<-- ANALYTICS CODE - DO NOT EDIT -->

<-- Google Analytics script BEGIN -->

<-- Google Analytics script END -->

<-- END OF ANALYTICS CODE -->

• A new Pacs Documentation page has been opened on the PacsDocumentation public wiki.

<-- COMMENT BOX CODE - DO NOT EDIT -->

<-- END OF COMMENT BOX CODE -->

• Changes between HIPE 5.2 and 5.3

• Changes between HIPE 5.1 and 5.2

See also the HIPE known issues and the Data products known issues pages for last-minute information on known problems.

• 'Varient' and 'overwrite' options removed from ExtractSpectrum task

• When reprocessing with the hifiPipeline task it is not recommend to set aux=True when using the HSA to retrieve your data as not all Auxiliary data products will be updated correctly. The default option of aux=False is valid for all observations created with SPG v4.0+ (which should be all HIFI observations in the HSA).

• A task to allow conversion of Kelvin to Jansky, ConverK2Jy, is introduced

• The fitting routine in fitBaseline has a condition set such that no polynomials of order n are fitted if less than n+1 channels are available. This prevents crashes of the task in the case of heavily flagged data.

• A bug that appeared in early versions of HIPE 5 preventing export of Spectral Scans now fixed
• HiClass now exports the source velocity as provided by the astronomer in the AOR into a new FITS header keyword. At this time, the CLASS software will not do anything with this information, but the requirement to provide ALL the information is satisfied by having the velocity information present in the FITS header somewhere. The keywords used, which are slightly different than the usual ones used to reflect the fact that only the velocity in the proposal is exported, are tabulated below.

These keyword takes care of the value, the unit and the reference frame. The PVELTYPE keyword will identify the type of 'velocity' (velocity, redshift, radio or optical convention).

• Updates to the HIFI Users Manual: HIFI cookbook, Data Primer, Running the HIFI Pipeline, HIFI Standing Wave Removal Tool , How to make Spectral Cubes and Exporting HIFI data to CLASS.

• A complete new GUI for the SpectrumFitter was added. Start it from the spectrum toolbox menu, select SpectrumFitterGUI. Besides the functionality that was available in the SpectrumFitterTool, the GUI has an export tab that enables you to export the residual, total model, and individual models as an object with the type of the input object and as text.

• The new spectrum toolbox panel is opened by clicking on in the toolbar
• Instead of dragging a selection of spectra to the dataset parameter, the task is applied on either the displayed or the selected spectra within the plot.

• The ra and dec in cube meta data is now set to be the centre of the cube. Previously the ra and dec in the first dataset of the HTP had been used.

• It is now possible to create a SimpleSpectrum from a HIFI product or dataset containing only one spectrum. A SimpleSpectrum is easily written to FITS and maintains metadata. The conversion is done as:

HIFI Data Processing Tools

• PolarPairs now has a tolerance of 100. Previously it as 5 and this sometimes caused the task to fail resulting in the output being identical to the first input product.

• A bug that appeared in early versions of HIPE 5 where domask=2 in did not work in FitBaseline (preventing automatic line masking) now fixed.
• A bug that appeared in early versions of HIPE 5 in which fitBaseline would not save on an intermediate quit is resolved.

• The HIFI cookbook can now be selected from the Help menu in the HIPE toolbar under Help -> Data Reduction -> HIFI

• Changes between HIPE 5.0 and 5.1

• The Spur table has been updated to reflect the improvements in band 1a. For data taken after OD 516 the spur table now only shows three entries:
  • 1a 537500.00 3000.0 LO_settling 1,2,3,4
  • 1a 540929.25 1000.0 Strong_spur 1
  • 1a 542306.25 1000.0 Weak_spur 1

HIFI Pipeline

hifiPipelineTask

Level 0 Pipeline

• H and V spectrometer astrometry

WBS Pipeline

• Fixes for frequency calibration in the case that line centres fall precisely in the middle of two pixels

Level 2 Pipeline

• DoMainBeamTemp now multiplies by the forward efficiency. In earlier versions of HIPE this was not an issue because a value of Feff=1 was assumed. This update was required as a consequence of the adoption of Feff=0.96 in the calibration tree.

HIFI Products

• Modifications to HifiSpectrumDatasets are now reflected in SpectrumExplorer when the HTP is viewed there.

HIFI Data Processing Tools

Standing Wave Removal

• Tooltips implemented in the FitHifiFringe GUI
• Improved error messaging from FitHifiFringe
• Bug that appeared in early versions of HIPE 5 where usermask max/min boxes in FitHifiFringe GUI disappeared fixed.

• A Deep Fourier Transform (DFT) module is available (not a part of FitHifiFringe)

Baseline removal

• Introduction of a new task, FitBaseline, to interactively fit and subtract or divide out baselines in HIFI spectra.
• Bug that appeared in early versions of HIPE 5 where domask=2 in did not work in FitBaseline (preventing automatic line masking) now fixed.

Deconvolution

• More understandable error message when deconvolution fails (usual due to poor data input).
• Plotting of DSB solution as deconvolutions runs now possible when diag_mode_on selected.

• Bug that appeared in early versions of HIPE 5 preventing export of Spectral Scans now fixed

HIFI Documentation

• HiClass now exports the correct frequencies for bands 6 and 7

What's New in HIPE 5

• Included new metadata slewTime returning the scheduled start time of the slew as a FineTime.

• Updated task sample code.

• • Wiki chapters

• • Standardised __ doc __ of general tasks (signature, description, parameter descriptions)

• Compress and Decompress tasks: they support zip, tar, gzip, tgz in a portable way. Decompress can guess the compression of a file.

• Perform bulk actions on several quality control reports simultaneously.
• Flags colour codes depending on their importance.

• New framework to process/combine data from several observations in the same pipeline.
  • This will allow, for instances, better results using MADmap algorithm with cross-scan data from two or more observations.

• • New start-time algorithm based on the DTCP time of the OD being processed instead of fixed timestamps.
  • New check on TM data based on the gap checker reports. Only those observations with complete tm data will be processed at this point.

• Obsids shown also in decimal format in all the application tables.
• Include proper motion data into the ObservationContext.
• SPG stand-alone application can now run on Mac OS with no Versant libraries available.
• On-demand request are now recorded into a report file.

• Improved error reporting in AsciiTableTool.
• FitsArchive:

• The Basic/Advanced selection page has been removed.
• The memory, proxy and Versant settings were restructured. There are no separate panels asking if you want to set a proxy or a Versant database. There is a dedicated page for the proxy setting and one for the Versant settings.
• Instead of the old instrument selection page, you are now offered the following list with check boxes for selection:

• The instrument selection window allows you to click in any combination of items to install those components.
• You can also deselect all the instruments to get a barebone HIPE framework only for FITS I/O and numeric analysis.
• If possible, the installer informs you of the amount of data to download for the selected combination of modules.

• Those using the expert build and wishing to run the pipeline from level zero will need to set the property var.hcss.instrument = hifi ' this so the system knows to use the HIFI binstruct settings for reading the packets. Those using the legacy hifi-new.props file will already have this property set.

• • For beam widths the fitted widths of circular Gaussians, corrected for the convolution with the Mars disk. Open question: is the fit of a circular Gaussian correct? Need to check by how much Mars moved during an integration.
  • For main beam efficiencies the values derived using the observed beam widths for the beam coupling. It is not yet decided if an average value or a linear interpolation between (and beyond) frequencies produces the best results
  • For aperture efficiencies the values are derived, again, using observed beam widths for the coupling of the beam to the source.

• DoPointing no longer fails when reapplied

• More understandable error message when deconvolution fails (usual due to poor data input).
• Plotting of DSB solution as deconvolutions runs now possible when diag_mode_on selected.

• New chapters in the HIFI User's Manual: What was done to my data? and HIFI Baseline Removal.
• Updates to the HIFI Users Manual: HIFI cookbook, Data Primer, Running the HIFI Pipeline, HIFI Standing Wave Removal Tool and Exporting HIFI data to CLASS.
• Updates to HIFI Pipeline Specification document: improved pipeline flow diagrams, improved information about Euler resampling used in DoFreq grid task of pipeline.
• Creation of FitHifiFringe URM entry.

• PACS "New Style Sliced" pipeline implementation is now activated.
  • Phot and spec ipipe script names were changed to be more logic and intuitive.
  • Pipeline Script access:

• Calibration framework:
  

• A new Pacs Documentation page has been opened on the public wiki.

• • • Now the standard processing script for scan maps

• • CorrectRaDec4SsoTask is now also accepting the new Horizons class definition. An example how to use it can be found in the attached.
  • photProject error map has been checked for NaN.
  • Exposure map has been removed in favour of the coverage map.
• Noise treatment
  • A new module herschel.pacs.spg.phot.PhotAddNoisePerPixelTask provides four ways to calculate the initial photometer noise.

• • New Horizons Product used in PACS pipeline (SSO observations).
  • Residual spatial offset PACS photometer blue/green vs red.
  • Corners shall follow changes in PhotAssignRaDec.
  • Separate blue and green in ArrayInstrument calfile and the code using it.

• • PACS EDP: phot pipeline adaptation to cope with variable scan velocity.

• • • noScanData: mask data where the ScanBBID is not set (includes turn arounds)

• • First draft of the unchopped mode pipeline added to the build.
    • A "newstyle" unchopped mode pipeline is in place which can be run automatically or interactively.
  • New task specSubtractOffPosition implemented for unchopped pipeline .
• New interactive tool SpectrumExplorer:
  • PACS Product interface to SpectrumExplorer
• Improved wavelength calibration:
  • New wavelength calibration applied in waveCalc task using the wavePolynomes cal file.

• • Calibration block trendAnalysis product added to the observation context.
• Extra options to specDiffChop (note these are expert options for the analysis of special observations!).

• Several fixes to specWaveRebin, etc. to reduce the amount of NaNs considerably in the rebinned data.
• SpecProject sets the flux to Jy.

• Residual spatial offset PACS photometer blue/green vs red --> PCalPhotometer_ArrayInstrument_FM_v6.fits.

• New Wavelength Calibration Product.

• Updated spectrometer spatial cal file --> PCalSpectrometer_ArrayInstrument_FM_v6.fits.
• The RSRF calibration tables contain "NaN" which are then introduced into the signal by rsrfCal.

• Compatibility of PACS products with HCSS tools improved: you can now properly select data points and use the fitter tool on PACS frames and cubes.

• Minor changes in logs and documentation.

• • Added time correlation task and concurrent glitch deglitcher task in scan map pipelines.
  • Comment out optical and electrical crosstalk corrections.
  • Removed plotting blocks.

• Added the new DTMosaic class that allows the user to plot data recorded by many detectors during one observation on a single page to compare data more effectively (mosaic plot). This features is available from HIPE command line or within DetectorTimelineExplorer.
• DetectorTimelineExplorer and DetectorTimelineExplorerComponent:
  • Added a new features that allows the creation of mosaic plot of the detectors via right-click on the desired array.
  • Added a new Color&Mask Preferences panel.
  • Improved the color scale (using the Color&Mask Preferences it is now possible to select the Min/Max value for the color scale).
  • Improved the mask visualization (using the Color&Mask Preferences it's now possible to visualize any specific mask value directly on the array display).

• Drag & drop a variable from the Variables view to the Console to write the variable name. The same behaviour can be achieved when dropping it to a Jython editor.
• Drag & drop selected text from a Jython editor to the Console or vice-versa to copy & paste that text.
• The small level-2 preview shown in the top-right corner of the Observation Context viewer can be enlarged by clicking on it.

• Title and label can now accept the LaTeX commands \textrm \textit \textbf \mathrm \mathit \mathbf. For example: V$_{\textrm{LSR}}$

• RadialVelocity now uses TimeCorrProduct to select pointing data.
• SSO now uses TimeCorrProduct to select pointing data.
• Improved documentation.

• A complete new GUI for the SpectrumFitter was added. Start it from the SpectrumExplorer right mouse menu, select SpectrumFitterGUI. Besides the functionality that was available in the SpectrumFitterTool, the GUI has an export tab that enables you to export the residual, total model, and individual models as an object with the type of the input object and as text.

• The SpectrumFitter (commandline version) has methods to export the data as the input type and as text.
• The SpectrumFitter (commandline version) has method that returns a Jython script.

• RMS has been renamed to QRMS.
• RMS implements a new functionality (see the User Reference Manual).

• GeoMean, which calculates the geometric mean of a array of numeric data types with 1 to 5 dimensions.
• Mode, which returns the mode(s), or the most common element(s), of an array of numeric data with 1 to 5 dimensions.
• Covariance, which returns the covariance between two random variables/vectors x and y with finite second moments. If one vector is longer than the other, only the values up to the length of the shorter vector will be taken into account.
• CovarianceMatrix, which returns the covariance matrix of the input M x N matrix. The result is a N x N matrix with each i, j value equal to the covariance of the ith and jth columns of the original matrix.

• • Median mode uses median absolute deviation instead of standard deviation.

• Images can be opened as RgbImage directly from the Navigator in HIPE.
• Regular image files (jpg, gif, png, ...) are shown as preview in the outline when being clicked on in the navigator view of HIPE.
• Two images can be compared by setting the opacity of the image. This can be done using the setOpacity(float) method or using a slider in the Image Display.

• Added the methods containsNorthCelestialPole and containsSouthCelestialPole to the Wcs.
• Regridding an image on the grid of another image using RegridTask.
• Speed improvement of createRgbImageTask.
• Update of documentation in the User Reference Manual.
• You can now crop an image by drawing a rectangle on it
• Correction of the calculation of the dimensions of a mosaic.

• Task & Tools Toolbox: All task and tool panels are displayed in the same toolbox panel and selectable through a drop-down list. The task GUI components have become clearer as well.

• Load selectable datasets from a product: When opening the spectrum explorer on a product, a selection table is displayed showing all datasets that are stored in this product along with certain characteristics.

• Realtime statistics information: Selecting the Statistics task within the Toolbox displays realtime statistics while dragging ranges in the plot.
• Added relativistic frequency to velocity transform: When changing the unit of the wave axis to velocity the transform is now relativistic.

• Gaussian resampling introduced.

• • Better error reporting
  • File dialogue windows can show hidden files
  • Versions of observations inside pools are now shown

• • Releases memory calling Java garbage collector if it is needed.
  • Can write compressed files.
  • Can save a binary representation of a product if the FITS save operation fails.

• Product comparison API is available in the ia_toolbox_util module.

• Include new metadata slewTime returning the scheduled start time of the slew as a FineTime.
  • Specific accessors, getSlewStartTime() and setSlewStartTime(), where added to the ObservationContext class.
  • For old observations in the HSA without this value, the method will return a null value indicating that this field was never set before.

• • Phot and spec ipipe script names were changed to be more logic and intuitive

• • Two-stage photometer pipeline with masked highpass filtering

• • Photometer mapping accepts a variable input pixelsize (parameter pixfrac in the PhotProjectTask and the MapIndexTask).
  • The MapIndexViewer has a Sigclip gui to display the effect of different Sigclip parameters on the second level deglitching. Now it is not necessary any

• • The MapIndexViewer shows mean/median +/- error permanently in the signal plot
  • The Full MapIndex needs less memory than before
  • Wcs4mapTask got a crota option to manually set the requested rotation angle of a map.
  • CorrectRaDec4SsoTask is now also accepting the new Horizons class definition. An example how to use it can be found in the attached
  • photProject error map has been checked for NaN
  • Exposure map has been removed in favour of the coverage map

• • A new module herschel.pacs.spg.phot.PhotAddNoisePerPixelTask provides 4 ways to calculate the initial photometer noise.

• • Residual spatial offset PACS photometer blue/green vs red
  • Corners shall follow changes in PhotAssignRaDec
  • Separate blue and green in ArrayInstrument calfile and the code using it
  • Astrometric offset between two scan maps in parallel mode on the same field.

• • PACS EDP: phot pipeline adaptation to cope with variable scan velocity

• • Sigclip can now be applied n-times in one go.

• • • A "newstyle" unchopped mode pipeline is in place which can be run automatically or interactively
  • New task specSubtractOffPosition implemented for unchopped pipeline

• • PACS Product interface to SpectrumExplorer

• • New wavelength calibration applied in waveCalc task using the wavePolynomes cal file

• • Calibration block trendAnalysis product added to the observation context

• • Extra option 'extend' added which allows for other pairwise differencing schemes. These schemes will subtract more datapoints from each other than the standard scheme.
  • Extra option 'chopNodScheme' which provides different scheme for determining which label is on-source and which is off-source for a given nodding position. Needed for some special test AORs.
• Different bands in a range scan are now stored in their own blocks in the block table. This means that it is now possible to slice SEDs per band.

• Residual spatial offset PACS photometer blue/green vs red --> PCalPhotometer_ArrayInstrument_FM_v6.fits

• New Wavelength Calibration Product
• Version 8 of the Pacs Spectrometer Array to Instrument and version 4 of the Module2Array cal file are inserted in the system. This generates correct values also for center chopper positions (unchopped and waveswitch mode).
• Updated spectrometer spatial cal file --> PCalSpectrometer_ArrayInstrument_FM_v6.fits
• The RSRF calibration tables contain "NaN" which are then introduced into the signal by rsrfCal

1. Flux Calibration updated for the nominal voltages.
2. Temperature Drift Correction to match the new flux conversion.

• Updated to match the new products and dependencies.
• The spireCal task now has an option to read from a local pool.

• A new SpireMultiObs task to process several observations through the pipeline at once.

Feedback

A feedback button has been added for documentation and HIPE functionality.

• SSO now uses TimeCorrProduct to select pointing data

• • For old observations in the HSA without this value, the method will return a null value indicating that this field was never set before

• • Standardised _ doc _ of general tasks (signature, description, parameter descriptions)

• New metadata into the QualityContext product. The field is fixed and the main interest is to include this information into the FITS file associated to the product.

• Perform bulk actions on several quality control reports simultaneously
• Flags colour codes depending on their importance

• H and V spectrometer astrometry

• Task & Tools Toolbox All task and tool panels are displayed in the same toolbox panel and selectable through a drop-down list. The task gui components have become more clear as well. Instead of dragging a selection of spectra to the dataset parameter, the task is applied on either the displayed or the selected spectra within the plot.
• Load selectable datasets from a product When opening the spectrum explorer on a Product, a selection table is displayed showing all datasets that are stored in this Product along with certain characteristics. Just as within the original spectrum selection panel, these datasets can be sorted by clicking on the headercolumns. To load a set of datasets instead of just one or all, select each row and right click to load the selected datasets.
• Realtime statistics information Selecting the Statistics task within the Toolbox displays realtime statistics while dragging ranges in the plot.
• Added relativistic frequency to velocity transform When changing the unit of the wave axis to velocity the transform is now relativistic

• It is now possible to remove the smoothed baseline calculated by FitHifiFringe from the spectrum, in addition to the fitted sine waves. This option (sub_base) should be used with care as real spectral features can be removed this way. For interactive fitting of polynomial baselines, the new task FitBaseline is recommended.

• Improved error messaging

• DoPointing no longer fails when reapplied

• Fixes for frequency calibration in the case that line centres fall precisely in the middle of two pixels

• hifiPipeline GUI layout improved to make more efficient use of space

• HiClass will now appear an an applicable task for HTP as well as ObservationContexts

• Introduction of a new task, FitBaseline, to interactively fit and subtract or divide out baselines in HIFI spectra.

• Gaussian resampling introduced

• Unnecessary and deprecated properties files have been removed from the system

• • for main beam efficiencies the values derived using the observed beam widths for the beam coupling. It is not yet decided if an average value or a linear interpolation between (and beyond) frequencies produces the best results
  • for aperture efficiencies the values are derived, again, using observed beam widths for the coupling of the beam to the source.

• H and V spectrometer astronometry

• The path/filename for a user defined/edited pipeline algorithm can now be passed to the hifiPipeline task via the command line, previously this was only possible in the GUI.

obs =hifiPipeline (obs=obs, level2AlgoPath="/home/Me/myLevel2Algo.py")

The older method of defining a function with the def keyword in Jython and passing that function to the hifiPipeline task still exists

obs =hifiPipeline (obs=obs, level2Algo=level2PipelineAlgo)

• hifiPipeline will make an ObservationContext even if creation of trend products fails

• More understandable error message when deconvolution fails (usual due to poor data input)
• Plotting of DSB solution as deconvolutions runs now possible when diag_mode_on selected

• Now possible to remove from the spectrum the smoothed baseline calculated by FitHIFIFringe. This option should be used with care as real spectral features can be removed this way
• Tooltips implemented

FitBaseline

• New task introduced to interactively fit and subtract or divide out baselines in HIFI spectra.

• New chapters in the HIFI User's Manual: "What was done to my data?" and "HIFI Baseline Removal"
• Updates to the HIFI Users Manual: "HIFI cookbook", "Data Primer", "Running the HIFI Pipeline", "HIFI Standing Wave Removal Tool", and "Exporting HIFI data to CLASS"
• Updates to HIFI Pipeline Specification document: improved pipeline flow diagrams, improved information about Euler resampling used in DoFreq grid task of pipeline
• Creation of FitHifiFringe URM entry

• Generally most HIFI properties have moved from the hcss.dp.hifi build to the hcss.icc.hifi (expert) area and it is recommended that all internal ICC users always install the expert build in order to be able to run the pipeline from level -1.0.

• Those using the expert build and wishing to run the pipeline from level zero will need to set the property var.hcss.instrument = hifi ' this so the system knows to use the HIFI binstruct settings for reading the packets. Note those using the legacy hifi-new.props file will already have this property set.

• The hcss.dp.hifi (non-expert) build has pre-configured pools to access the HSA and the local pool via the Product Browser and PAL Storage Manager

• The minimum and maximum chopper threshold values have been changed.
  • HkThresholdsContext_H fpuChopperMin -> -0.04
  • HkThresholdsContext_V fpuChopperMin -> -0.04
  • HkThresholdsContext_H fpuChopperMax -> 0.06
  • HkThresholdsContext_H fpuChopperMax -> 0.06

• The upper limit of the mixer current quality control check for HEBs (bands 6 and 7) have been decreased from 0.08 mA to 0.055 mA, where tests have shown the mixers start to be under-pumped.

• Update to calibration tree for results of Mars beam measurements

• • apertureEfficiency-H
  • apertureEfficiency-V
  • beamEfficiency-H
  • beamEfficiency-V
  • beamWidths-H
  • beamWidths-V

The approach so far is to strictly use observed numbers (see tables below), without involving fitting results or empirical formulas. In particular this means: * for beam widths the fitted widths of circular Gaussians, corrected for the convolution with the Mars disk. See Raphael Moreno's report for details. Open question: is the fit of a circular Gaussian correct? Need to check by how much Mars moved during an integration. * for main beam efficiencies the values derived by Raphael (case c, i.e. using the observed beam widths for the beam coupling). Do we want to interpolate linearly between (and beyond) frequencies or do we want to use an average value per band? * for aperture efficiencies the values derived by me, again using observed beam widths for the coupling of the beam to the source.

In addition, the already available entries for the forward efficiencies (H and V) are updated from 1.0 to 0.96 (in all bands, all frequencies) to be compatible with above results. This requires the change of constant 1.0 to 0.96 .