $
dataset, FIRST=dimFirst
compile_opt idl2, hidden
sUnitAttr = '%units'
if sPlaneType eq 'yscan' then begin
sUnitAttr = '%zUnits'
endif
bXOffset = !false
if hStream.haskey('properties') then begin
h = hStream['properties']
if h.haskey('%renderer') then bXOffset = (h['%renderer'] eq 'waveform')
if h.haskey('%String:renderer') then bXOffset = (h['%String:renderer'] eq 'waveform')
endif
if ~(hPkt.haskey(sPlaneType)) then return, iOffset
; make sure planes are always a list
lPlanes = hPkt[sPlaneType]
if typename(lPlanes) ne 'LIST' then lPlanes = list( hPkt[sPlaneType] )
; make decoders, variables and dimensions for each plane
for iPlane=0, n_elements(lPlanes) - 1 do begin
hPlane = lPlanes[iPlane]
decoder = obj_new('das2decoder', iOffset, hPlane)
iOffset += decoder.chunkSize()
var = _das2_varFromHdr( hPlane, idxmap, decoder)
; Add to an existing dimension (if min/max) or start a new one.
; When we restructure das2 streams, dimensions (or some other plane
; grouping mechanism) need to be added
propSrc = _das2_getProp(hPlane, 'source')
propOp = _das2_getProp(hPlane, 'operation')
; get an existing dim...
dim = !null
if propSrc ne !null then begin
if dataset.dims.haskey(propSrc.value) then $
dim = dataset.dims[propSrc.value]
endif
; ...or create a new one
if dim eq !null then begin
;printf, -2, sDatAxis, sKind, hPlane
dim = _das2_dimFromHdr(hStream, sDatAxis, hPlane, sKind)
if (sPlaneType eq 'x') && bXOffset && (dimFirst eq !null) then $
dim.vars['reference'] = var $
else $
dim.vars['center'] = var
; determining good names for dimensions is not easy...
sName = !null
if sSrc ne !null then sName = sSrc
if sName eq !null then begin
if hPlane.haskey('%name') then sName = hPlane['%name']
endif
if sName eq !null then begin
if hPlane.haskey(sUnitAttr) then sName = _das2_nameFromUnits(hPlane[sUnitAttr])
endif
if sName eq !null then sName = string(sPlaneType, iPlane, format='%s_%d')
if dataset.dims.haskey(sName) then $
message, 'Error in das2 stream, name '+sName+' repeats in the same packet"
dataset.dims[sName] = dim
;printf, -2, dataset
; Set the name based on the first plane of each kind.
; WARNING ASSUMES: _das2_addDimsFromHdr is called in order from
; <x> ... to <yscan> This is not the best
; idea in the world, but the das2.2 stream format has no concept of
; a dataset name.
if dimFirst eq !null then dataset.name = sName
endif else begin
message, 'Peaks-averages datasets such as Voyager SA are not yet supported'
sRole = _das2_op2role(propOp)
dim.vars[sRole] = var
endelse
; Save off the first <x><y><z><yscan> dim in case we need to add an offset
; variable to it from the ytags
if dimFirst eq !null then begin
dimFirst = dim
endif
endfor
return, iOffset
nd
--------------------------------------------------------------------------- ;
+
Make a new dataset object given stream and packet headers.
:Private:
:Params:
hStrmHdr: in, required, type=hash
The parsed stream XML header as returned from xml_parse
hPktHdr: in, required, type=hash
The parsed packet XML header as returned from xml_parse
:Keywords:
DEBUG: in, optional, private, type=bool
If true print debugging information
:Author:
Chris Piker (hence the snark)
(See external/das2pro/src/das2_parsestream.pro)
Parse data for a single packet using the given dataset structure and append
it to the variable arrays
note: since das2 streams don't put a length on the data packets there's
no way to cross check that the stream definition matches the length
of the data. All we can do is try to parse it without running of
the end
:Private:
(See external/das2pro/src/das2_parsestream.pro)
Convert all the variable value lists to arrays
(See external/das2pro/src/das2_parsestream.pro)
Look to see if a message was an exception, if it was, format it as a string and return it.
(See external/das2pro/src/das2_parsestream.pro)
Create a dataset structure from a list of packets.
:Private:
:Params:
pkts : in, required, type=byte array
A byte vector containing all the packets in a das2 stream, the
header packet (id [00]) should not be included.
:Returns:
list of dataset structures
:History:
Jul. 2018, D. Pisa : original
Nov. 2018, D. Pisa : fixed object.struct conversion for ypackets
May 2018, C. Piker : refactored
(See external/das2pro/src/das2_parsestream.pro)
Given a string, parse typical delimited ASCII date/time and
return year, month, day of month, day of year, hour, minute, second.
This function loads JULDAY and CALDAT as dependencies and prints error or
debug messages to stderr.
:Params:
timestr: in, required, type=string
Either 4-digit or 2-digit years can be parsed. 4-digit years must be in
the range range 1900 through 2199, and 2-digit years are only allowed in
the range 1960 through 2059.
year: out, required, type=integer
Year like 2000
month: out, required, type=integer
Month of year (1 through 12)
day: out, required, type=integer
Day of month (1 through 31)
doy: out, required, type=integer
Day of year (1 through 366)
hour: out, required, type=integer
Hour of day (0 through 23)
minute: out, required, type=integer
Minute of hour (0 through 59)
second: out, required, type=double
Second of minute (0.0d0 to 61.0d0)
:Keywords:
julian: out, optional, type=double
Double precision fractional Julian Day
debug: in, optional, type=bool
Turn on certian debug output strings
:Returns:
integer: 0 on success, 1 on failure
:Examples:
status = das2_parsetime (timestr, year, month, day_month, day_year $
[, hour, minute, second] [, julian=julian] [, /debug])
:History:
2012-10-30, L. Granroth: Rewrite using regex
2019-05-23, L. Granroth: Extended to handle nanoseconds
(See external/das2pro/src/das2_parsetime.pro)
Provide feedback that the stream is downloading :Private:
(See external/das2pro/src/das2_readhttp.pro)
Request data from a specific das2 server using native HTTP GET parameters
:Params:
sUrl: in, required, type=string
The full HTTP GET url to retrieve, it doesn't need to be URL encoded
this function will handle that step
:Keywords:
params: in, optional, type=list
ascii: in, optional, type=boolean
extras: in, optional, type=list
verbose: in, optional, type=boolean
:Returns:
A list of dataset objects. Each dataset object corresponds to a single
packet type in the input. If no datasets could be retrieved the function
return an empty list, so ret_val.length = 0.
:Requires:
xml_parse: IDL 8.6.1
IDLnetURL: IDL 6.4
:Example:
sServer = 'http://planet.physics.uiowa.edu/das/das2Server'
sDataset = 'Cassini/RPWS/HiRes_MidFreq_Waveform'
sBeg = '2008-08-10T09:06'
sEnd = '2008-08-10T09:13'
sParams = '10khz'
sFmt = '%s?server=dataset&dataset=%s&start_time=%s&end_time=%s¶ms=%s'
sUrl = string(sServer, sDataset, sMin, sMax, sParams, format=sFmt)
lDs = das2_readhttp(sUrl, /messages=sMsg)
if lDs eq !null then begin
print, sMsg
stop
endif
print, n_elements(lDs), /format="%d datesets read"
print, lDs[0]
:History:
Jul. 2018, D. Pisa : original
May 2019, C. Piker: refactored
(See external/das2pro/src/das2_readhttp.pro)
Description: Convert a string time value to a cdf_tt2000 epoch time
tt2000 is the number of nano-seconds since 2000-01-01T11:58:55.816
:Params:
sTime: in, required
This is a string time value such as '2007-001T14:00'
This function works by calling das2_parsetime to parse time components
from a variety of string formats, and then calling cdf_tt2000 to get
the epoch.
2019-01-24 L. Granroth, original
2019-05-23 C. Piker, nano second accuracy fix
(See external/das2pro/src/das2_text_to_tt2000.pro)