File: CpptrajDevelopmentGuide.lyx

package info (click to toggle)
cpptraj 5.1.0%2Bdfsg-4
links: PTS, VCS
area: main
in suites: sid, trixie
size: 138,900 kB
sloc: cpp: 145,621; ansic: 34,635; sh: 11,365; f90: 971; makefile: 770; awk: 242
file content (3977 lines) | stat: -rw-r--r-- 90,747 bytes
#LyX 2.3 created this file. For more info see http://www.lyx.org/
\lyxformat 544
\begin_document
\begin_header
\save_transient_properties true
\origin unavailable
\textclass article
\use_default_options true
\maintain_unincluded_children false
\language english
\language_package default
\inputencoding auto
\fontencoding global
\font_roman "default" "default"
\font_sans "default" "default"
\font_typewriter "default" "default"
\font_math "auto" "auto"
\font_default_family default
\use_non_tex_fonts false
\font_sc false
\font_osf false
\font_sf_scale 100 100
\font_tt_scale 100 100
\use_microtype false
\use_dash_ligatures false
\graphics default
\default_output_format default
\output_sync 0
\bibtex_command default
\index_command default
\paperfontsize default
\spacing single
\use_hyperref true
\pdf_bookmarks true
\pdf_bookmarksnumbered false
\pdf_bookmarksopen false
\pdf_bookmarksopenlevel 1
\pdf_breaklinks false
\pdf_pdfborder false
\pdf_colorlinks false
\pdf_backref false
\pdf_pdfusetitle true
\papersize default
\use_geometry false
\use_package amsmath 1
\use_package amssymb 1
\use_package cancel 1
\use_package esint 1
\use_package mathdots 1
\use_package mathtools 1
\use_package mhchem 1
\use_package stackrel 1
\use_package stmaryrd 1
\use_package undertilde 1
\cite_engine basic
\cite_engine_type default
\biblio_style plain
\use_bibtopic false
\use_indices false
\paperorientation portrait
\suppress_date false
\justification true
\use_refstyle 0
\use_minted 0
\index Index
\shortcut idx
\color #008000
\end_index
\secnumdepth 3
\tocdepth 3
\paragraph_separation indent
\paragraph_indentation default
\is_math_indent 0
\math_numbering_side default
\quotes_style english
\dynamic_quotes 0
\papercolumns 1
\papersides 1
\paperpagestyle default
\tracking_changes false
\output_changes false
\html_math_output 0
\html_css_as_file 0
\html_be_strict false
\end_header

\begin_body

\begin_layout Title
CPPTRAJ Development Notes
\end_layout

\begin_layout Author
Daniel R.
 Roe (daniel.r.roe@gmail.com)
\begin_inset Newline newline
\end_inset

Jason M.
 Swails (Code Docs)
\end_layout

\begin_layout Date
2010-07-21
\begin_inset Newline newline
\end_inset

Last Updated: 2019-12-13
\begin_inset Newpage pagebreak
\end_inset


\end_layout

\begin_layout Abstract
CPPTRAJ is code used for processing MD trajectory data as well as other
 types of data, derived from trajectories or otherwise.
 CPPTRAJ is a complete rewrite of the PTRAJ code in primarily C++, with
 the intent being to make the code more readable, leak-free, and thread-safe.
 The biggest functional change from PTRAJ is the ability to load and process
 trajectories with different topology files in the same run.
\end_layout

\begin_layout Abstract
This guide assumes that the reader has at least a basic familiarity with
 C and C++ object-oriented programming.
 If you aren't sure what a constructor is or how pointers work you may have
 a difficult time coding in Cpptraj.
 There are several good introduction to C/C++ tutorials on the web that
 may be helpful.
 
\begin_inset Newpage pagebreak
\end_inset


\end_layout

\begin_layout Abstract
\begin_inset CommandInset toc
LatexCommand tableofcontents

\end_inset


\begin_inset Newpage pagebreak
\end_inset


\end_layout

\begin_layout Part
Introduction
\begin_inset CommandInset label
LatexCommand label
name "part:Coding-Conventions"

\end_inset


\end_layout

\begin_layout Section
Coding Conventions
\end_layout

\begin_layout Standard
It is important to maintain a consistent coding style within cpptraj so
 that it remains easy to modify and understand.
 By following code conventions, it will be easier to read code written by
 anybody and determine what is happening.
\end_layout

\begin_layout Itemize
Code blocks are indented using 2 spaces.
 
\series bold
DO NOT USE TABS
\series default
 since these are in general not portable between different editors.
\end_layout

\begin_layout Itemize
Try to keep the maximum length of lines between 80 and 100 characters long.
\end_layout

\begin_layout Itemize
Whenever possible, put separate code on separate lines.
 Exceptions can be made for very simple statements such as logic evaluations
 and simple initializations.
 For example,
\end_layout

\begin_layout LyX-Code
double x1 = 0.0; x2 = 0.0; x3 = 0.0;
\end_layout

\begin_layout Standard
is OK, but
\end_layout

\begin_layout LyX-Code
double x1 = var1 * var2; double x2 = var3 / var4;
\end_layout

\begin_layout Standard
is not.
 There are two reasons: 1) When separate statements share a line it makes
 using debuggers more difficult, and 2) when separate statements share a
 line it is harder to read.
\end_layout

\begin_layout Itemize
C++ files have '.cpp' suffix, C files have '.c' suffix, header files have
 '.h' suffix.
\end_layout

\begin_layout Itemize
All header files should have a '#define' guard to prevent multiple inclusion.
 The define guard has format:
\end_layout

\begin_layout LyX-Code
#ifndef INC_<basefilename>_H
\end_layout

\begin_layout LyX-Code
#define INC_<basefilename>_H
\end_layout

\begin_layout LyX-Code
...
\end_layout

\begin_layout LyX-Code
#endif
\end_layout

\begin_layout Itemize
'using namespace' should be used sparingly and NEVER in a header file.
\end_layout

\begin_layout Itemize
The order of #include directives should be (in general): C includes, C++
 includes, class definition, any other Cpptraj includes.
\end_layout

\begin_layout Itemize
Use of STL classes/methods is acceptable; use C99 conventions to maximize
 portability.
 The only external libraries that should be used are NetCDF and ARPACK/LAPACK/BL
AS (both included with AmberTools), i.e.
 no Boost etc.
\end_layout

\begin_layout Itemize
Do not use iostream for basic IO.
 All console output should be performed with the functions in CpptrajStdio.h
 (chiefly mprintf() and mprinterr() for STDOUT and STDERR respectively).
 All file IO should be performed with CpptrajFile or the derived classes
 BufferedLine and BufferedFrame.
 This choice has been made mainly for performance reasons (C file routines
 are in general much faster than iostream), but also so that all IO is centraliz
ed (e.g.
 CpptrajFile will automatically detect if an input file is compressed).
 This is also so output can be easily controlled; for example, using mprintf
 will make sure that during MPI only the master writes.
\end_layout

\begin_layout Itemize
Warnings should be written to STDOUT with mprintf with prefix 'Warning:';
 errors should be written to STDERR with mprinterr with prefix 'Error:'.
\end_layout

\begin_layout Itemize
Classes:
\begin_inset Separator latexpar
\end_inset


\end_layout

\begin_deeper
\begin_layout Itemize
Class types are named using 
\emph on
CapWords
\emph default
 (no spaces or underscores, start of each word is a capital letter).
\end_layout

\begin_layout Itemize
Files containing a class should be named after the class (e.g.
 'class TrajectoryFile {};' in TrajectoryFile.cpp).
\end_layout

\begin_layout Itemize
Classes which inherit should be named after their base class (e.g.
 'class Action_Distance : public Action { };').
\end_layout

\begin_layout Itemize
Public class methods should be listed first; protected methods/variables
 second; private methods/variables last.
 All class member variables should be private if possible.
\end_layout

\begin_layout Itemize
Public class methods are named using 
\emph on
CapWords
\emph default
.
\end_layout

\begin_layout Itemize
Private class methods are named using 
\emph on
mixedCase.
\end_layout

\begin_layout Itemize
Class variables that are 
\family typewriter
private
\family default
 or 
\family typewriter
protected
\family default
 are named using 
\emph on
mixedCase_
\emph default
 (with a trailing underscore).
\end_layout

\end_deeper
\begin_layout Itemize
Abbreviations: 1st letter in each word is capitalized.
 For instance, 
\family typewriter
Data File List
\family default
 may be abbreviated 
\family typewriter
DatFilList
\family default
 or 
\family typewriter
DFL.
\end_layout

\begin_layout Itemize
Variables that have function scope (or lower) and all public variables for
 classes are named using 
\emph on
mixedCase
\emph default
 (same as 
\emph on
CapWords
\emph default
 except the first letter is lower-case).
\end_layout

\begin_layout Itemize
No one-letter variable names except in loop scopes (e.g.
 for (int i = 0; i < N; ++i) { } ), and even then they should be short loops
 (no more than 10 lines or so).
\end_layout

\begin_layout Itemize
All identifiers in an enumerated type are named using all 
\emph on
CAPS
\emph default
, and the first identifier should be explicitly initialized (e.g.
 enum DirectionType { DX = 0, DY, DZ };).
\end_layout

\begin_layout Itemize
There is a 
\emph on
doxygen
\emph default
 rule file to automatically generate code documentation using 
\emph on
doxygen
\emph default
, so please construct comments in such a doxygen-compatible manner (e.g.
 JavaDoc etc).
 See http://www.stack.nl/~dimitri/doxygen/manual.html for instructions.
\end_layout

\begin_layout Subsection
Versioning
\end_layout

\begin_layout Standard
The internal versioning for CPPTRAJ is supposed to go like this: 
\end_layout

\begin_layout LyX-Code
V<major>.<minor>.<revision> 
\end_layout

\begin_layout Description
<major> Incremented whenever there is a major API change, e.g.
 changing the Action base class, etc.
\end_layout

\begin_layout Description
<minor> Incremented whenever there are changes to behavior, e.g.
 syntax, functionality, or output.
\end_layout

\begin_layout Description
<revision> All other changes (so at least each pull request).
\end_layout

\begin_layout Standard
Whenever a number that precedes <revision> is incremented, all subsequent
 numbers should be reset to 0.
\end_layout

\begin_layout Section
Building Cpptraj and Documentation
\end_layout

\begin_layout Standard
Cpptraj is automatically built as part of AmberTools, or it can be built
 standalone using the configure script in the 
\family typewriter
$AMBERHOME/AmberTools/src/cpptraj
\family default
 or 
\family typewriter
$CPPTRAJHOME
\family default
 directory.
 The standalone build is particularly useful for development and testing.
 Type './configure –help' for a list of configure options.
 In order to build Cpptraj standalone one needs to specify the location
 of the NetCDF, zlib, bzlib2, and BLAS/LAPACK/ARPACK libraries if they aren't
 in your system path; configure will use the ones in $AMBERHOME if 
\family typewriter
'-amberlib'
\family default
 is specified.
 The -noX options can be used to disable use of certain libraries.
 
\end_layout

\begin_layout Standard
For example, to build cpptraj standalone:
\end_layout

\begin_layout LyX-Code
./configure -amberlib gnu
\end_layout

\begin_layout LyX-Code
make install OR cd src && make install
\end_layout

\begin_layout Standard
To build the documentation using 
\emph on
doxygen
\emph default
, you must have 
\emph on
doxygen
\emph default
 installed, and you must have configured AmberTools.
 Run the command:
\end_layout

\begin_layout LyX-Code
make docs
\end_layout

\begin_layout Standard
to build the documentation.
 PDF files and HTML files are generated during this process, showing class
 inheritance and descriptions from comments written in doxy-format.
 Open the file 
\family typewriter
$AMBERHOME/AmberTools/src/cpptraj/doc/html/index.html
\family default
 to see the class heirarchy and descriptions.
\end_layout

\begin_layout Part
General Layout and Concepts
\end_layout

\begin_layout Standard
Cpptraj currently lives in 3 key classes:
\end_layout

\begin_layout Description

\series bold
Cpptraj
\series default
 Defined in main.cpp, controls overall flow (e.g.
 it is responsible for deciding whether to execute in batch mode or interactive
 mode).
\end_layout

\begin_layout Description

\series bold
CpptrajState
\series default
 Defined in 
\series bold
Cpptraj
\series default
, holds all of the data, Actions, Analyses, etc.
 
\end_layout

\begin_layout Description
Command 
\begin_inset Quotes eld
\end_inset

Static
\begin_inset Quotes erd
\end_inset

 class used by 
\series bold
Cpptraj
\series default
 to process user input.
 The file Command.cpp also contains all of the logic for executing commands.
\end_layout

\begin_layout Section
CpptrajState
\end_layout

\begin_layout Standard
The main components of 
\series bold
CpptrajState
\series default
 are:
\end_layout

\begin_layout Description
DataSetList
\begin_inset space ~
\end_inset

DSL_; Hold all DataSets.
 This is essentially how different components can talk to each other, e.g.
 an Action creates a DataSet in the DataSetList, which can then be used
 by a subsequent Analysis.
\end_layout

\begin_layout Description
DataFileList
\begin_inset space ~
\end_inset

DFL_; Hold all DataFiles.
 These are either for writing out DataSets or general text output, primarily
 from Actions/Analyses.
\end_layout

\begin_layout Description
TrajinList
\begin_inset space ~
\end_inset

trajinList_; Hold all input trajectories to be processed during a run.
 Whenever a user inputs a 'trajin' or 'ensemble' command, the trajectory/ensembl
e in question is added to this list.
 When a 'run' command is executed, these are the trajectories that are read
 in a frame at a time so that Actions in the ActionList can process them.
\end_layout

\begin_layout Description
TrajoutList
\begin_inset space ~
\end_inset

trajoutList_; Hold output trajectories to be written during a run.
 This output occurs after all Actions have been processed.
\end_layout

\begin_layout Description
ActionList
\begin_inset space ~
\end_inset

actionList_; Hold all Actions to be executed during a run.
 By default, whenver an Action command is issued the Action in question
 is initialized and queued up in the ActionList, to be processed during
 the next run.
\end_layout

\begin_layout Description
AnalysisList
\begin_inset space ~
\end_inset

analysisList_; Hold all Analyses to be executed after a run or when a 'runanalys
is' command is given.
 Similar to ActionList, whenever an Analysis command is issued the Analysis
 in question is initialized and queued up in the AnalysisList.
\end_layout

\begin_layout Section
Actions
\end_layout

\begin_layout Standard
Actions are how Cpptraj derives data from input trajectories.
 There are two basic classes for Actions:
\end_layout

\begin_layout Subsection
Action
\end_layout

\begin_layout Standard
The abstract base class that defines the Action interface.
 Consists of 4 functions:
\end_layout

\begin_layout Description
Init(): Initialize Action, set up DataSets/DataFiles, etc.
\end_layout

\begin_layout Description
Setup(): Set up Action for a given Topology.
\end_layout

\begin_layout Description
DoAction(): Perform Action on given Frame.
\end_layout

\begin_layout Description
Print(): Perform any post-processing or output that occurs outside the main
 DataSet/DataFile framework.
\end_layout

\begin_layout Subsection
ActionInit (ActionState.h)
\end_layout

\begin_layout Standard
Used to interface with Init(); contains pointers to the master DataSetList
 and DataFileList in CpptrajState.
\end_layout

\begin_layout Description
DataSetList&
\begin_inset space ~
\end_inset

DSL(),
\begin_inset space ~
\end_inset

DataSetList
\begin_inset space ~
\end_inset

const&
\begin_inset space ~
\end_inset

DSL()
\begin_inset space ~
\end_inset

const Reference to the master DataSetList.
 The appropriate version should be used automatically.
\end_layout

\begin_layout Description
DataSetList*
\begin_inset space ~
\end_inset

DslPtr() For Actions that require access to the master DataSetList after
 Init() (e.g.
 hbond, which cannot set up hbond time series until hbonds are actually
 detected in DoAction()), they can store a pointer to the master DataSetList
 like so:
\begin_inset Separator latexpar
\end_inset


\end_layout

\begin_deeper
\begin_layout Standard
(In header): DataSetList* masterDSL_;
\end_layout

\begin_layout Standard
(In Action::Init): masterDSL_ = init.DslPtr();
\end_layout

\end_deeper
\begin_layout Description
DataFileList&
\begin_inset space ~
\end_inset

DFL(),
\begin_inset space ~
\end_inset

DataFileList
\begin_inset space ~
\end_inset

const&
\begin_inset space ~
\end_inset

DFL()
\begin_inset space ~
\end_inset

const Reference to master DataFileList.
 The appropriate version should be used automatically.
\end_layout

\begin_layout Subsection
ActionSetup (ActionState.h) 
\end_layout

\begin_layout Standard
Used to interface with Setup(); contains pointers to current Topology and
 CoordinateInfo, as well as expected number of frames associated with current
 Topology.
\end_layout

\begin_layout Subsection
ActionFrame (ActionState.h)
\end_layout

\begin_layout Standard
Used to interface with DoAction(); contains pointer to current Frame.
\end_layout

\begin_layout Part
Key Classes And Functions
\end_layout

\begin_layout Section
Math-related Classes
\end_layout

\begin_layout Subsection
Vec3
\end_layout

\begin_layout Standard
An array of 3 doubles, used to hold XYZ coords.
 Used for vector math.
\end_layout

\begin_layout Subsection
Matrix_3x3
\end_layout

\begin_layout Standard
A 3x3 array of doubles, useful for performing rotations etc.
 Used for basic matrix math.
 Can be diagonalized via an internal routine (no need for external math
 library).
\end_layout

\begin_layout Subsection
ComplexArray
\end_layout

\begin_layout Standard
Used to hold an array of complex numbers.
 Implemented as a double array instead of using the STL Complex class so
 that it easily interface with external routines.
\end_layout

\begin_layout Subsection
PubFFT
\end_layout

\begin_layout Standard
Interface to FFT routines (either pubfft, which are the FFT routines used
 by Amber, or FFTW depending on how CPPTRAJ is configured).
 Currently only 1D forward and backwards FFTs are supported.
 Makes use of ComplexArray.
\end_layout

\begin_layout Subsection
Corr.h: CorrF_Direct, CorrF_FFT
\end_layout

\begin_layout Standard
Classes used to calculate auto/cross correlation functions from arrays of
 complex numbers (ComplexArray).
\end_layout

\begin_layout Section
Some Key Classes and Functions
\end_layout

\begin_layout Standard
The following is a brief list of some of the more commonly-used classes
 and functions in Cpptraj.
 Classes are more or less self-documented to a certain extent; this section
 will be focused on how these classes are/should be used.
\end_layout

\begin_layout Subsection
ArgList
\end_layout

\begin_layout Standard
The ArgList class is used throughout Cpptraj.
 It is the main way that user input is translated to actions, analyses,
 trajectory IO, etc.
 Basically, the ArgList class takes a string and separates it into tokens
 based on a given delimiter or delimiters.
 For example, the string:
\end_layout

\begin_layout LyX-Code
myString = 
\begin_inset Quotes eld
\end_inset

trajin mytraj.nc 1 100 10
\begin_inset Quotes erd
\end_inset

;
\end_layout

\begin_layout Standard
can be separated via a space (' ') delimeter (the default) into 5 tokens
 like so:
\end_layout

\begin_layout LyX-Code
ArgList myArgs(myString);
\end_layout

\begin_layout Standard
The resulting ArgList internally looks something like:
\end_layout

\begin_layout LyX-Code
0: trajin
\end_layout

\begin_layout LyX-Code
1: mytraj.nc
\end_layout

\begin_layout LyX-Code
2: 1
\end_layout

\begin_layout LyX-Code
3: 100
\end_layout

\begin_layout LyX-Code
4: 10
\end_layout

\begin_layout Standard
A custom delimeter string containing 1 or more characters can also be used.
 For example, the following string:
\end_layout

\begin_layout LyX-Code
myString = 
\begin_inset Quotes eld
\end_inset

d01,d02,d03,d04
\begin_inset Quotes erd
\end_inset

;
\end_layout

\begin_layout Standard
can be separated via a comma (',') delimiter into 4 tokens like so:
\end_layout

\begin_layout LyX-Code
ArgList myArgs(myString, 
\begin_inset Quotes eld
\end_inset

,
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout Standard
The resulting ArgList internall lookws something like:
\end_layout

\begin_layout LyX-Code
0: d01
\end_layout

\begin_layout LyX-Code
1: d02
\end_layout

\begin_layout LyX-Code
2: d03
\end_layout

\begin_layout LyX-Code
3: d04
\end_layout

\begin_layout Standard
These tokens (or arguments) are stored internally as an STL vector of strings.
 ArgList provides many functions to access user arguments.
 A second array of boolean values records whether an argument has been accessed.
 This concept is functionally similar to the argumentStack in Ptraj; however,
 it avoids the constant memory allocation/deallocation when arguments are
 added/accessed, and allows an argument list to be re-used if desired.
 The two main ways arguments are usually accessed are through 
\begin_inset Quotes eld
\end_inset

GetNextX
\begin_inset Quotes erd
\end_inset

 and 
\begin_inset Quotes eld
\end_inset

GetKeyX
\begin_inset Quotes erd
\end_inset

 functions.
 
\end_layout

\begin_layout Standard
\begin_inset Quotes eld
\end_inset

GetNext
\begin_inset Quotes erd
\end_inset

 functions return the next argument of the desired type.
 For example, using the ArgList created in the first example from 
\begin_inset Quotes eld
\end_inset

trajin mytraj.nc 1 100 10
\begin_inset Quotes erd
\end_inset

 and assuming all arguments are unmarked, GetStringNext() would return 
\begin_inset Quotes eld
\end_inset

trajin
\begin_inset Quotes erd
\end_inset

, while getNextInteger() would return 
\begin_inset Quotes eld
\end_inset

1
\begin_inset Quotes erd
\end_inset

; in both cases the argument returned would be marked, so that a subsequent
 call to GetStringNext() would return 
\begin_inset Quotes eld
\end_inset

mytraj.nc
\begin_inset Quotes erd
\end_inset

 and so on.
 Another very commonly used 
\begin_inset Quotes eld
\end_inset

GetNext
\begin_inset Quotes erd
\end_inset

 function is the 
\begin_inset Quotes eld
\end_inset

GetMaskNext()
\begin_inset Quotes erd
\end_inset

 function, which returns the next atom mask expression (so noted because
 it will begin with ':', '@', '*'); an example of this will be shown below.
\end_layout

\begin_layout Standard
\begin_inset Quotes eld
\end_inset

GetKey
\begin_inset Quotes erd
\end_inset

 functions return an argument next to a specified 
\begin_inset Quotes eld
\end_inset

key
\begin_inset Quotes erd
\end_inset

 string.
 Take for example the argument list created from 
\begin_inset Quotes eld
\end_inset

rmsd R1 @CA ref [myref] out rmsd.dat
\begin_inset Quotes erd
\end_inset

:
\end_layout

\begin_layout LyX-Code
0: rmsd
\end_layout

\begin_layout LyX-Code
1: R1
\end_layout

\begin_layout LyX-Code
2: @CA
\end_layout

\begin_layout LyX-Code
2: ref
\end_layout

\begin_layout LyX-Code
3: [myref]
\end_layout

\begin_layout LyX-Code
4: out
\end_layout

\begin_layout LyX-Code
5: rmsd.dat
\end_layout

\begin_layout Standard
If we want to access a specific argument, we use a 
\begin_inset Quotes eld
\end_inset

GetKey
\begin_inset Quotes erd
\end_inset

 function.
 For example, if we want to know the filename specified by 'out', we would
 use GetStringKey(
\begin_inset Quotes eld
\end_inset

out
\begin_inset Quotes erd
\end_inset

); this would return 
\begin_inset Quotes eld
\end_inset

rmsd.dat
\begin_inset Quotes erd
\end_inset

, and mark both 
\begin_inset Quotes eld
\end_inset

out
\begin_inset Quotes erd
\end_inset

 and 
\begin_inset Quotes eld
\end_inset

rmsd.dat
\begin_inset Quotes erd
\end_inset

.
 Similarly, GetStringKey(
\begin_inset Quotes eld
\end_inset

ref
\begin_inset Quotes erd
\end_inset

) would return 
\begin_inset Quotes eld
\end_inset

[myref]
\begin_inset Quotes erd
\end_inset

.
 At this point we could also use the GetMaskNext() function to get the atom
 mask expression 
\begin_inset Quotes eld
\end_inset

@CA
\begin_inset Quotes erd
\end_inset

.
\end_layout

\begin_layout Subsubsection
ArgList Example
\end_layout

\begin_layout LyX-Code

\end_layout

\begin_layout LyX-Code
ArgList myArgs(
\begin_inset Quotes eld
\end_inset

test_command cutoff 2.0 nval 3 name MyTest :2-30@CA extra
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout LyX-Code
double Cut = myArgs.getKeyDouble(
\begin_inset Quotes eld
\end_inset

cutoff
\begin_inset Quotes erd
\end_inset

, 0.0); // Value 2.0
\end_layout

\begin_layout LyX-Code
int Nval = myArgs.getKeyInteger(
\begin_inset Quotes eld
\end_inset

nval
\begin_inset Quotes erd
\end_inset

, 0);      // Value 3
\end_layout

\begin_layout LyX-Code
int Ntypes = myArgs.getKeyInteger(
\begin_inset Quotes eld
\end_inset

ntypes
\begin_inset Quotes erd
\end_inset

, 0);  // Value 0
\end_layout

\begin_layout LyX-Code
std::string Name = myArgs.GetStringKey(
\begin_inset Quotes eld
\end_inset

name
\begin_inset Quotes erd
\end_inset

);  // Value 
\begin_inset Quotes eld
\end_inset

MyTest
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout LyX-Code
std::string Out = myArgs.GetStringKey(
\begin_inset Quotes eld
\end_inset

out
\begin_inset Quotes erd
\end_inset

);    // Empty
\end_layout

\begin_layout LyX-Code
std::string maskExp = myArgs.GetMaskNext();      // Value 
\begin_inset Quotes eld
\end_inset

:2-30@CA
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout LyX-Code
std::string mask2Exp = myArgs.GetMaskNext();     // Empty
\end_layout

\begin_layout LyX-Code
// At this point only 
\begin_inset Quotes eld
\end_inset

extra
\begin_inset Quotes erd
\end_inset

 will be unmarked.
\end_layout

\begin_layout Subsection
Topology
\end_layout

\begin_layout Standard
The Topology class describes how a system is laid out in terms of Atoms,
 Residues, and Molecules (all of which are classes themselves).
 It may also hold parameters which describe interactions between Atoms (e.g.
 bonds, angles, dihedrals, etc).
 The Topology class is chiefly used in Trajectory input/output and setting
 up atom masks (see below).
\end_layout

\begin_layout Standard
The Topology class has several routines that return strings of atom and
 residue names:
\end_layout

\begin_layout Description
std::string
\begin_inset space ~
\end_inset

TruncResAtomName(int
\begin_inset space ~
\end_inset

atom) Format: 
\begin_inset Quotes eld
\end_inset

<res name><res num>@<atom name>
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout Description
std::string
\begin_inset space ~
\end_inset

AtomMaskName(int
\begin_inset space ~
\end_inset

atom) Format: 
\begin_inset Quotes eld
\end_inset

:<res num>@<atom name>
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout Description
std::string
\begin_inset space ~
\end_inset

TruncAtomNameNum(int
\begin_inset space ~
\end_inset

atom) Format: 
\begin_inset Quotes eld
\end_inset

<atom name>_<atom num>
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout Description
std::string
\begin_inset space ~
\end_inset

TruncResNameNum(int
\begin_inset space ~
\end_inset

residue) Format: 
\begin_inset Quotes eld
\end_inset

<res name>:<res num>
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout Subsubsection
Examples
\end_layout

\begin_layout Standard
Iterate over all atoms in a certain residue.
 This can be accomplished like so:
\end_layout

\begin_layout LyX-Code
// Iterate over all atoms in residue 4, print charge.
\end_layout

\begin_layout LyX-Code
for (int atom_index = Top.Res(4).FirstAtom;
\end_layout

\begin_layout LyX-Code
         atom_index != Top.Res(4).LastAtom;
\end_layout

\begin_layout LyX-Code
       ++atom_index)
\end_layout

\begin_layout LyX-Code
  mprintf(
\begin_inset Quotes eld
\end_inset

Atom %i charge= %g
\backslash
n
\begin_inset Quotes erd
\end_inset

, atom_index+1, Top[atom_index].Charge());
\end_layout

\begin_layout Standard
Iterate over all atoms bonded to a certain atom and pick out the hydrogens:
\end_layout

\begin_layout LyX-Code
// Iterate over all atoms bonded to atom 66
\end_layout

\begin_layout LyX-Code
for (Atom::bond_iterator bond_atom = Top[66].bondbegin();
\end_layout

\begin_layout LyX-Code
                         bond_atom != Top[66].bondend();
\end_layout

\begin_layout LyX-Code
                       ++bond_atom)
\end_layout

\begin_layout LyX-Code
  if (Top[*bond_atom].Element() == Atom::HYDROGEN)
\end_layout

\begin_layout LyX-Code
    mprintf(
\begin_inset Quotes eld
\end_inset

Hydrogen %s bonded to atom %s
\backslash
n
\begin_inset Quotes erd
\end_inset

,
\end_layout

\begin_layout LyX-Code
            Top.AtomMaskName(*bond_atom).c_str(),
\end_layout

\begin_layout LyX-Code
            Top.AtomMaskName(66).c_str());
\end_layout

\begin_layout Subsection
AtomMask/CharMask
\end_layout

\begin_layout Standard
The AtomMask and CharMask classes keep track of what atoms for a given Topology
 are selected based on a given mask expression.
 The AtomMask class holds information on selected atoms only, while CharMask
 has the state of all atoms (selected/not selected).
 AtomMask is as an integer mask, where the atom numbers currently selected
 are stored as an array of integers.
 Since one is usually only interested in selected atoms, most times AtomMask
 is all that is needed and so is the most used mask class in Cpptraj; for
 example, all routines that take masks in the Frame class use the AtomMask
 class.
 
\end_layout

\begin_layout Standard
The CharMask class has an internal character array where the state of each
 atom is stored.
 It has functions that can then be used to interrogate if a certain atom
 or atoms are selected or not.
\end_layout

\begin_layout Standard
In typical use, there are 3 phases to using AtomMask or CharMask: 1) initializat
ion with a mask expression, 2) setup via a Topology class, and 3) iteration
 over the mask/interrogation of the mask.
 Initialization with a mask expression performs all necessary tokenization
 of the mask expression string and prepares the mask to be set up, but does
 not actually select atoms.
 The mask expression can be used during AtomMask construction or passed
 in via SetMaskString():
\end_layout

\begin_layout LyX-Code
AtomMask* Mask = new AtomMask(
\begin_inset Quotes eld
\end_inset

@CA
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout LyX-Code
AtomMask Mask(
\begin_inset Quotes eld
\end_inset

@CA
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout LyX-Code
AtomMask Mask; Mask.SetMaskString(
\begin_inset Quotes eld
\end_inset

@CA
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout Standard
Setup occurs via a Topology class (since in order to set up a mask you need
 to know atom names/numbers, residue name/number/types etc).
 This can be done using SetupIntegerMask() or SetupCharMask() to set up
 an integer mask (more common) or a char mask:
\end_layout

\begin_layout LyX-Code
AtomMask iMask(
\begin_inset Quotes eld
\end_inset

@CA
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout LyX-Code
Top.SetupIntegerMask( iMask );
\end_layout

\begin_layout LyX-Code
CharMask cMask(
\begin_inset Quotes eld
\end_inset

:1-20
\begin_inset Quotes erd
\end_inset

);
\end_layout

\begin_layout LyX-Code
Top.SetupCharMask( Mask );
\end_layout

\begin_layout Standard
Once a mask has been setup the Nselected() function returns the number of
 selected atoms, while the None() function returns true if no atoms were
 selected.
 If necessary, one can convert between the mask types post-setup by using
 AtomMask::ConvertToCharMask() / CharMask::ConvertToIntMask() routines:
\end_layout

\begin_layout LyX-Code
AtomMask mask( CharMask.ConvertToIntMask(), CharMask.Natom() )
\end_layout

\begin_layout LyX-Code
CharMask mask( AtomMask.ConvertToCharMask(), AtomMask.Nselected() )
\end_layout

\begin_layout Standard
The final stage is to make use of the atom mask.
 One can iterate over selected atoms in an integer atom mask using the STL-like
 const_iterator variable and begin() and end() functions - this is the recommend
ed way to use atom masks:
\end_layout

\begin_layout LyX-Code
for (AtomMask::const_iterator atomnum = Mask.begin();
\end_layout

\begin_layout LyX-Code
                              atomnum != Mask.end(); 
\end_layout

\begin_layout LyX-Code
                            ++atomnum)
\end_layout

\begin_layout LyX-Code
  mprintf(
\begin_inset Quotes eld
\end_inset

Selected atom %i
\backslash
n
\begin_inset Quotes erd
\end_inset

, *atomnum);
\end_layout

\begin_layout Standard
One can also access members of the integer array directly via the bracket
 ('[]') operator:
\end_layout

\begin_layout LyX-Code
for (int maskidx = 0; maskidx < Mask.Nselected(); ++maskidx)
\end_layout

\begin_layout LyX-Code
  mprintf(
\begin_inset Quotes eld
\end_inset

Selected atom %i
\backslash
n
\begin_inset Quotes erd
\end_inset

, Mask[atomidx]);
\end_layout

\begin_layout Standard
To see if atom(s) are selected in a CharMask, use the AtomInCharMask() and
 AtomsInCharMask() functions.
 The former returns true if a specified atom is selected, the latter returns
 true if any atoms within a given range are selected.
 For example:
\end_layout

\begin_layout LyX-Code
for (int atom = 0; atom < Top.Natom(); ++atom)
\end_layout

\begin_layout LyX-Code
  if (Mask.AtomInCharMask(atom)) 
\end_layout

\begin_layout LyX-Code
    mprintf(
\begin_inset Quotes eld
\end_inset

Selected Atom %i
\backslash
n
\begin_inset Quotes erd
\end_inset

, atom);
\end_layout

\begin_layout LyX-Code
for (int rnum = 0; rnum < Top.Nres(); ++res)
\end_layout

\begin_layout LyX-Code
  if (Mask.AtomsInCharMask( Top.Res(rnum).FirstAtom(),
\end_layout

\begin_layout LyX-Code
                            Top.Res(rnum).LastAtom() ))
\end_layout

\begin_layout LyX-Code
    mprintf(
\begin_inset Quotes eld
\end_inset

Selected Residue %i
\backslash
n
\begin_inset Quotes erd
\end_inset

, rnum);
\end_layout

\begin_layout Subsection
Frame
\end_layout

\begin_layout Standard
The Frame class is in many ways the workhorse of Cpptraj, as it holds all
 XYZ coordinates for a given input frame, and optionally box coordinates,
 masses, replica indices, temperature, time, and/or velocities.
 Note that although mass is stored in Topology, it is also stored in Frame
 since many calculations require it (center of mass, mass-weighted RMSD,
 etc).
 Coordinates and velocities are stored with double precision.
 Many routines are available to do things like calculate the center of mass
 of atoms, rotate, translate, scale, and so on.
 A major use of the Frame class is to perform RMSD calculations.
\end_layout

\begin_layout Standard
Frames are typically set up in two phases.
 The first phase is memory allocation, which occurs via constructors or
 the SetupFrameX routines.
 This should be done as little as possible since memory allocation is relatively
 expensive.
 The second phase is actually setting the coordinates, which occurs via
 the SetX routines.
 For example, the following code will set up a Frame (newFrame) with the
 coordinates from another Frame (oldFrame) based on a previously set up
 AtomMask (mask) and Topology (top) corresponding to oldFrame:
\end_layout

\begin_layout LyX-Code
Frame newFrame;
\end_layout

\begin_layout LyX-Code
// Allocate memory, copy in masses based on mask.
\end_layout

\begin_layout LyX-Code
newFrame.SetupFrameFromMask( mask, top.Atoms() );
\end_layout

\begin_layout LyX-Code
// Set only coordinates from oldFrame based on mask.
\end_layout

\begin_layout LyX-Code
newFrame.SetCoordinates( oldFrame, mask );
\end_layout

\begin_layout Standard
Alternatively, this can be done in one step with a constructor:
\end_layout

\begin_layout LyX-Code
Frame newFrame( oldFrame, mask );
\end_layout

\begin_layout Standard
The advantage of separating out Setup and Set is that memory reallocation
 is kept to a minimum.
 For example, if we wanted to use newFrame to hold a different set of coordinate
s (of the same size as newFrame or smaller) we might do something like:
\end_layout

\begin_layout LyX-Code
newFrame.SetCoordinates( differentFrame );
\end_layout

\begin_layout Standard
If the new coordinates might be bigger than the current size of newFrame,
 we could explicitly call a SetupFrameX routine; this will only reallocate
 if the new size is greater than the current maximum size.
 A Frame remembers the largest size it was ever allocated for, so reallocation
 is kept to a minimum.
\end_layout

\begin_layout Standard
There are two basic ways to access coordinates within Frame:
\end_layout

\begin_layout Description
const
\begin_inset space ~
\end_inset

double*
\begin_inset space ~
\end_inset

XYZ(atom) return pointer to beginning of XYZ coordinates for given atom
 (max Natom()).
\end_layout

\begin_layout Description
const
\begin_inset space ~
\end_inset

double*
\begin_inset space ~
\end_inset

CRD(coord) return pointer to given coordinate (max size()).
\end_layout

\begin_layout Standard
Note that you can get pointers to the raw coordinates, but it is not recommened
 to use these in general.
\end_layout

\begin_layout Subsubsection
Using Frame for RMSD calculations
\end_layout

\begin_layout Standard
Unlike some of the other functions of Frame, the RMSD functions do not take
 a mask - it is assumed all atoms in the Frame are involved in the RMSD
 calculation.
 This is done for performance reasons.
 If a subset of atoms is desired for an RMSD calculation the reference and
 target Frames should be modified beforehand.
 Since the reference structure usually does not change it is often beneficial
 to pre-center the reference at the origin.
 For example, given a reference Frame (Ref), a Farget frame (Tgt), and an
 AtomMask (mask):
\end_layout

\begin_layout LyX-Code
bool useMass = false;
\end_layout

\begin_layout LyX-Code
top.SetupIntegerMask( mask );
\end_layout

\begin_layout LyX-Code
Frame selectedRef, selectedTgt;
\end_layout

\begin_layout LyX-Code
// Set up and pre-center reference.
\end_layout

\begin_layout LyX-Code
selectedRef.SetupFrameFromMask( mask, top.Atoms() );
\end_layout

\begin_layout LyX-Code
selectedRef.SetCoordinates( Ref, mask );
\end_layout

\begin_layout LyX-Code
// refTrans will contain translation from origin to reference.
\end_layout

\begin_layout LyX-Code
Vec3 refTrans = selectedRef.CenterOnOrigin( useMass );
\end_layout

\begin_layout LyX-Code
// Set up target.
\end_layout

\begin_layout LyX-Code
selectedTgt.SetupFrameFromMask( mask, top.Atoms() );
\end_layout

\begin_layout LyX-Code
selectedTgt.SetCoordinates( Tgt, mask );
\end_layout

\begin_layout LyX-Code
// Calculate RMSD.
 tgtTrans is translation from target to origin.
\end_layout

\begin_layout LyX-Code
Vec3 tgtTrans;
\end_layout

\begin_layout LyX-Code
Matrix_3x3 rot_matrix;
\end_layout

\begin_layout LyX-Code
double rmsd = selectedTgt.RMSD_CenteredRef( selectedRef,
\end_layout

\begin_layout LyX-Code
                                            rot_matrix,
\end_layout

\begin_layout LyX-Code
                                            tgtTrans,
\end_layout

\begin_layout LyX-Code
                                            useMass );
\end_layout

\begin_layout LyX-Code
// Best-fit rotate/translate current Target to Reference.
\end_layout

\begin_layout LyX-Code
Tgt.Trans_Rot_Trans( tgtTrans, rot_matrix, refTrans );
\end_layout

\begin_layout Standard
Now for subsequent RMS calculations to the same Reference, only the selected
 Target frame needs to have its coordinates set:
\end_layout

\begin_layout LyX-Code
selectedTgt.SetCoordinates( Tgt2, mask );
\end_layout

\begin_layout LyX-Code
rmsd = selectedTgt.RMSD_CenteredRef( selectedRef, ...
\end_layout

\begin_layout Subsection
Box
\end_layout

\begin_layout Standard
The Box class holds all unit cell information.
 It has the unit cell vectors (stored in a Matrix_3x3 class in row-major
 order), as well as the fractional cell matrix (for converting from Cartesian
 to fractional coordinates), the box lengths and angles, and the unit cell
 volume.
 There are in general two ways to set up Box with unit cell information:
 the SetupX routines (e.g.
 SetupFromXyzAbg()) and AssignX routines (e.g.
 AssignFromXyzAbg()).
 The SetupX routines perform some extra checks and will return 1 if something
 is wrong with the box.
 The AssignX routines do not perform these checks and so should be used
 where performance is needed (or the state of the box doesn't matter).
\end_layout

\begin_layout Standard
Both the SetupX and AssignX will set the unit and fractional cell matrices,
 the box lengths and angles, and the cell volume each time they are called.
 The ImageOption class can then be used to determine what kind of imaging
 can be performed on the unit cell using the distance calculation routines
 (DIST2(), DIST()) found in DistRoutines.h.
 For example, this is how distances are calculated using ImageOption imageOpt_
 in the distance action.
\end_layout

\begin_layout Standard
1.
 In Init(), determine if imaging should be used if possible:
\end_layout

\begin_layout LyX-Code
imageOpt_.InitImaging( !(actionArgs.hasKey("noimage")) );
\end_layout

\begin_layout Standard
2.
 In Setup(), determine if imaging will be possible based on if box info
 is present:
\end_layout

\begin_layout LyX-Code
imageOpt_.SetupImaging( setup.CoordInfo().TrajBox().HasBox() );
\end_layout

\begin_layout Standard
3.
 In DoAction(), determine what imaging to use and then calculate distance
 using DIST() from DistRoutines.h:
\end_layout

\begin_layout LyX-Code
if (imageOpt_.ImagingEnabled())
\end_layout

\begin_layout LyX-Code
  imageOpt_.SetImageType( frm.Frm().BoxCrd().Is_X_Aligned_Ortho() );
\end_layout

\begin_layout LyX-Code
double Dist = DIST(imageOpt_.ImagingType(), a1, a2_, frm.Frm().BoxCrd());
\end_layout

\begin_layout Section
Console and File Input/Output
\end_layout

\begin_layout Subsection
Output to STDOUT/STDERR: CpptrajStdio.h
\end_layout

\begin_layout Standard
The file CpptrajStdio.cpp contains functions used to write output to standard
 output (STDOUT) and standard error (STDERR).
 This is accomplished with the C printf-like functions mprintf() and mprinterr()
 respectively:
\end_layout

\begin_layout LyX-Code
mprintf(const char* format, ...); // STDOUT
\end_layout

\begin_layout LyX-Code
mprinterr(const char* format, ...); // STDERR
\end_layout

\begin_layout Standard
The 'm' prefix stands for 
\begin_inset Quotes eld
\end_inset

master
\begin_inset Quotes erd
\end_inset

, and ensures that when CPPTRAJ is running via MPI that only the master
 thread is able to write with these functions (note the same does NOT apply
 for OpenMP).
 The syntax is the same as basic printf - a format string followed by any
 variables.
 There are numerous resources that describe printf syntax in detail.
 Some useful syntax is listed here:
\end_layout

\begin_layout Description
%i Integer
\end_layout

\begin_layout Description
%f Floating point number
\end_layout

\begin_layout Description
%g Use scientific or floating point representation, whichever is shorter.
\end_layout

\begin_layout Description
%s Character string
\end_layout

\begin_layout Description
%c Single character
\end_layout

\begin_layout Description

\backslash
n Newline
\end_layout

\begin_layout Description

\backslash
t Tab
\end_layout

\begin_layout Standard
For example:
\end_layout

\begin_layout LyX-Code
#include 
\begin_inset Quotes eld
\end_inset

CpptrajStdio.h
\begin_inset Quotes erd
\end_inset


\end_layout

\begin_layout LyX-Code
int myInteger = 3;
\end_layout

\begin_layout LyX-Code
double myDouble = 5.43;
\end_layout

\begin_layout LyX-Code
string myString = 
\begin_inset Quotes eld
\end_inset

easy
\begin_inset Quotes erd
\end_inset

;
\end_layout

\begin_layout LyX-Code
mprintf(
\begin_inset Quotes eld
\end_inset

Using printf is %s; %i is an integer and %f is a double.
\backslash
n
\begin_inset Quotes erd
\end_inset

,
\end_layout

\begin_layout LyX-Code
        myString.c_str(), myInteger, myDouble);
\end_layout

\begin_layout Standard
Note that the string function c_str() must be used to print C++ strings.
 Output:
\end_layout

\begin_layout LyX-Code
Using printf is easy; 3 is an integer and 5.430000 is a double.
\end_layout

\begin_layout Subsection
CpptrajFile
\end_layout

\begin_layout Standard
The CpptrajFile class provides basic file input and output operations.
 It can handle reading and writing both Gzip and Bzip2 compressed files,
 and through the FileName class performs tilde-expansion on file names (via
 globbing) as well as separates the file name into its base name, extension,
 and compressed extension.
 The file can be opened immediately, or set up first and then opened later.
 Once it has been set up it can be opened or closed multiple times.
 The CpptrajFile class destructor will automatically close the file if it
 is open at time of destruction.
 
\end_layout

\begin_layout Subsection
BufferedLine
\end_layout

\begin_layout Standard
The BufferedLine class is a child of CpptrajFile used for text files that
 will be read in line by line (note that writing is not possible with this
 class).
 The class has an internal buffer, which chunks of the input file are read
 into.
 The Line() routine can be used to read that chunk line by line; this avoids
 potentially expensive file IO.
 When the chunk is empty a new chunk is read in.
 The line can be further split into Tokens (similar to ArgList) and read
 one token at a time; this can be useful for e.g.
 determining the number of columns in a file.
\end_layout

\begin_layout Subsection
BufferedFrame
\end_layout

\begin_layout Standard
The BufferedFrame class is a child of CpptrajFile used for highly-formatted
 text files that will be read/written multiple lines at a time (such as
 the Amber ASCII trajectory format).
 This class is set up for a certain total number of elements of a certain
 character width with a certain number of elements per line, which can then
 be read to or written from a character buffer in one entire chunk.
 
\end_layout

\begin_layout Subsection
FileName, FileName.h
\end_layout

\begin_layout Standard
The FileName class is used by all file-related classes.
 It is more powerful than a string and will automatically do tilde expansion
 and split the name into path, base name, extension, etc.
 FileName.h also contains the File namespace which includes File::Exists()
 for testing whether a file can be opened and File::NameArray and File::ExpandTo
Filenames() for getting an array of files using wildcard matching.
\end_layout

\begin_layout Section
Trajectory Input/Output
\end_layout

\begin_layout Standard
Trajectory input and output (IO) is handled via high-level and low-level
 classes.
 At the highest level trajectory input is provided by Trajin-derived classes
 (Trajin_Single and Trajin_Multi) for reading one frame at a time (i.e.
 during 'trajin' runs) and EnsembleIn-derived classes (EnsembleIn_Single,
 which is experimental currently, and EnsembleIn_Multi) for reading multiple
 frames at a time (i.e.
 during 'ensemble' runs).
 Trajectory output is currently handled by Trajout_Single for writing one
 frame at a time to a single file, and EnsembleOut-dervived classes for
 writing multiple frames out at a time.
 At the lower level IO is handled by format-specific classes which inherit
 from TrajectoryIO (TrajectoryIO.h), which are called Traj_X by convention
 (e.g.
 Traj_AmberNetcdf for Amber NetCDF trajectories).
 TrajectoryIO classes are contained and set up within the higher level classes.
 Note that there is currently no Trajout_Multi (write a single frame to
 multiple files) since this functionality is already handled by TrajoutList.
\end_layout

\begin_layout Subsection
Trajin_Single
\end_layout

\begin_layout Standard
This is for reading in a single frame at a time from a single file.
\end_layout

\begin_layout Subsection
Trajin_Multi
\end_layout

\begin_layout Standard
This is for reading in a single frame at a time from multiple files (e.g.
 getting a frame at a specifed temperature from a T-REMD ensemble).
\end_layout

\begin_layout Subsection
EnsembleIn_Single
\end_layout

\begin_layout Standard
For reading in multiple frames at a time from a single file.
 Currently experimental.
\end_layout

\begin_layout Subsection
EnsembleIn_Multi
\end_layout

\begin_layout Standard
For reading in multiple frames at a time from multiple files, optionally
 sort the frames.
\end_layout

\begin_layout Subsection
Trajout_Single
\end_layout

\begin_layout Standard
For writing out a single frame at a time to a single file.
\end_layout

\begin_layout Subsection
EnsembleOut_Single
\end_layout

\begin_layout Standard
For writing out multiple frames at a time to a single file.
 Currently experimental.
\end_layout

\begin_layout Subsection
EnsembleOut_Multi
\end_layout

\begin_layout Standard
For writing out multiple frames at a time to multiple files.
 Used for 'ensemble' run trajectory output and LES trajectory splitting.
\end_layout

\begin_layout Section
Topology Input/Output
\end_layout

\begin_layout Standard
The 
\emph on
ParmIO
\emph default
 class is a base class for all topology file formats.
 This provides an easy mechanism for extracting the system topology from
 any number of file formats.
 The 
\emph on
ParmFile
\emph default
 class is a wrapper around the 
\emph on
ParmIO
\emph default
 classes that hides the implementation details for each data file type from
 you.
 You should interact with 
\emph on
ParmIO
\emph default
 objects through 
\emph on
ParmFile
\emph default
 handlers.
 
\emph on
ParmFile
\emph default
 provides the ability to both read and write topology file objects of any
 class.
\end_layout

\begin_layout Standard
One thing that sets 
\emph on
ParmIO
\emph default
 and 
\emph on
ParmFile
\emph default
 apart from 
\emph on
TrajectoryIO/TrajectoryFile
\emph default
 and 
\emph on
DataIO/DataFile
\emph default
 is its connection with the 
\emph on
Topology
\emph default
 class.
 
\emph on
Topology
\emph default
 objects contain as much of the information in the Amber topology file as
 can be parsed from the information present in the ParmIO object (and figured
 out based on atomic arrangements).
 A 
\emph on
Topology
\emph default
 instance is the first argument passed to the 
\emph on
ParmFile::Read
\emph default
 function, followed by the name of the topology file.
 Unlike the 
\emph on
DataFile
\emph default
 and 
\emph on
TrajectoryFile
\emph default
 classes, 
\emph on
ParmFile
\emph default
 does not have a reference to the 
\emph on
ParmIO
\emph default
 object to forward read/write information to.
 It exists simply to fill the 
\emph on
Topology
\emph default
 class with the relevant data structures and inform it how to do the rest.
 The 
\emph on
Topology
\emph default
 class is format-independent, providing a layer of abstraction to make other
 parts of the code that require topology information less error-prone while
 coding.
\end_layout

\begin_layout Standard
Every 
\emph on
ParmIO
\emph default
 subclass implements a 
\emph on
ReadParm
\emph default
 method that takes a 
\emph on
Topology
\emph default
 instance as the first argument and fills as much of the information there
 as possible.
 Afterwards, the CommonSetup method of the Topology class is called to finish
 setup and determine bond information (from atom distances if not present
 directly in the file format) and molecule information (based on the bonded
 structure).
 The currently available types of topologies are summarized in 
\begin_inset CommandInset ref
LatexCommand formatted
reference "tbl:Cpptraj-Parm-Formats"

\end_inset

.
\end_layout

\begin_layout Standard
\begin_inset Float table
wide false
sideways false
status open

\begin_layout Plain Layout
\begin_inset Caption Standard

\begin_layout Plain Layout
Topology file formats currently implemented in Cpptraj.
 The first column has the 
\emph on
ParmIO
\emph default
 class name as well as the 
\emph on
ParmFormatType
\emph default
 enumeration type that corresponds to that class inside 
\emph on
ParmFile
\emph default
 in parentheses.
\end_layout

\end_inset


\end_layout

\begin_layout Plain Layout
\align center
\begin_inset Tabular
<lyxtabular version="3" rows="5" columns="2">
<features tabularvalignment="middle">
<column alignment="center" valignment="top">
<column alignment="center" valignment="top">
<row>
<cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout

\emph on
ParmIO
\emph default
 Subclass (
\emph on
ParmFormatType
\emph default
)
\end_layout

\end_inset
</cell>
<cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Description
\end_layout

\end_inset
</cell>
</row>
<row>
<cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Parm_Amber (AMBERPARM)
\end_layout

\end_inset
</cell>
<cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Amber style topology file (OLD and NEW styles)
\end_layout

\end_inset
</cell>
</row>
<row>
<cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Parm_CharmmPsf (CHARMMPSF)
\end_layout

\end_inset
</cell>
<cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
CHARMM PSF topology file format (used by NAMD, too)
\end_layout

\end_inset
</cell>
</row>
<row>
<cell alignment="center" valignment="top" topline="true" leftline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Parm_Mol2 (MOL2FILE)
\end_layout

\end_inset
</cell>
<cell alignment="center" valignment="top" topline="true" leftline="true" rightline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
TRIPOS Mol2 file
\end_layout

\end_inset
</cell>
</row>
<row>
<cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
Parm_PDB
\end_layout

\end_inset
</cell>
<cell alignment="center" valignment="top" topline="true" bottomline="true" leftline="true" rightline="true" usebox="none">
\begin_inset Text

\begin_layout Plain Layout
PDB File
\end_layout

\end_inset
</cell>
</row>
</lyxtabular>

\end_inset


\end_layout

\begin_layout Plain Layout
\begin_inset CommandInset label
LatexCommand label
name "tbl:Cpptraj-Parm-Formats"

\end_inset


\end_layout

\end_inset


\end_layout

\begin_layout Section
The DataSet and DataFile Framework
\end_layout

\begin_layout Standard
One of the goals in writing cpptraj was to try and generalize data collection
 and output, so that any action could output any generated data in any format
 known to cpptraj without having to write any extra code.
 For example, data generated by a distance calculation can be output in
 columns, as a Grace file, as a Gnuplot file, or as all three.
 To that end Actions and Analyses have access to the main DataSetList in
 CpptrajState (usually named DSL inside Actions) and the main DataFileList
 (usually named DFL inside actions).
 A DataSet can be generated by an Action, but because it is held outside
 the Action in the master DataSetList it can persist after the Action that
 generated it has been destroyed and be used in subsequent Analyses etc.
 DataSets have a base type which determines what kind of data it can hold,
 and can also belong to a group; DataSets in the same group behave in a
 similar fashion.
 For example, all DataSets in the COORDINATES group hold coordinates, etc.
\end_layout

\begin_layout Subsection
The MetaData class
\end_layout

\begin_layout Standard
DataSets are associated with meta data which is used to both describe and
 categorize the data.
 DataSet selection uses meta data; meta data is also used to sort DataSets.
 This information is contained within the MetaData class.
 A DataSet can have a name, aspect, index, type etc.
 This allows for DataSets to e.g.
 be marked as an alpha torsion (M_TORSION, ALPHA).
 There are currently 9 MetaData variables:
\end_layout

\begin_layout Description
name_ The DataSet name; this is the most general level of classification.
 Used in searches.
\end_layout

\begin_layout Description
fileName_ If the data was read in from a file, the name of that file.
 Used in searches; both the full path and base name can be used to match
 (e.g.
 for reference frame data sets).
\end_layout

\begin_layout Description
aspect_ A 
\begin_inset Quotes eld
\end_inset

sub-name
\begin_inset Quotes erd
\end_inset

, used to differentiate between different aspects of similar data.
 For example, DataSets generated by the 'nastruct' Action have different
 aspects for nucleic acid 
\begin_inset Quotes eld
\end_inset

stretch
\begin_inset Quotes erd
\end_inset

, 
\begin_inset Quotes eld
\end_inset

shear
\begin_inset Quotes erd
\end_inset

, 
\begin_inset Quotes eld
\end_inset

stagger
\begin_inset Quotes erd
\end_inset

, etc.
 Used in searches.
\end_layout

\begin_layout Description
legend_ This is the name that will be used when writing data out to a file.
 Not used in searches.
 A default one is created if one is not provided.
\end_layout

\begin_layout Description
idx_ (Index) A number which can be used to further differentiate data.
 It could correspond to residue, atom, etc.
 Used in searches.
\end_layout

\begin_layout Description
ensembleNum_ (Ensemble number) For use during ensemble processing; this
 is set by the DataSetList to differentiate data from different ensembles.
\end_layout

\begin_layout Description
scalarmode_ Internal categorization of the type of data, e.g.
 a distance, angle, torsion, etc.
 Certain functions will use this data - for example data sets marked as
 angle, torsion, or pucker will take periodicity into account when averaging
 (DataSet_1D::Avg()).
\end_layout

\begin_layout Description
scalartype_ Internal sub-type; so in the case of a torison what type of
 torsion it is (phi, psi etc) or in the case of a matrix what kind of matrix,
 etc.
\end_layout

\begin_layout Description
timeSeries_ Whether the DataSet is a time series; used by DataSetList to
 determine whether to call the Allocate() function.
\end_layout

\begin_layout Standard
MetaData can be set using the DataSet::SetMeta() routine, but it is recommended
 that this be done sparingly once a DataSet is part of the master DataSetList
 since this could create conflicts.
\end_layout

\begin_layout Subsection
The TextFormat class
\end_layout

\begin_layout Standard
This class is used when writing the data set out to text files.
 It creates a printf-like format string of a given type with specified width
 and precision.
\end_layout

\begin_layout Subsection
Brief DataSet/DataFile Example
\end_layout

\begin_layout Standard
A simple usage example is given here.
 Say for example we want to track a distance in an Action.
 The first step is to create the DataSet in the Init() routine.
\end_layout

\begin_layout LyX-Code
dist_ = DSL->AddSet(DataSet::DOUBLE, MetaData(actionArgs.GetStringNext(),
\end_layout

\begin_layout LyX-Code
                                              MetaData::M_DISTANCE), "Dis");
\end_layout

\begin_layout LyX-Code
if (dist_==0) return 1;
\end_layout

\begin_layout Standard
In the first line a DataSet class of type DOUBLE is added to the master
 DataSetList.
 The various types are enumerated in DataSet::DataType (DataSet.h).
 The DataSet will be named whatever the next string is in the actionArgs
 ArgList.
 If there is no name, a default one will be created based on the given default
 "Dis" and the DataSet's overall position in the DataSetList (so in this
 case the default could be something like Dis_00000).
 The DataSet is also given the scalarMode M_DISTANCE, which is information
 that other Actions/Analyses can use (like Analysis_Statistics).
 What is returned is a pointer to the DataSet; DataSet is actually a base
 class that specific DataSet types inherit (in this case DataSet_double).
 In this way the interface is generalized.
\end_layout

\begin_layout Standard
The next step is to add the DataSet to the DataFileList.
\end_layout

\begin_layout LyX-Code
DataFile* outfile = DFL->AddFile(distanceFile, actionArgs);
\end_layout

\begin_layout LyX-Code
if (outfile != 0) outfile->AddDataSet( dist_ );
\end_layout

\begin_layout Standard
In the first line a pointer to a new or exsiting (if already created somewhere
 else) DataFile is returned only if the string distanceFile is not empty;
 this allows specification of output files to be optional.
 The actionsArgs ArgList is passed in so that the DataFile outfile can process
 any DataFile-related arguments.
 In the second line the DataSet is added to the DataFile.
 In this way output from multiple actions can be combined rather than overwritte
n.
 The machinery of the DataFileList takes care of output (formatting etc)
 from there.
\end_layout

\begin_layout Standard
The final phase is actually adding data to the DataSet.
 So for example in the action() routine you could have:
\end_layout

\begin_layout LyX-Code
double distance = sqrt( DIST2_NoImage( V1, V2 ) );
\end_layout

\begin_layout LyX-Code
dist_->Add(frameNum, &distance);
\end_layout

\begin_layout Standard
In the first line the value 'distance' is being calculated.
 In the next line the value from 'distance' is being added to DataSet dist_
 with frame number 'frameNum' (automatically set within Action).
 Notice that the address of 'distance' is passed rather than the value;
 this is a necessity from the generalization of the DataSet interface.
 DataSet has no idea a prior what the data type might be, so in the Add
 routine the value is cast to what the underlying DataSet implementation
 expects.
 This allows the Add routine to be used for double, float, int, string,
 etc.
 This DataSet could also be cast back to it's actual type to access other
 routines, e.g.:
\end_layout

\begin_layout LyX-Code
DataSet_double& ds_dist = static_cast<DataSet_double&>( *dist_ );
\end_layout

\begin_layout Subsection
DataSet_1D / SCALAR_1D
\end_layout

\begin_layout Standard
Base class for 1D scalar data sets (like DataSet_double, DataSet_Mesh, etc).
 Basically hold a series of numbers.
\end_layout

\begin_layout Subsection
DataSet_2D / MATRIX_2D
\end_layout

\begin_layout Standard
Base class for 2D matrices.
\end_layout

\begin_layout Subsection
DataSet_3D / GRID_3D
\end_layout

\begin_layout Standard
Base class for 3D grids.
\end_layout

\begin_layout Subsection
DataSet_Coords / COORDINATES
\end_layout

\begin_layout Standard
Base class for DataSets which hold coordinates.
\end_layout

\begin_layout Part
Adding New Functionality
\end_layout

\begin_layout Standard
Most development for Cpptraj will likely be in adding new functionality;
 actions, analyses, and trajectory/topology/data file formats.
 This part of the manual will provide guidance and some helpful hints to
 this end.
 In general, adding new functionality is done by writing an implementation
 of the desired class type (e.g.
 for actions, inherit from the Action class) and then adding that class
 to the container for that specific functionality (e,g, in the case of actions,
 ActionList).
\end_layout

\begin_layout Standard
Note that there is a script provided that can generate default templates
 for various class types, located at $CPPTRAJHOME/devtools/Template.sh:
\end_layout

\begin_layout LyX-Code
Usage: ./devtools/Template.sh <name> [<type>] 
\end_layout

\begin_layout LyX-Code
  <type>: Action Analysis Exec Traj DataIO DataSet
\end_layout

\begin_layout Section
Adding Actions - Example
\end_layout

\begin_layout Standard
All actions inherit from the Action abstract base class.
 The Action class itself inherits from the DispatchObject class so that
 it can be associated with an allocator (to create the action) and a help
 function.
 There are four functions that every action must implement: Init(), Setup(),
 DoAction(), and Print().
 Init() is called when the action is first created, and processes input
 arguments, sets up DataSets/DataFiles, deals with reference frames, and
 sets the debug level.
 Setup() is called to set the action up for a specific topology, and so
 handles anything Topology-related (such as parsing atom masks).
 The DoAction() function is called to actually perform the action on input
 coordinate frames.
 The Init(), Setup(), and DoAction() functions return a special type of
 integer, Action::RetType, which described the result of the action:
\end_layout

\begin_layout Description
Action::OK Action is successful.
\end_layout

\begin_layout Description
Action::ERR Action is not successful.
\end_layout

\begin_layout Description
Action::USE_ORIGINAL_FRAME Action requests that the original unmodified
 topology/frame be used (see e.g.
 Action_Unstrip in Action_Strip.h).
\end_layout

\begin_layout Description
Action::SUPPRESS_COORD_OUTPUT Action requests that further processing of
 the current coordinate frame be skipped (see e.g.
 Action_RunningAvg).
\end_layout

\begin_layout Description
Action::SKIP Non-fatal problem occurred during setup; skip Action until
 next Setup call.
\end_layout

\begin_layout Description
Action::MODIFY_TOPOLOGY Setup routine has modified the Topology/CoordinateInfo.
\end_layout

\begin_layout Description
Action::MODIFY_COORDS DoAction routine has modified the Frame.
\end_layout

\begin_layout Standard
The final function is Print(), which is called after all trajectory processing
 is complete and performs any additional calculation or output necessary.
 This function can be blank if such functionality is not needed, but it
 still must be implemented.
\end_layout

\begin_layout Standard
In addition to Action, there are currently two additional action-related
 classes that actions may want to inherit from.
 The ImagedAction class is for classes that may need to calculate imaged
 distances, and the ActionFrameCounter class is for actions that may want
 to process subsets of input frames (see e.g.
 the Action_Matrix action).
\end_layout

\begin_layout Standard
As an example, we will go through the creation of a simplified version of
 the Action_Distance class for calculating distances; this will cover using
 the DataSet, DataFile, AtomMask, and ImagedAction classes as well.
\end_layout

\begin_layout Subsection
Create the Class Header
\end_layout

\begin_layout Standard
As mentioned in the style guide, header files should be named after the
 class, so the Action_Distance class will go in a file named 
\begin_inset Quotes eld
\end_inset

Action_Distance.h
\begin_inset Quotes erd
\end_inset

.
 The first thing to do is create a 
\begin_inset Quotes eld
\end_inset

header guard
\begin_inset Quotes erd
\end_inset

 - this will prevent issues with multiple inclusion.
 The header guard should be named after the class and header file, so for
 Action_Distance.h:
\end_layout

\begin_layout LyX-Code
#ifndef INC_ACTION_DISTANCE_H
\end_layout

\begin_layout LyX-Code
#define INC_ACTION_DISTANCE_H
\end_layout

\begin_layout Standard
Next comes the class description.
 Since distance calculations may involve imaging we also include the ImagedActio
n class as a variable to simplify image handling:
\end_layout

\begin_layout LyX-Code
class Action_Distance: public Action {
\end_layout

\begin_layout Standard
Following the style guide, we first implement any public methods.
 For actions this is at least the constructor, the allocator (named Alloc()
 by convention), and the Help() function.
 The allocator and help functions need to be static so that they can be
 called without instantiating the class.
 
\end_layout

\begin_layout LyX-Code
public:
\end_layout

\begin_layout LyX-Code
  Action_Distance();  ///< Constructor
\end_layout

\begin_layout LyX-Code
  /// Allocator
\end_layout

\begin_layout LyX-Code
  static DispatchObject* Alloc() { return (DispatchObject*)new Action_Distance()
; }
\end_layout

\begin_layout LyX-Code
  static void Help(); ///< Help function
\end_layout

\begin_layout Standard
The implemented functions Init(), Setup(), DoAction(), and Print() can be
 either public or private, although the preference is private.
\end_layout

\begin_layout Standard
The private section is where all functions and variables specific to the
 class will go.
 First, we add entries for the functions inherited from the Action base
 class which must be implemented.
 Since we will not need to do any post-processing for this action, the Print()
 function is empty:
\end_layout

\begin_layout LyX-Code
  Action::RetType Init(ArgList&, ActionInit&, int);
\end_layout

\begin_layout LyX-Code
  Action::RetType Setup(ActionSetup&);
\end_layout

\begin_layout LyX-Code
  Action::RetType DoAction(int, ActionFrame&);
\end_layout

\begin_layout LyX-Code
  void Print() {}
\end_layout

\begin_layout Standard
Next we define the class variables.
 For Action_Distance we will want a DataSet to hold the calculated distances,
 two AtomMasks to describe the points between which the distance should
 be calculated, and a variable to indicate whether the distance should be
 mass-weighted.
\end_layout

\begin_layout LyX-Code
private:
\end_layout

\begin_layout LyX-Code
  DataSet* dist_;      ///< Will hold DataSet of calculated distances.
\end_layout

\begin_layout LyX-Code
  bool useMass_;       ///< If true, mass-weight distances.
\end_layout

\begin_layout LyX-Code
  AtomMask Mask1_;     ///< First atom selection.
\end_layout

\begin_layout LyX-Code
  AtomMask Mask2_;     ///< Second atom selection
\end_layout

\begin_layout LyX-Code
  ImagedAction image_; ///< Holds imaging info
\end_layout

\begin_layout Standard
All variables related to imaging are already include via the ImagedAction
 class.
 
\end_layout

\begin_layout Standard
Last, end the class definition and finish the header guard:
\end_layout

\begin_layout LyX-Code
};
\end_layout

\begin_layout LyX-Code
#endif
\end_layout

\begin_layout Subsection
Create the Class Implementation
\end_layout

\begin_layout Standard
Following the naming scheme, the class implementation will go into Action_Distan
ce.cpp.
 The first part of this file will have the necessary #include directives.
 We need <cmath> for the square root function, Action_Distance.h for the
 class definition, and CpptrajStdio.h for wiriting to the console.
 
\end_layout

\begin_layout LyX-Code
#include <cmath>
\end_layout

\begin_layout LyX-Code
#include "Action_Distance.h"
\end_layout

\begin_layout LyX-Code
#include "CpptrajStdio.h"
\end_layout

\begin_layout Standard
First we will need to create the class constructor.
 It is encouraged that users make use of initalizer lists (which tend to
 be more efficient) for this purpose.
 In this case we have two non-class variables: dist_, which is a pointer
 to a DataSet, and useMass_, which is boolean:
\end_layout

\begin_layout LyX-Code
Action_Distance::Action_Distance() : dist_(0), useMass_(true) {}
\end_layout

\begin_layout Standard
Next, ensure that the Help() function has an implementation.
 Note that in cpptraj 
\begin_inset Quotes eld
\end_inset

mprintf
\begin_inset Quotes erd
\end_inset

 is used over 
\begin_inset Quotes eld
\end_inset

printf
\begin_inset Quotes erd
\end_inset

 for making any future IO modifications easier:
\end_layout

\begin_layout LyX-Code
void Action_Distance::Help() {
\end_layout

\begin_layout LyX-Code
  mprintf("distance [<name>] <mask1> <mask2> [out <filename>] [geom] [noimage]
\backslash
n");
\end_layout

\begin_layout LyX-Code
}
\end_layout

\begin_layout Subsubsection
Init() - Parse user arguments, set up DataSets/DataFiles etc
\end_layout

\begin_layout Standard
Init() is called when the action is created and is responsible for parsing
 the Argument list (ArgList) and inital setup.
 Init() has the same input arguments for every action:
\end_layout

\begin_layout LyX-Code
Action::RetType 
\end_layout

\begin_layout LyX-Code
  Action_Distance::Init(ArgList& actionArgs, ActionInit& init, int debugIn)
\end_layout

\begin_layout LyX-Code
{
\end_layout

\begin_layout Standard
The input arguments are as follows: 
\series bold
actionArgs
\series default
 (ArgList class) contains arguments from user input, 
\series bold
init
\series default
 (ActionInit class) contains the master DataSetList and master DataFileList,
 and 
\series bold
debugIn
\series default
 is the current debug level for all actions.
 It is up to the action implementation whether it wants to record the debug
 level or not.
\end_layout

\begin_layout Standard
Typical order of argument processing is keywords, masks, DataSet name.
 First we will process the keywords 'noimage', 'geom', and 'out <filename>'.
\end_layout

\begin_layout Standard
In order to determine whether the action will try to use imaging we call
 the InitImaging() function (inherited from the ImagedAction class).
 If the ArgList actionArgs contains the string 
\begin_inset Quotes eld
\end_inset

noimage
\begin_inset Quotes erd
\end_inset

, false will be sent to InitImaging to disable imaging:
\end_layout

\begin_layout LyX-Code
  image_.InitImaging( !(actionArgs.hasKey("noimage")) );
\end_layout

\begin_layout Standard
Next we set useMass_.
 If actionArgs contains the string 
\begin_inset Quotes eld
\end_inset

geom
\begin_inset Quotes erd
\end_inset

, useMass_ will be set to false:
\end_layout

\begin_layout LyX-Code
  useMass_ = !(actionArgs.hasKey("geom"));
\end_layout

\begin_layout Standard
Next, we will try to create an output DataFile:
\end_layout

\begin_layout LyX-Code
  DataFile* outfile = init.DFL().AddDataFile( actionArgs.GetStringKey("out"),
\end_layout

\begin_layout LyX-Code
                                              actionArgs );
\end_layout

\begin_layout Standard
The behavior of AddDataFile() depends on the result from actionArgs.GetStringKey(
); if 
\begin_inset Quotes eld
\end_inset

out
\begin_inset Quotes erd
\end_inset

 is present in actionArgs, the next string (presumably <filename>) is returned
 and passed to AddDataFile(), and a DataFile will be returned corresponding
 to <filename>.
 If 
\begin_inset Quotes eld
\end_inset

out
\begin_inset Quotes erd
\end_inset

 is not present nothing will be returned, no file will be set up, and outfile
 will be null (0).
 
\end_layout

\begin_layout Standard
Next, we will get two atom mask expressions.
 We will require that the user must specify two masks, so if either of the
 strings is empty return an error:
\end_layout

\begin_layout LyX-Code
  std::string mask1 = actionArgs.GetMaskNext();
\end_layout

\begin_layout LyX-Code
  std::string mask2 = actionArgs.GetMaskNext();
\end_layout

\begin_layout LyX-Code
  if (mask1.empty() || mask2.empty()) {
\end_layout

\begin_layout LyX-Code
    mprinterr("Error: distance: Requires 2 masks
\backslash
n");
\end_layout

\begin_layout LyX-Code
    return Action::ERR;
\end_layout

\begin_layout LyX-Code
  }
\end_layout

\begin_layout Standard
Now we can use the mask expression strings to initialize the two AtomMask
 classes (note that this tokenizes the mask expressions but does not yet
 set them up since we need topology information to do that):
\end_layout

\begin_layout LyX-Code
  Mask1_.SetMaskString(mask1);
\end_layout

\begin_layout LyX-Code
  Mask2_.SetMaskString(mask2);
\end_layout

\begin_layout Standard
Next we will use the master DataSetList (init.DSL()) to create a DataSet
 to store the calculated distances.
 We will use a version of DataSetList::AddSet() that allows us to specify
 the DataSet type, MetaData, and a default name if no name is specified
 (
\begin_inset Quotes eld
\end_inset

Dis
\begin_inset Quotes erd
\end_inset

).
 If any errors occur in creating the DataSet, NULL (0) will be returned.
 Note that the string returned by actionArgs.GetStringNext() is implicitly
 converted to a MetaData class.
\end_layout

\begin_layout LyX-Code
  dist_ = init.DSL().AddSet(DataSet::DOUBLE, actionArgs.GetStringNext(), "Dis");
\end_layout

\begin_layout LyX-Code
  if (dist_==0) return Action::ERR;
\end_layout

\begin_layout Standard
If a DataFile was previously set up, we now add the DataSet to this DataFile:
\end_layout

\begin_layout LyX-Code
  if (outfile != 0) outfile->AddDataSet( dist_ );
\end_layout

\begin_layout Standard
Last, we print out some information regarding how the Action has been initialize
d and return Action::OK to indicate successful intialization:
\end_layout

\begin_layout LyX-Code
  mprintf("    DISTANCE: %s to %s",Mask1_.MaskString(), Mask2_.MaskString());
\end_layout

\begin_layout LyX-Code
  if (!image_.UseImage())
\end_layout

\begin_layout LyX-Code
    mprintf(", non-imaged");
\end_layout

\begin_layout LyX-Code
  if (useMass_)
\end_layout

\begin_layout LyX-Code
    mprintf(", center of mass");
\end_layout

\begin_layout LyX-Code
  else 
\end_layout

\begin_layout LyX-Code
   mprintf(", geometric center");   mprintf(".
\backslash
n");
\end_layout

\begin_layout LyX-Code
  return Action::OK;
\end_layout

\begin_layout LyX-Code
}
\end_layout

\begin_layout Standard

\series bold
IMPORTANT:
\series default
 Note that this is the only time in which the master DataSetList is passed
 to the Action.
 If the Action will need to set up DataSets later (because e.g.
 they may depend on what's in the Topology, like in the case of the 
\series bold
\emph on
multidihedral
\series default
\emph default
 command), it should save a pointer to the master DataSetList using the
 
\series bold
init.DslPtr()
\series default
 function, e.g.
\end_layout

\begin_layout LyX-Code
masterDSL_ = init.DslPtr();
\end_layout

\begin_layout Subsubsection
Setup() - Set up Topology-related parts of the Action
\end_layout

\begin_layout Standard
Setup() is called whenever the action needs to be set up for a given Topology
 file.
 Any component of the action that depends on Topology (in this case the
 AtomMasks and the Imaging) is handled here.
 The arguments to Setup() are:
\end_layout

\begin_layout LyX-Code
Action::RetType Action_Distance::Setup(ActionSetup& setup) {
\end_layout

\begin_layout Standard
Note that the 
\series bold
setup
\series default
 variable (ActionSetup class) contains a pointer to the current Topology
 and current trajectory CoordinateInfo, as well as the number of expected
 frames associated with this Topology during the current run.
 Actions that want to modify the current Topology or CoordinateInfo for
 subsequent Actions can do so using the 
\series bold
setup
\series default
 variable (see e.g.
 Action_Strip).
 
\end_layout

\begin_layout Standard
First, we setup the AtomMasks.
 Each AtomMask is passed to the current topology using the SetupIntegerMask()
 function, which will create an integer array containing only the selected
 atoms based on the mask expression.
 If we needed to know both selected and unselected atoms we could use the
 SetupCharMask() function instead.
\end_layout

\begin_layout LyX-Code
  if (setup.Top().SetupIntegerMask( Mask1_ )) return Action::ERR;
\end_layout

\begin_layout LyX-Code
  if (setup.Top().SetupIntegerMask( Mask2_ )) return Action::ERR;
\end_layout

\begin_layout Standard
After this, we print some information about what atoms are selected (note
 we could also use the MaskInfo() function of AtomMask for this).
 For calculating distance, we need to make sure atoms were actually selected
 (using the None() function of AtomMask).
 If no atoms were selected this may be because the mask is only valid for
 certain Topologies during the run, so in that case make it a non-fatal
 error (i.e.
 a Warning) and return Action::SKIP:
\end_layout

\begin_layout LyX-Code
  mprintf("
\backslash
t%s (%i atoms) to %s (%i atoms)",Mask1_.MaskString(), Mask1_.Nselected(),
\end_layout

\begin_layout LyX-Code
          Mask2_.MaskString(),Mask2_.Nselected());
\end_layout

\begin_layout LyX-Code
  if (Mask1_.None() || Mask2_.None()) {
\end_layout

\begin_layout LyX-Code
    mprintf("
\backslash
nWarning: distance: One or both masks have no atoms.
\backslash
n");
\end_layout

\begin_layout LyX-Code
    return Action::SKIP;
\end_layout

\begin_layout LyX-Code
  }
\end_layout

\begin_layout Standard
Next we determine if imaging can actually be performed based on the box
 information present in the current trajectory's CoordinateInfo; if there
 is no box information imaging cannot be performed.
 We do this with the image_.SetupImaging() function (ImagedAction class).
 The image_.ImagingEnabled() function will let us know if imaging for this
 Topology is possible or not:
\end_layout

\begin_layout LyX-Code
  image_.SetupImaging( setup.CoordInfo().TrajBox().Type() );
\end_layout

\begin_layout LyX-Code
  if (image_.ImagingEnabled())
\end_layout

\begin_layout LyX-Code
    mprintf(", imaged");
\end_layout

\begin_layout LyX-Code
  else
\end_layout

\begin_layout LyX-Code
    mprintf(", imaging off");
\end_layout

\begin_layout LyX-Code
  mprintf(".
\backslash
n");
\end_layout

\begin_layout Standard
Now all Topology-dependent aspects of the action are set up.
 Return Action::OK.
\end_layout

\begin_layout LyX-Code
  return Action::OK;
\end_layout

\begin_layout LyX-Code
}
\end_layout

\begin_layout Standard

\series bold
IMPORTANT:
\series default
 Note that this is the only time in which a Topology is passed to the Action.
 If the Action requires Topology information later (such as in DoAction()
 or Print()) it should save a pointer to the Topology using the 
\series bold
setup.TopAddress()
\series default
 function, e.g.
\end_layout

\begin_layout LyX-Code
currentParm_ = setup.TopAddress();
\end_layout

\begin_layout Paragraph
Modification of Topology Info
\end_layout

\begin_layout Standard
If there will be a modification of Topology/Frame information (e.g.
 removing atoms, adding velocity information to the Frame, etc) the Action
 must return Action::MODIFY_TOPOLOGY.
 If the CoordinateInfo will be modified, the Action should have a copy of
 the CoordinateInfo in the Action class itself; 
\series bold
it must not reside only in Action::Setup(), otherwise it will be lost when
 setup exits
\series default
.
\end_layout

\begin_layout Standard
For example, say we are adding time information to a frame that currently
 does not have it.
 The code might look something like:
\end_layout

\begin_layout LyX-Code
cInfo_ = setup.CoordInfo(); 
\end_layout

\begin_layout LyX-Code
if (!cInfo_.HasTime())
\end_layout

\begin_layout LyX-Code
  cInfo_.SetTime( true );
\end_layout

\begin_layout LyX-Code
setup.SetCoordInfo( &cInfo_ );
\end_layout

\begin_layout Standard
Here, cInfo_ is a CoordinateInfo variable in the Action class.
\end_layout

\begin_layout Subsubsection
DoAction() - Process input Frame
\end_layout

\begin_layout Standard
Coordinates are read in a frame at a time and stored in a Frame class, which
 is then passed to each action in the ActionList.
 The DoAction() function is called to process a coordinate Frame.
 The arguments are:
\end_layout

\begin_layout LyX-Code
Action::RetType Action_Distance::DoAction(int frameNum, ActionFrame& frm)
 {
\end_layout

\begin_layout Standard
The first argument 
\series bold
frameNum
\series default
 is the current frame number (starting at 0).
 Note that the 
\series bold
frm
\series default
 variable (ActionFrame class) contains a pointer to the current Frame.
 Actions that want to alter the current Frame beyond just manipulating coordinat
es for subsequent Actions (e.g.
 changing the Frame size or adding velocity info etc) can do so via the
 
\series bold
frm
\series default
 variable (see e.g.
 Action_Closest).
 
\end_layout

\begin_layout Standard
There are several variables needed for calculating the distance.
 First, we have two Vec3 classes (Vec3.h, which is already included from
 other headers) to store the XYZ coordinates of the points:
\end_layout

\begin_layout LyX-Code
  Vec3 a1, a2;
\end_layout

\begin_layout Standard
If we are performing non-orthorhombic imaging we need to store the matrices
 which perform conversion from Cartesian to fractional coordinates and vice
 versa (using Matrix_3x3 classes, Matrix_3x3.h).
 Note that these are called 'ucell' and 'recip' respectively throughout
 CPPTRAJ, as these were the names used for the analogous structures in PTRAJ.
\end_layout

\begin_layout LyX-Code
  Matrix_3x3 ucell, recip;
\end_layout

\begin_layout Standard
Finally, we need a double to store the actual result of the distance calculation
:
\end_layout

\begin_layout LyX-Code
  double Dist;
\end_layout

\begin_layout Standard
In the first part of the actual calculation, we calculate the centers of
 the coordinates in Mask1_ and Mask2_, either mass-weighted or not depending
 on useMass_, using the appropriate functions from the Frame class (Frame.h):
\end_layout

\begin_layout LyX-Code
  if (useMass_) {
\end_layout

\begin_layout LyX-Code
    a1 = frm.Frm().VCenterOfMass( Mask1_ );
\end_layout

\begin_layout LyX-Code
    a2 = frm.Frm().VCenterOfMass( Mask2_ );
\end_layout

\begin_layout LyX-Code
  } else {
\end_layout

\begin_layout LyX-Code
    a1 = frm.Frm().VGeometricCenter( Mask1_ );
\end_layout

\begin_layout LyX-Code
    a2 = frm.Frm().VGeometricCenter( Mask2_ );
\end_layout

\begin_layout LyX-Code
  }
\end_layout

\begin_layout Standard
Note that here we are using the Frm() function, which returns a constant
 (i.e.
 non-modifiable) reference to the current Frame; if we wanted to actually
 manipulate the coordinates we would have to call ModifyFrm().
 
\end_layout

\begin_layout Standard
Next, we get the distance between the coordinates stored in a1 and a2.
 For non-orthorhombic imaging we first need to convert the current box coordinat
es (stored in the Frame class in double precision as 3 lengths and 3 angles)
 into the coordinate conversion matrices using the ToRecip() function of
 the Box class (Box.h).
 Then, depending on the type of imaging that needs to be performed we call
 the appropriate distance calculation routine (DIST2_XXX, found in DistRoutines.h
):
\end_layout

\begin_layout LyX-Code
  switch ( image_.ImageType() ) {
\end_layout

\begin_layout LyX-Code
    case NONORTHO:
\end_layout

\begin_layout LyX-Code
      frm.Frm().BoxCrd().ToRecip(ucell, recip);
\end_layout

\begin_layout LyX-Code
      Dist = DIST2_ImageNonOrtho(a1, a2, ucell, recip);
\end_layout

\begin_layout LyX-Code
      break;
\end_layout

\begin_layout LyX-Code
    case ORTHO:
\end_layout

\begin_layout LyX-Code
      Dist = DIST2_ImageOrtho(a1, a2, frm.Frm().BoxCrd());
\end_layout

\begin_layout LyX-Code
      break;
\end_layout

\begin_layout LyX-Code
    case NOIMAGE:
\end_layout

\begin_layout LyX-Code
      Dist = DIST2_NoImage(a1, a2);       break;
\end_layout

\begin_layout LyX-Code
  }
\end_layout

\begin_layout LyX-Code
  Dist = sqrt(Dist);
\end_layout

\begin_layout Standard
Last, we add the result to the DataSet and return Action::OK.
 Since DataSet is just an interface we pass in the address of Dist (&Dist)
 to let the underlying DataSet framework take care of the fact that it is
 a double.
\end_layout

\begin_layout LyX-Code
  dist_->Add(frameNum, &Dist);
\end_layout

\begin_layout LyX-Code
  return Action::OK;
\end_layout

\begin_layout LyX-Code
}
\end_layout

\begin_layout Subsubsection
Print() - Any post-processing
\end_layout

\begin_layout Standard
The Print() function is called once all input frames have been read in,
 and is used if there is anything that should be printed outside the normal
 DataFile/DataSet framework (e.g.
 hydrogen bond averages in the hbond action) or if there are any additional
 calculations that need to be performed (e.g.
 finishing up matrix calculations in the matrix action).
 In this example we're only calculating a simple distance; the output is
 handled by the DataFile/DataSet framework, so we implement a blank Print()
 function in the header:
\end_layout

\begin_layout LyX-Code
void Print() {}
\end_layout

\begin_layout Subsection
Add the Action to the Command class
\end_layout

\begin_layout Standard
Now that the class implementation is complete, we need to let cpptraj know
 how to call it.
 This is currently done using a 
\begin_inset Quotes eld
\end_inset

static
\begin_inset Quotes erd
\end_inset

 Class, Command (Command.cpp), which is initialized by the Cpptraj class
 via Command::Init() when the program starts.
 Command::Init() makes use of the Command::AddCmd() function to add the
 Command, set its destination, and any associated keywords.
 The Command::AddCmd() function looks like:
\end_layout

\begin_layout LyX-Code
void Command::AddCmd(DispatchObject* oIn, Cmd::DestType dIn, int nKeys,
 ...)
\end_layout

\begin_layout Standard
where 
\series bold
oIn
\series default
 is a pointer to the DispatchObject (Exec-, Action-, Analysis-, or Deprecated-de
rived class), 
\series bold
dIn
\series default
 determines how the Command will be processed, 
\series bold
nKeys
\series default
 is the number of keywords associated with the command, and the remaining
 arguments are the command keywords.
 For example:
\end_layout

\begin_layout LyX-Code
Command::AddCmd( new Action_Rmsd(), Cmd::ACT, 2, "rms", "rmsd" );
\end_layout

\begin_layout Standard
Adds a new instance of the Action-derived class Action_Rmsd, sets its destinatio
n as Action (Cmd::ACT), and sets 2 associated command keys, 
\begin_inset Quotes eld
\end_inset

rms
\begin_inset Quotes erd
\end_inset

 and 
\begin_inset Quotes eld
\end_inset

rmsd
\begin_inset Quotes erd
\end_inset

.
\end_layout

\begin_layout Standard
To make navigation of Commands.cpp easier, you can search for ACTION (or
 ANALYSIS if adding an Analysis) to go where things need to be added.
 First add the class to Commands.cpp with the appropriate '#include'.
 Includes should be in alphabetical order within their given section.
\end_layout

\begin_layout LyX-Code
#include "Action_Dihedral.h"
\end_layout

\begin_layout LyX-Code

\series bold
#include "Action_Distance.h"
\end_layout

\begin_layout LyX-Code
#include "Action_Hbond.h"
\end_layout

\begin_layout Standard
Then add the command to Command::Init() using Command::AddCmd(), e.g.:
\end_layout

\begin_layout LyX-Code
Command::AddCmd( new Action_Dipole(), Cmd::ACT, 1, "dipole" );
\end_layout

\begin_layout LyX-Code

\series bold
Command::AddCmd( new Action_Distance(), Cmd::ACT, 1, "distance" );
\end_layout

\begin_layout LyX-Code
Command::AddCmd( new Action_DistRmsd(), Cmd::ACT, 2, "drms", "drmsd" );
 
\end_layout

\begin_layout Standard
\begin_inset CommandInset index_print
LatexCommand printindex
type "idx"
name "Index"
literal "true"

\end_inset


\end_layout

\end_body
\end_document