Differences

This shows you the differences between two versions of the page.

--- 2015:groups:tools:lhaad:proposal2 [2015/06/18 15:21]
nishita.desai created
+++ 2015:groups:tools:lhaad:proposal2 [2015/06/22 15:36] (current)
daniel.schmeier
@@ Line 1: / Line 1: @@
+Modifications:
+(19-Jun)
+  - Removed end blocks
+  - Added function arguments (e.g arg1, arg2,..)
+  - Renamed SR to CUT
+  - Inheritance explicitly defined by key 'contains'
+  - Changed cut language to select/reject
+  - Added new info block "units"
+The idea is to write a plain text file that can be read easily into
+xml if needed (and hence universally readable).  The name of the block become sthe xml tag.  Anything that follows is turned into attributes. e.g. "name ATLAS" inside the block becomes <experiment name="ATLAS">.  Anything inside the block starting with "#" is added as a comment, i.e. the block
+<code autoconf>
+info analysis
+  # Details about experiment
+  id ATLAS-2014-XXX
+  publication JHEP11(2014)118
+  sqrtS 8.0
+  lumi 20.0
+</code>
+converts to
+<code autoreconf>
+<info name="analysis" id="ATLAS-2014-XXX" publication="JHEP11(2014)118" sqrtS="8.0" lumi="20.0" description="Details about analysis">
+</code>
+There are five types of pre-defined blocks:
+  - **detectorobj** (defines reconstructed objects)
+  - **function** (defines functions that act on these)
+  - **cut** (blocks of cuts)
+  - **binning** (following Daniel's idea, for complex descriptions of orthogonal signal regions)
+  - **info** (for meta information chosen by the implementer, e.g. information about publication, conf-notes, theory papers used for interpretation, benchmarks, etc.)
+For presenting results, we need two more blocks, e.g.
+  - **results** (for expected, observed, S95, etc.)
+  - **table** (for providing digitised histograms etc.)
+For any block, a link to code or external documentation can be provided.  We can also include a MCsamples block to document the programs used to generate signal and background samples.
+====== Part I: Preamble (define detector objects, variables) ======
 <code autoconf>
-# The idea is to write a plain text file that can be read easily into
-# xml (and hence universally readable).  The name of the block becomes
-# the xml tag.  Anything that follows is turned into
-# attibutes. e.g. "name ATLAS" inside the block becomes <experiment
-# name="ATLAS">.  Anything inside the block starting with "#" becomes
-# a comment, i.e. the first block below becomes <analysis
-# name="ATLAS" analysis="ATLAS-2014-XXX" cme="8.0" lumi="20.0"
-# description="Details about analysis">
-# DETECTOROBJ: JET, ELEC, MUON, MET
-# BASE-VARIABLES: PX, PY, PZ, E, PT, ETA, PHI, BETA
-# USER-DEFINED VARS: MT(), MT2(), DPHIJMET(J, MET)
-analysis
+info analysis
   # Details about experiment
-  name ATLAS
+  id ATLAS-2014-XXX
-  analysis ATLAS-2014-XXX
   publication JHEP11(2014)118
-  cme 8.0
+  sqrtS 8.0
   lumi 20.0
-end anslysis
+info units
+  energy GeV
+  length mm
+  xsec pb
+  # If using function repository, need to clarify if phi is between 0-2pi or -pi to pi.
 detectorobj jet
@@ Line 29: / Line 61: @@
   ptmin 20
   etamax 2.5
-  pileup_sub _link_
   # Can provide code and/or link to writeup with details of implementation
-  code = _code_link
+  code _code_link
   doc _link_to_file_with_details
-end detectorobj jet
+  doc _link_to_doc_about_pilup_removal
-function isolation 2
+function isolation
-  # Human readable descripttion: returns (sum of pT in cone of 0.1 around
+  # Takes 2 arguments -- object, conesize; returns (sum of pT in cone of 0.1 around
   # electron)/(pt of lepton) takes in lepton and conesize
-  code = _code_link
+  arg1 detectorobj
+  arg2 float
+  return float
+  code _code_link
   doc _link_to_file_with_details
-end function
-function overlap_removal 2
+function overlap_removal
-  # takes in jet list and lepton, calculates based on recipe
+  # takes in jet list and lepton, reduces them to non-overlapping objects
-  code = _code_link
+  arg1 detectorobj-list
+  arg2 detectorobj-list
+  code _code_link
   doc _link_to_file_with_details
-end function ovelap_removal
 detectorobj elec_l
   # Can provide code and/or link to writeup with details of implementation
   cut isolation<0.1
-  cut overlap_removal=true
+  cut overlap_removal
-  code = _code_link
+  code _code_link
   doc _link_to_file_with_details
-end detectorobj elec_l
-detectorobj elec_m elec_l
+detectorobj elec_m
   # electron-medium inherits from loose.
   # Can provide code and/or link to writeup with details of implementation
-  code = _code_link
+  contains elec_l
+  code _code_link
   doc _link_to_file_with_details
-end detoctorobj elect_t
-detectorobj lep elec_m mu_m
+detectorobj lep
   # lep inherits from medium electrons or muons
-end detectorobj
+  contains elec_m
+  contains mu_m
 detectorobj met
@@ Line 71: / Line 105: @@
   code _code_link
   doc _link_to_file_with_details
-end detectorobj met
-trigger e_trig
+cut e_trigger
   # Human readable info
   # Each signal analysis will need atleast one trigger to be satisfied
@@ Line 80: / Line 113: @@
   code _code_link
   doc _link_to_file_with_details
-end trigger e_trig
-trigger m_trig
+cut mu_trigger
   # human readable info
   eff 0.8
   code _code_link
   doc _link_to_file_with_details
-end trigger m_trig
-trigger met_trig
+cut met_trigger
   # human readable info
   eff 0.8
   code _code_link
   doc _link_to_file_with_details
-end trigger met_trig
-function pt 2
+function mt
-  # p_T is a function (wrapper) that takes two inputs (1) name of list
+  # mT requires one lepton and met i.e. mt(lep.1,met)
-  # of detectorobj and (2) the index in that list.  E.g. pt(jet2) =
+  arg1 detectorobj
-  # pt(jets,2); pt(elec_tight,1) = pt(elec_t,1) In principle, all
+  arg2 detectorobj
-  # common p(), eta(), phi() etc. need not be explicitly defined, this
+  return float
-  # is just an example
   code _code_link
   doc _link_to_file_with_details
-end function pt
-function mt 2
-  # mT requires one lepton and met i.e. mt(lep(1),met)
-  code _code_link
-  doc _link_to_file_with_details
-end function mt
-function mt2 3
-  code _code_link
-  doc _link_to_file_with_details
-end function mt2
-function b-tag 1
+function b-tag
   # takes in a jet and returns whether it is b-tagged
+  arg1 detectorobj
+  return bool
   eff 0.7
   code _code_link
   doc _link_to_file_with_details
-end function b-tag
-function select 2
+function select
   # select(listname, criterion)
   # e.g. select(jet, b-tag) returns number of objects (i.e. jets) give
   # true for the test (i.e. b-tag).  Should be made a global function like pT?
+  arg1 detectorobj-list
+  arg2 function
+  return int
   code _code_link
   doc _link_to_file_with_details
-end function select
-MCsamples
-  # Human readable info regarding what programs were used to make what
-  # sample suggested format: Signal program extra details (extras are
-  # stored in description for xml
-  signal Sherpa1.4.1 stop-pair
-  background POWHEG ttbar mtop = 172.5 NLO
-  doc _link_to_file_with_details
-end MCsamples
-cut cleaning
-  # List of preselection cuts: Give human readable descriptions (for
-  # your collaborators) and efficiency for theorists
-  code _link_to_code
-  doc _link_to_doc
-end cut
-cut preselect
-  # Pre-selection cuts
-  eff 0.95
-  trigger trigger_e.OR.trigger_mu.OR.trigger_met
-  cut pt(lep(1))>25
-  cut size(lep)=1
-  cut size(jet)>2
-  cut met>100
-end preselect
 detectorobj hadtop
   # Define hadtop
-end detectorobj hadtop
-function deltaR 2
+function deltaR
   # takes in two objects; returns deltaR
-end function deltaR
+  arg1 detectorobj
+  arg2 detectorobj
+  return float
-detectorobj bjet
-  #
-  eff 0.7
-  code _code_
-  doc _doc_link_
-end detectorobj bjet
 function HT
   code _code_
   doc _doc_link_
-end function HT
 function metBYsqrtHT
   # returns MET/sqrt(HT)
-end function metBYsqrtHT
+  return float
-function delPhi 2
+function delPhi
   # returns minimum delta phi between objects
-end function delPhi
+  arg1 detectorobj
+  arg2 detectorobj
+  return float
+</code>
-SR tN_diag
+====== PART 2: uses the objects defined above to define the signal regions ======
-  cut preselect
-  cut size(jet)>4
+<code autoconf>
-  cut pt(jet,1)>60
-  cut pt(jet,2)>60
+cut cleaning
-  cut pt(jet,3)>40
+  # List of preselection cuts: Give human readable descriptions (for
-  cut pt(jet,4)>25
+  # your experimental collaborators) and efficiency for theorists (if applicable)
-  cut mt(lep(1),met)>=60
+  code _link_to_code
-  cut m(hadtop)>=130
+  doc _link_to_doc
-  cut m(hadtop)<=205
-  cut size(tau)=0
+cut preselect
-  cut deltaR(bjet(1),lep(1))<2.5
+  # Pre-selection cuts
-  cut metBYsqrtHT>5
+  eff 0.95
-  cut delPhi(jet(1),met)>0.8
+  trigger e_trigger.OR.mu_trigger.OR.met_trigger
-  cut delPhi(jet(2),met)>0.8
+  select lep.1.pt>25
-  hist SHAPE_SR1
+  reject lep.size>1
-end SR
+  select jet.size>2
+  select met>100
+cut tN_diag
+  contains preselect
+  select jet.size>4
+  select jet.1.pt>60
+  select jet.2.pt>60
+  select jet.3.pt>40
+  select jet.4.pt>25
+  select mt(lep.1,met)>=60
+  select m(hadtop)>=130
+  select m(hadtop)<=205
+  reject size(tau)>0
+  select deltaR(bjet.1,lep.1)<2.5
+  select metBYsqrtHT>5
+  select delPhi(jet.1,met)>0.8
+  select delPhi(jet.2,met)>0.8
+</code>
+====== Part 3: Present results of the analysis.  Give numbers, exclusions, digitised histograms, link to plots, etc. ======
+We need to think about information that is absolutely necessary for validation and whether to put it in the same kind of info block (which are currently entirely optional).
+<code autoconf>
+info MCsamples
+  # Human readable info regarding what programs were used to make what
+  # sample suggested format: Signal program extra details (extras are
+  # stored in description for xml
+  signal Sherpa1.4.1 stop-pair
+  background POWHEG ttbar mtop = 172.5 NLO
+  doc _link_to_file_with_details
 result tN_diag
@@ Line 214: / Line 240: @@
   obs
   s95
-  hist SHAPE1
+  table data1
-  hist SHAPE2
-end result tN_diag
+table data1
+  # Not real data in example; could be used to provide histograms/tables/likelihoods
+  bins nn
+  data xlabel ylabel yerr
+.0 2.0 1.0
+.0 1.0 1.0
+.0 10.0 1.0
+.0 15.0 1.0
 </code>

Les Houches

User Tools

Site Tools

Differences

Page Tools