(r19) AMC13RawDataAnalysis < BUCMSPublic

Tags: view all tags
---++ *New analysis code being developed for 2014 data format*

New code is being developed by Daniel Arcaro with help from Eric Hazen and Dan Gastler to analyze
the new/final data format documented here: AMC13CommonFirmwareProposal.
Also Ted Laird has written a tool to convert to root format:

<pre>
P.S. If you end up dumping the AMC13 monitor buffers, but would like
to convert the dumps to .root files with TTrees of
std::vector<uint64_t>, so that you can use "ordinary" analysis
programs to analyze them, feel free to use this program:
<a href="https://github.com/elaird/hcalraw/blob/master/cpp/badcoffee.cpp">badcoffee.cpp</a>

Bug reports are quite welcome.  (to Ted Laird)
</pre>

---+++ *Finding the Data*

See [[https://twiki.cern.ch/twiki/bin/view/Main/EricFindingCMSData][EricFindingCMSData]] for some help.

---+++ *Converting Data*

The two options for file types processed by this code are binary and root files. The code to switch from one to another or vice versa is here:

Binary to Root: https://github.com/elaird/hcalraw (From Ted Laird)
   * The main tool for making badcoffee binary files is hcalraw/cpp/badcoffee.cpp. Usage is cat a.dat | ./badcoffee fed#.

Root to Binary: http://edf.bu.edu/svn/edf/arcaro/cms/trunk/pythontools/%GREEN%rootdumper/%ENDCOLOR%  or http://edf.bu.edu/svn/edf/arcaro/cms/trunk/analyzer/%PURPLE%dumpraw/%ENDCOLOR%
   * The %PURPLE%dumpraw%ENDCOLOR% code is a cmssw analyzer and requires a cmssw environment. The %GREEN%rootdumper%ENDCOLOR% only requires root.
   * The %GREEN%rootdumper%ENDCOLOR% is built by running the Makefile in rootdumper/cpp. 
      * Run the code using ./oneRun.py --rootFile=HTB_007850.root --feds=928 --nEvents=101 --outputFile=data1.dat
   * The %PURPLE%dumpraw%ENDCOLOR% analyzer can be built with the same directions as the CMSSW section below.
      * Run the code using cmsRun .../python/ConfFile_cfg.py


---+++ *Unpacker*

Unpacker consists of three classes to parse AMC13 data format: FedEvent, FedBlock, and FedAMC. Various tools have also been created to utilize these classes and analyze the data. All code can be found here: http://edf.bu.edu/svn/edf/arcaro/cms/trunk/

Class documentation: http://ohm.bu.edu/~djarcaro/Documentation/unpacker.pdf 

---++++ %BLUE% Unpacker Build and Run Instructions %ENDCOLOR%
Check out and build:
<verbatim>
$svn co http://edf.bu.edu/svn/edf/arcaro/cms/trunk/unpacker and http://edf.bu.edu/svn/edf/arcaro/cms/trunk/tools
$cd tools
$make
</verbatim>
List of tools and their uses:
   * dump64: Prints the raw data to screen in hex with 64bits per line. Used to sort through data by hand.
      * use:  $ ./dump64 file.dat (usually followed by | less)
   * dump: Main dump tool to parse events and check for errors.
      * use: $ ./dump file.dat -v 1 -e 1 -c
      * -v flag for amount of information about the event printed to screen
      * -c flag to turn CRC checking on or off
      * -e flag to print specific errors to screen

---++++ %BLUE% Unpacker histogram option %ENDCOLOR%
If root is installed dumptmp.cc can be made by $make root=1. dumptmp is identical to dump but also creates a root file with a histogram of the error array for easier visualization with large files. 

---+++ *CMSSW*
The dump software has also been implemented in a cmssw analyzer to extract raw data from global/local root files and unpack it. This requires a cmssw environment (most likely lxplus). The code can be found here http://edf.bu.edu/svn/edf/arcaro/cms/trunk/analyzer/ and .../trunk/cmsswtools. General instructions to make an analyzer can be found here: https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookWriteFrameworkModule.

Recipe to build the analyzer code:
<verbatim>
$ cmsrel CMSSW_#_#_#    ($ scram list   if you want to see the version choices) (The version should match the one that created the root file. Use edmProvDump name.root to check the version.) 
$ cd CMSSW_#_#_#/src
$ cmsenv (equivalent: eval `scramv1 runtime -csh`)
$ svn co http://edf.bu.edu/svn/edf/arcaro/cms/trunk/analyzer/
$ svn co http://edf.bu.edu/svn/edf/arcaro/cms/trunk/cmsswtools
$ scram b
</verbatim>

To run the analyzer first set the parameters in the python file analyzer/parse/python/ConfFile_cfg.py. The most important parameters (besides the file name) are the HcalTBSource vs PoolSource and the rootType string input. Change these to correspond to either a local or global root file. The python script is set to run over multiple files in a directory. To then run the analyzer:

<verbatim>
$ cmsRun analyzer/parse/python/ConfFile_cfg.py.
</verbatim>
 
All of this should be done using the same version of CMSSW as what the root file was produced with. To determine this use $ edmProvDump rootfile.root to see the CMSSW version number and other attributes of the file. There are also other edm functions that are quite useful (of course only on lxplus).  

---+++ *Bugs*

---+++ *To Do List*

   * Add a doxygen configuration file and "doc" target to makefile so one can do 'make doc' and get the doxygen output.

---+++ New Features

---+++ DONE:

   * In *FedEvent::PreParse* the expression =if(GetNAMC() > 0 || GetNAMC() < 13)= is always true.  I think you mean to use =&&= instead of =||=.
   * In *FedEvent::PreParse* there are two possibilities for size: a) 2 and b) 4...14.  It should throw an exception if any other value is passed.  The comments should reflect this.

   * Check for consistency of LMSEPVC bits through an entire event
      * MS should be "10" for first block, "11" for intermediate blocks, "01" for final blocks.
      * C should be 1 when MS="01".
      * EP should always be "11" I think
   * Check Blk_No, AmcNo, BoardID in block headers for consistency throughout event
   * Add option to dump entire AMC payloads


-- Main.EricHazen - 04 Aug 2014
Attachments
Topic attachments
I	Attachment	History	Action	Size	Date	Who	Comment
pdf	unpacker.pdf	r1	manage	125.4 K	14 Aug 2015 - 16:00	DanielArcaro	Documentation for classes in unpacker
Topic revision: r19 - 23 Oct 2015 - DanielArcaro