r43 - 30 Oct 2012 - 13:24:03 - DavideRizzoYou are here: TWiki >  Public Web  >  HipeWhatsNew > HipeWhatsNew7x

What's New in HIPE 7

PDF Version Click here

help Interested in what's new for other versions of HIPE? See this page for links to all the What's New documents from HIPE 4.0 onwards.

This document lists the changes in HIPE version 7.0 with respect to the 6.x series. Additional pages list changes between minor versions of the 7.x series.

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

Core system

HIPE

Console

  • Incremental search (Ctrl+I) available.
  • Console help text enhanced.
  • New internal commands !clear, !help and !history.

History

  • Detailed error information can be opened with a double click in the Trace column of the History view.

Status bar

  • Progress bar shows the current line number of the executing script.
  • New bin-shaped button to the right of the memory bar, at the bottom of the HIPE window, to run the garbage collector and free memory manually.

Generic viewers

Dataset viewer

  • Multiple cells can be selected and copied in the TableDataset viewer.

Jython editor

  • Find/Replace dialog enhanced. It can now show now whitespace characters.
  • Autoindentation of Jython code improved.
  • Indent and unindent buttons added to the toolbar.
  • If the Monospaced font is not really monospaced, HIPE proposes an alternate font.

Astronomical utilities

RadialVelocity task and utility class renamed to RadialV and now using the proper Doppler sign convention.

Numeric routines

Wavelet transform

  • new friendly interfaces are now available:
    • algo = ContinuousWavelet?("MexicanHat")
    • algo = DiscreteWavelet?("db5")
    • algo = StationaryWavelet?("coif3")

Images

Display

  • Adapted the popup menu to make it possible to work with axes, add a compass and add a scale.
  • Display now returns the commands done using the popup menu to the HIPE console. This makes it very easy to reconstruct an image plot from within a script.
  • When dragging source positions to an image, the circles are locked. When dragging contours to an image, the contours are also locked.
  • Added method addArcSecs to add a line of a given length in arcseconds to an image.
  • When using the opacity slider with a cube, the next image is used as the underlying image, not the next layer of the cube anymore.

Analysis

  • You can now load all monochrome or indexed gif files using importRgbImageTask.

Spectra

Display

Analysis

  • Addition to SpectrumFitter: create ASCII file with fit parameters. Use in the command line:
    • sf = SpectrumFitter(...)
    • ...
    • m = sf.addModel(...)
    • ...
    • sf.doFit( ) or sf.doGlobalFit( )
    • sf.saveFitParametersAsASCII(fileName) # writes parameters of all models to fileName
    • m.toASCII(dirName) # writes parameters to dirName/[modelname].txt
  • The SpectrumFitterGUI has, in the export TAB, also the possibility to export fit parameters to file.
  • Addition to SpectrumFitterGUI: user preference for initial models. If switched on, you can select from:
    • Instrument specific initial model.
    • User specific initial models - you can specify the models.
  • Addition to SpectrumFitterGUI, MultiTAB:
    • Some information on what will happen.
    • Fit parameters are written to file.
    • A count-down progress bar.

Calibration sources

Integrated the Calibration Source Database into HIPE. There is a Calibrators view in the Work Bench perspective, next to the Navigator view. You may have to reset the perspective for it to appear.

The Calibrators view allows access to the database of calibration sources at ESAC, used by HSC calibration scientists. It contains spectra of planets, asteroids, and fixed targets. The data can be viewed in the usual HIPE viewers for spectra and images, and manipulated using Jython commands.

Product Access Layer

  • The HTTP pool now supports digest as well as basic authentication.

Tasks

The default Task GUI has been improved with the following upgrades:

  • Parameter error detection (red X) also checks the validator of the parameter.
  • A new Help button gives easy direct access to the URM description of the task.
  • White space in layout moved to the lower part of the panel.
  • Editors based on text fields now require half the width.
  • Labels of parameters have now a maximum size to reserve space for the editors.
  • Support for grouping inputs in multiple tabs. For tasks that have inputs related to different aspects or simply a lot of them, they can be presented in several tabs.

There is now a detailed outline shown in the Outline view when a task is selected in the Tasks view. This allows inspection of the structure of a task without the need to either use the Console view or open its GUI in the Editor view.

Data input-output

  • simpleFitsReader and simpleFitsWriter now also support reading/writting datasets (not just products).
  • The Export view now shows the tags associated with each observation version.
  • The localStoreWriter task now allows to add a tag to the saved product.
  • localStoreWriter and localStoreCopier now use a better integrated GUI to select and create a pool.
  • File filters for file dialogs are now self-explanatory (accepted extensions ...).
  • The clear task now has a simplified syntax to delete all variables (clear(), without any parameter, will delete all variables).

Virtual Observatory

The VO external tools have been made accessible from the Tools menu.

We have also added a few pre-set applications (Topcat, Aladin and VOSpec) to the Send To menu. This menu is accessible by right-clicking on a variable. The pre-set applications are visible, even if the application in question is not running and/or HIPE is not connected to the VO. Using these pre-set is transparent to the user: When sending a variable, the application will be automatically started first and the data will be sent.

Documentation

Plug-ins

It is now possible for plug-in authors to publish new versions of plug-ins, which will be automatically picked up by existing users of the plug-in. If a new version of a plug-in is found by HIPE, and the author has indicated that the update is compatible with the user's version of HIPE, then HIPE will ask the user if the plug-in should be updated. The update mechanism is automatic, though HIPE must be restarted for the changes to be completed.

HIFI

HIFI Pipeline

  • doPipelineConfiguration now passes changes to the level 2 pipeline. To use this task, pass an HTP to the GUI (doPipelineConfiguration), modify the pipeline steps as required. Press accept (default name of output is out) then pass this new configuration to the pipeline task:
obs_new = hifiPipeline(obs=obs, params=out, save=False)

HIFI Data Processing Tools

Standing Wave Removal

  • new default doplot=2 option opens no more than 2 fitHifiFringe plot windows at the time and allows for inspection before processing next spectrum
  • averscan option fitHifiFringe has been removed as standing wave characteristics vary too much between scans
  • fitHifiFringe now shows up as applicable task to both HTPs and ObsContexts?

Baseline Removal

  • New non-stop option in interactive mode of fitBaseline allows automatic processing of remaining spectra
  • The htpcopy option of fitBaseline has been removed as modified HTPs are now properly displayed in SpectrumExplorer?

Deconvolution

  • doDeconvolution now produces a result even if deconvolution fails. In this case an error message is given
  • doDeconvolution now read flags from USB, rather than LSB. Channel flags have always been read from the USB so this change makes it easier for you to pass flagged data to the task

HiClass export tool

  • Obsid now exported again to CLASS header

PACS

Pipeline

Photometry

Improvements for SSO scan maps
Assuming the object move linear a linear interpolation between start and end position is done. This method provides 'smoother' coordinates especially for point source.

Old behavior :

frames = correctRaDec4Sso(frames, horizonsProduct=horizonsProduct, timeOffset=timeOffset, orbitEphem=orbitEphem)
frames = correctRaDec4Sso(frames, horizonsProduct=horizonsProduct, timeOffset=timeOffset, linear=1, orbitEphem=orbitEphem)
New implementation :
frames = correctRaDec4Sso(frames, horizonsProduct=horizonsProduct, timeOffset=timeOffset, linear=0, orbitEphem=orbitEphem) 

Aberration correction
Change the oder applying SIAM and Aberration correction in photAddInstantPointing and specAddInstantPointing . This shall have very minor impact on the coordinates, but correct reverse order of uplink.

Spectroscopy

  • All level-2 tasks now do an internal loop over line ID. They behave now the same as the level-0/1 tasks: If you do not select a particular line/range before then the level-2 tasks just process all lines/ranges. As a consequence the level-2 products have changed a bit:
  • We no longer have a list context for each line ID within another list context: All slices of a level-2 product are now stored in a single list context. This is much more practical when selecting slices based on arbitrary criteria.
  • The pacsSliceContext task now contains all the rules in the ipipe scripts as default rules: Also, it slices by line ID, raster column, raster line, nodding position, nod cycle number, is out of field and band. The rules in the scripts are also now optional and merely for information.
  • The interactive pipeline scripts (ipipe) are now tested in the nightly tester
  • The interactive pipeline scripts (ipipe) are moved in the build from the directory scripts/pacs/spg/pipeline/ipipe/spec to scripts/pacs/scripts/ipipe/spec since they had to be moved to the package pacs_scripts

Calibration

Photometry

Flux calibration
The photometer flux calibration has changed (new response calibration file "FM, 6"). Absolute fluxes of point-sources will change by less than 5%, extended sources will change by 10-20%. Aperture photometry of point-sources requires from now on the application of new aperture correction factors, taking into account that there is still significant flux beyond 60 arcsec.
These new correction factors (together with the old ones) are stored in: PCalPhotometer_ApertureCorrection_FM_v2.fits and can be retrieved via the new task photApertureCorrectionPointSource to be applied after annularSkyAperturePhotometry .
See also the update of PACS ipipe script Menu -> Pipeline -> PACS -> Photometer -> Copped point source -> standard pipeline -> L3_pointSourceAperturePhotometry.py
You need to update your local calibration tree Menu to activate it (Menu -> Tools -> PACS_CAL -> run Updater).
You may check by printing the calibration tree :
HIPE> c=getCalTree()
HIPE> print c.photometer

MadMap
New calibration files containing the actual 'best' calibration data for MadMap : PCalPhotometer_InvnttBL_FM_v4.fits
PCalPhotometer_InvnttBS_FM_v4.fits
PCalPhotometer_InvnttRed_FM_v4.fits
You need to update your local calibration tree Menu to activate it (Menu -> Tools -> PACS_CAL -> run Updater).
You may check by printing the calibration tree :
HIPE> c=getCalTree()
HIPE> print c.photometer

Spectroscopy

  • Wrong coordinates in Spectrometer results: update of spec spatial cal file
  • Improvements in wavelength calibration scatter for spectral pixels in band B2B -> wavelength calibration file updated

Analysis tools

Slicing

  • Saving intermediate data on machines with low memory is now done more clever: The slices are only saved to the temporary pool when the memory usage is more than 75% and the sliced product contains more than one slice.
  • The temporary pools of HIPE are now written under a directory 'null value' which is created where you start HIPE ( HcssSpr:12911 ).
    Workaround: Set the following property in your user.props file:
hcss.ia.pg.temporalpoolparent = ${user.home}/.hcss/lstore

These temporary pools are called tmp_product_sink_NNNNN

Photometry

  • A new Mask Array Viewer displays Masks with their correct size (also the timeline).
  • Add start time of last cooler recycling to meta data of products A new calibration product was introduced : CoolerRecyclingTimes
    • A convenience tool times = extractCoolerRecTimes(obs, calTree) returns the last cooler recycling of an observation
    • This file will be updated on a regular base

Spectroscopy

  • Map2signalCubeTask gets an additional shortcut: m2sc (besides map2signalCube)
  • Spectrum Explorer : New extension panel to edit masks is now fully working
  • Spectrum Explorer : Performance improvements

Product changes

  • calTree version is now in meta data of products

Photometry

Spectroscopy

  • Frames and PacsCube? meta data is made consistent for every product change (like select or join)
  • Additional meta keywords added: onOffSource, minWave, maxWave, nodCycle, lineId
  • All masks which are not excluded in rebinning are propagated from PacsCube? to PacsRebinnedCube?. They can be obtained with the getFlag() method which returns a Flag dataset, which is similar to a PACS mask.
  • We have now two more sliced products: SlicedPacsCube?? and SlicedPacsRebinnedCube??. SlicedPacsCube?? offers the same functionality as SlicedFrames??, but SlicedPacsRebinnedCube?? has limited functionality, since it does not have a blocktable. Add, remove and replace work, but selection should be done with the pacsSelectSlices task.

Each slice is described by the following slice info meta keywords (recognizable by description prefix "SLICE_INFO:"):

Keyword Values Description
lineId 0..N Numerical ID of a distinct wavelength range (usually a line scan)
lineDescription string Which spectral line has been observed (or tells that it is a range scan)
nodPosition A/B Nodding position
nodCycleNum 1..N Nodding cycle number
isOffPosition false/true Is it an off position according to the IsOutOfField? status?
onOffSource 0,1,2 Based on new ONSOURCE/OFFSOURCE Status keywords: 0 = not science, 1 = on-source, 2 = off-source
rasterId N M first value: raster column, second value: raster line
calBlock false/true Belongs to slice to a calibration block?
minWave float Minimum wavelength of range
maxWave float Maximum wavelength of range
aotMode string AOT mode

Also, since these keywords are now always consistent with the BlockTable? and Status contents they can and should be used for slice selection. The new task pacsSelectSlices does this meta-keyword based selection. It has the same interface as the selectSlices Python function.

  • All level-2 tasks now do an internal loop over line ID. They behave now the same as the level-0/1 tasks: If you do not select a particular line/range before then the level-2 tasks just process all lines/ranges. As a consequence the level-2 products have changed a bit:
  • The wavelength grid is now also stored in a list context, since one needs to pass several ranges from task to task.
  • We no longer have a list context for each line ID within another list context: All slices of a level-2 product are now stored in a single list context. This is much more practical when selecting slices based on arbitrary criteria.
  • Status item "GratingCycle" and "IndexInCycle" is now created for unchopped modes

Documentation

SPIRE

Calibration Products

  • Spectrometer calibration products for nominal detector settings have been updated to make use of in-flight bolometer model parameters.
  • An updated TempDriftCorr product allows a new algorithm in the Signal Jump Detector module to be used (nominal mode only).

Calibration Framework

  • The spireCal task has new features and modified behaviour:
    • It will look for a calibration tree in a local pool by default. The default is the one that was last saved if applicable; otherwise the default for the HIPE version.
    • It will only download from the HSA if explicitly requested via the calTree parameter.
    • Optionally the tree can be saved to a local pool. In this case that one will become the default.
    • The trees available in the archive can be listed.
  • The API has new methods for showing differences between two calibration trees and for verifying that expected metadata is present and correct.

Common Pipeline

  • The AssociateSkyPositionTask now creates a new table "avgVelocity" in photometer level1 scanmap products. The table contains two columns: "sampleTime" (TAI) and "avgVelocity" (deg/sec)
  • The AssociateSkyPositionTask now now accepts two new optional parameters:
    • scanVelocityTolerance: double value, in arcsec/sec, defining the scan velocity tolerance relative the the mode's expected scan velocity. Signal samples outside this range will have their NON_NOMINAL_VELOCITY mask set. (default: 1.0"/sec)
    • scanVelocityLimits: Double1d([lowerVelocityLimit, upperVelocityLimit]), in arcsec/sec, defining a range of scan Velocities. Signal samples with angular velocity outside this range will have their NON_NOMINAL_VELOCITY mask set. Setting this parameter, parameter "scanVelocityTolerance" will be ignored. Setting any velocity limit to a negative value, the task will ignore the scanVelocityLimits parameter and revert to using "scanVelocityTolerance".

Photometer Pipeline

  • The LpfResponseCorrectionTask now accepts a new optional, boolean input parameter "doMirrorSignal". Setting this parameter to true, the task will perform the correction, by mirroring the signal time-lines in the y-axis. The effect of this is to avoid any edge-ringing, resulting from the default correction method when applied to timelines with a large slope. The draw-back of the new method is that it can double the processing time of the Task, however, not to a detrimental effect to single observation processing (It will definitely be noticeable when processing a large number of observations).
  • The mosaicing of maps from overlapping observations in a proposal is now fully functional (this is actually available from 6.1).

Spectrometer Pipeline

Level 0 to Level 0.5

No changes.

Level 0.5 to Level 1

Changes to existing Tasks
  • WaveletDeglitcherTask now offers the option to make the number of corrected samples dependent on the glitch size by setting the optionReconstruction parameter to "adaptative" or "adaptativeLinear".
  • SpecFluxConversionTask no longer automatically removes non-seeing channels. One must now use FilterChannelsTask to remove these channels from SpectrometerDetectorSpectrum products.
  • SpecFluxConversionTask now precedes SpecTelescopeCorrectionTask in the Standard pipelines.
  • SpecTelescopeCorrectionTask now, by default, subtracts only a model of the Herschel telescope and not the model multiplied by the RSRF for Telescope (Sky) port.

Level 1 to Level 2

New Tasks
  • SpireProduct2SimpleSpectrumTask to convert SPIRE Level-1 and Level-2 data to SimpleSpectrum products
  • GriddingProjectionTask for spectral cube creation. Given SpirePreprocessedCube input, GriddingProjectionTask produces a spectral cube output. This task uses HIFI's GriddingTask.

Changes to existing Tasks
  • NaiveProjectionTask now creates a spectral cube containing error instead of weight.
  • AverageSpectraTask no longer provides functionality to remove out-of-band spectral samples. One must use RemoveOutOfBandTask instead.
  • AverageSpectraTask no longer provides functionality to reject outlier from the average computation.

Interactive Analysis and Tools

  • New task DestriperTask to iteratively remove the relative offsets between Level-1 detector timelines of the same detector array, and thus removing the stripes in the Level-2 map. Signal timelines are re-sampled from a preliminary naive map and subtracted from the original timelines. The resulting offset functions are modeled as polynomials with default degree (n=0) and subtracted. A new naive map is generated from the new difference of original signal and polynomial for the next iteration. The bright source threshold is used to suppress the influence of strong spikes in the data. The primary output is a SpireListContext containing updated PointedPhotTimelines. The additional outputs are a SimpleImage map, the convergence ratio, Tod etc.
  • New tasks BaselineRemovalMedianTask and BaselineRemovalPolynomialTask operate on SPIRE Level-1 data to remove a baseline from the data. This is necessary due to the residual baseline that remains in the Level-1 data, which can produce stripes in the map if not removed. For each bolometer, the signal timeline is taken, either for each scan line separately, or for the whole observation as one long timeline. Then either the median value is subtracted (BaselineRemovalMedianTask?), or a polynomial is fit to the data and subtracted. These tasks can be thought of as a "lite" version of the DestriperTask: they are simpler and faster, but for some observations containing a lot of structure (such as Galactic cirrus), there will be significant improvement if the DestriperTask is used.

Documentation

  • Pipeline Specification Manual: Entries on spectroscopy have been reviewed, examples have been validated.


blog comments powered by Disqus

Edit | Attach | Printable | Raw View | Backlinks: Web, All Webs | History: r43 < r42 < r41 < r40 < r39 | More topic actions
 
Herschel Twiki