IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Changeset 13869


Ignore:
Timestamp:
Jun 18, 2007, 5:34:03 PM (19 years ago)
Author:
eugene
Message:

more updates

File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/doc/manual/manual.tex

    r13847 r13869  
    1  %%% $Id: manual.tex,v 1.7 2007-06-15 20:52:48 eugene Exp $
     1 %%% $Id: manual.tex,v 1.8 2007-06-19 03:34:03 eugene Exp $
    22\documentclass[panstarrs,spec]{panstarrs}
    33
     
    275275processing. \note{halt does not current stop pcontrol}. 
    276276
    277 \subsubsection{ippTools}
     277\subsubsection{ippTools \& ippScripts}
     278
     279The IPP uses a \code{mysql} database to keep track of the data to be
     280analysed and the results of analysis steps.  The IPP user interacts
     281with this database via a set of programs called \code{ippTools} and
     282\code{ippScripts}.  The difference between these is that the
     283\code{ippTools} are somewhat lower-level programs to interact directly
     284with the database, while the ippScripts perform a more complex
     285operation, sometimes wrapping multiple calls to commands within
     286\code{ippTools}.   
     287
     288\begin{itemize}
     289\item \code{ipp_serial_inject.pl} : inject data into the IPP database.
     290\item \code{dettool -definebyquery} : define a new detrend run.  With
     291  this command, the user defines a set of criteria which define a
     292  collection of input images to be used to generate a detrend image.
     293  The parameters which define the output detrend image also
     294  specified.
     295\item \code{dettool -updatedetrun} : manually set the state of a
     296  detrend run. 
     297\item \note{extend this list...}
     298\end{itemize}
     299
     300The ippTools accept a common set of command-line arguments which
     301modify their behavior:
     302\begin{itemize}
     303\item \code{-pretend} : show the expected result, but do not apply the
     304  results to the database
     305\item \code{-simple} return results with one line for each output
     306  component.  Without this flag, the results are returned with the
     307  versbose, but self-documenting, psMetadataConfig format.
     308\item \code{-dbname} specify the database to use
     309\item \note{extend this list...}
     310\end{itemize}
    278311
    279312\subsection{Operational Scenarios}
    280313
    281314\subsubsection{Reducing an Observing Run}
     315
     316For a single observing run, the user will typically have data from
     317several nights, both science and raw detrend images.  The following
     318steps illustrate the analysis with data from a sample observing run.
     319We assume the user has already installed the IPP software and the
     320external software (mysql, apache, php, etc), set up the configuration
     321for the camera, defined a project for this analysis, started pantasks,
     322and loaded the analysis tasks.  This example shows the steps to build
     323the detrend data and process the science data.
     324
     325The first step is to inject all of the data into the IPP database with
     326the command \code{ipp_serial_inject.pl}.  It is possible to define
     327abstract data storage paths in the user's \code{.ipprc} file in the
     328\code{DATAPATH} section. Each line gives the relationship between an
     329abstract data location and the real-world UNIX path to that data.  The
     330script \code{ipp_serial_inject.pl} will interpret the location of the
     331data in terms of the DATAPATH hierarchy when identifying the data for
     332the database. When injecting the data, it is convenient to define a
     333working directory with the \code{--workdir} option to
     334\code{ipp_serial_inject.pl}.  The science processing steps will use
     335this location to store the output data products.
     336
     337% complete command here
     338
     339As the images are injected, they will be placed in the table of
     340\code{new} exposures.  If \code{pantasks} is running with the modules
     341loaded, the images will be processed for registration.  In this step,
     342a minor analysis of the image statistics and an examination of the
     343header information is performed, and the results used to provide
     344further information about the images in the database.  As images are
     345registered, they migrate from the \code{new} to the \code{raw}
     346tables.  These tables can be view in the \code{ippMonitor} under the
     347\code{Load & Setup} pages.
     348
     349Once the images have been injected, the user will want to create, in
     350order, bias, dark, shutter, flat, and fringe frames.  As these are
     351created, they may be used to generate the next level of image.  It is
     352important not to define, eg, a flat run before the lower-precedence
     353elements are defined.  At the moment, it is not possible to define a
     354detrend analysis which blocks until a previous detrend analysis is
     355complete.  To start, define a bias based on a query to the database:
     356
     357% example command hrere
     358
     359\code{dettool -dbname name -definebyquery -det_type bias etc}
     360
     361For a small observing run, it is probably possible to include all bias
     362images in the initial pass.  IPP will iteratively reject complete
     363input images which are outliers from the ensemble, as well as
     364individual pixels.  In the \code{ippMonitor}, rejected images will be
     365greyed out in the residual pages.  Examine the resulting residual
     366images.  If there are patterns to the residuals, it may be necessary
     367to define subgroups, perhaps by time or ccd temperature.  For example:
     368
     369% example command here.
     370
     371After a bias or set of biases are defined, follow the same process for
     372a dark frame.  It may be necessary to supply a non-linear dark current
     373correction (perhaps one for each amplifier or chip).  The script
     374\code{ipp_darkstats.pl} can be used to examine this trend. 
     375
     376If a shutter correction is necessary, make sure to obtain data which
     377can be used to generate the correction: a series of flat-field images
     378with a range of exposure times, particularly near the short end of the
     379range.  If the shutter correction will be skipped, set the SHUTTER
     380entries in the camera's ppImage.config file to FALSE.
     381
     382A flat-field image is next.  Initially, we must define a raw
     383flat-field image, from dome or twilight flat-field images.  Below, we
     384will discuss the generation of a flat-field correction image from
     385dithered science images.  Do not forget to specify the filter for each
     386flat-field image
     387
     388% example command.
     389
     390Up to now, we have been ignoring the bad pixel mask.  At this point,
     391the user may generate a bad pixel mask from the collection of
     392processed flat-field images.  Use the script \code{UNKNOWN} to build a
     393mask image.  Turn on the use of the mask by setting \code{MASK} to
     394true in the \code{ppImage.config} file.  It is probably a good idea to
     395now re-run the previous stages with the mask applied: this will result
     396in better statistics for the residual images.
     397
     398If certain filters require a fringe image, the user may inform the IPP
     399of this fact by setting the FRINGE.FILTERS entry in the
     400\code{ppImage.config} file.  Use a set of night-time sky images to
     401generate a fringe frame. \note{how to specify more than one fringe
     402  mode?}
     403
     404Activate the chip-level science analysis tasks in pantasks: chip.on.
     405With this off, no data will be processed before the detrend images are
     406ready.  Once this is turned on, the science images will start to be
     407processed \note{discuss blocks; re-queue; versions}.
     408
     409\note{specifying the dvo output database}
    282410
    283411\subsubsection{Pipeline Reduction : CFHT Queue Runs}
     
    498626\subsection{psconfig}
    499627
     628\note{use the write-up in psconfig for this section}
     629
     630\begin{verbatim}
    500631* pscheckperl : search for and install, if needed, external Perl modules
    501632* pschecklibs : search for and install, if needed, external C libraries
     
    504635* psdist : build IPP distributions (requires CVS access)
    505636* tagsets : tables defining the C and Perl components to be built
     637\end{verbatim}
    506638
    507639The IPP is a large and complex software system.  A major goal of the
     
    13391471\subsection{missing libX11.so}
    13401472
    1341 RedHat RHEL 3 & 4 (likely other RedHat variants as well) for \emph{amd64} seem
     1473RedHat RHEL 3 \& 4 (likely other RedHat variants as well) for \emph{amd64} seem
    13421474to often be installed without a \code{libX11.so} under
    13431475\code{/usr/X11R6/lib64/}.  This will cause 64bit builds trying to link against
     
    13471479yourself.  Eg..
    13481480
    1349 \begin{verbatium}
     1481\begin{verbatim}
    13501482su -
    13511483cd /usr/X116/lib64
    13521484ln -s libX11.so.6.2 libX11.so
    1353 \end{verbatium}
     1485\end{verbatim}
    13541486
    13551487or with \code{sudo}:
    13561488
    1357 \begin{verbatium}
     1489\begin{verbatim}
    13581490cd /usr/X116/lib64
    13591491sudo ln -s libX11.so.6.2 libX11.so
    1360 \end{verbatium}
     1492\end{verbatim}
    13611493
    13621494
     
    13651497installing IPP software.  Note that \code{\$prefix/lib} will also need to be manually added the enviroment variable \code{LD_LIBRARY_PATH} if your not using jhbuild. E.g for ksh/bash users:
    13661498
    1367 \begin{verbatium}
     1499\begin{verbatim}
    13681500export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/[foo]/[yourinstallprefix]/lib
    1369 \end{verbatium}
     1501\end{verbatim}
    13701502
    13711503
Note: See TracChangeset for help on using the changeset viewer.