KVDataSetManager Class Reference

Detailed Description

Manage all datasets contained in a given data repository.

KVDataSetManager handles a collection of datasets contained in a data repository, which may be analysed with KaliVeda. The list of all known datasets is defined in $KVROOT/KVFiles/.kvrootrc. Some of these datasets may not be physically present in the repository. This is checked by method CheckAvailability(). This will search in the data repository for the subdirectories corresponding to each dataset, and afterwards KVDataSet::IsAvailable() will return kTRUE or kFALSE for each dataset depending on the result of the search.

ACTIVE DATA REPOSITORY/DATASET MANAGER

Each data repository has a data set manager. At any time, one repository is designated as being 'active' or 'current', and it's address is held in the global pointer gDataRepository. One can then access the corresponding dataset manager through gDataSetManager.

Definition at line 39 of file KVDataSetManager.h.

#include <KVDataSetManager.h>

Inheritance diagram for KVDataSetManager:

Public Member Functions
	KVDataSetManager ()
virtual	~KVDataSetManager ()
virtual void	CheckAvailability ()
virtual Bool_t	CheckUser (const Char_t *groupname, const Char_t *username="")
KVDataAnalysisTask *	GetAnalysisTaskAny (const Char_t *keywords) const
virtual KVDataSet *	GetAvailableDataSet (Int_t) const
KVDataSet *	GetDataSet (const Char_t *)
	Return pointer to DataSet using name.
KVDataSet *	GetDataSet (Int_t) const
	Return pointer to DataSet using index in list of all datasets, index>=0.
virtual Int_t	GetNavailable () const
virtual Int_t	GetNtotal () const
virtual KVDataAnalysisTask *	GetTask (const Char_t *name)
	Return pointer to named data analysis task.
const KVNameValueList &	GetUserGroups ()
virtual Bool_t	Init (KVDataRepository *=0)
virtual void	Print (Option_t *opt="") const
virtual void	Update ()

Private Member Functions
virtual Bool_t	CheckCacheStatus ()
const KVSeqCollection *	GetAnalysisTaskList () const
virtual KVDataSet *	NewDataSet ()
	Creates and returns pointer to new data set object.
virtual Bool_t	OpenAvailableDatasetsFile ()
virtual Bool_t	ReadAvailableDatasetsFile ()
virtual Bool_t	ReadDataSetList ()
virtual Bool_t	ReadTaskList ()
virtual void	ReadUserGroups ()

Private Attributes
Bool_t	fCacheAvailable
	kTRUE if caching is activated for parent repository
TString	fCacheFileName
	name of cache file ( = [repository name].available.datasets)
std::ifstream	fDatasets
	for reading cached repository available datasets file
KVUniqueNameList	fDataSets
	list of datasets handled by manager
std::vector< Int_t >	fIndex
	array of indices of available datasets
UInt_t	fMaxCacheTime
	maximum allowed age of cache file in seconds
Int_t	fNavailable
	number of available datasets
KVDataRepository *	fRepository
	the repository for which data sets are handled
KVUniqueNameList	fTasks
	list of all known analysis tasks
KVNameValueList	fUserGroups
	list of user groups

Friends
class	KVDataSet

Constructor & Destructor Documentation

KVDataSetManager()

KVDataSetManager::KVDataSetManager

(

)

Definition at line 33 of file KVDataSetManager.cpp.

~KVDataSetManager()

KVDataSetManager::~KVDataSetManager ( )

virtual

Definition at line 48 of file KVDataSetManager.cpp.

Member Function Documentation

CheckAvailability()

void KVDataSetManager::CheckAvailability ( )

virtual

Check availability of datasets in repository associated to this data set manager

If caching is activated for the parent repository, i.e. if

[repository name].DataRepository.CacheAvailable: yes

then instead of directly checking the existence of the directories for each dataset, we use the cached information written in the file KVBase::WorkingRepository()/[repository name].available.datasets unless (1) it doesn't exist, or (2) the file is older than the maximum cache time (in seconds) defined by

[repository name].DataRepository.MaxCacheSeconds:

In either of these 2 cases, we check the existence of the directories and update/ create the cache file.

If the repository appears to be empty (perhaps because we are using a remote access protocol to check it, and the protocol has some problems...), then as a last resort we we will use the cache if it exists, whatever its age.

Reimplemented in KVRemoteDataSetManager.

Definition at line 279 of file KVDataSetManager.cpp.

CheckCacheStatus()

Bool_t KVDataSetManager::CheckCacheStatus ( )

privatevirtual

We check the status of the available datasets cache file. We return kTRUE if the file exists & was last modified less than fMaxCacheTime seconds ago.

Definition at line 585 of file KVDataSetManager.cpp.

CheckUser()

Bool_t KVDataSetManager::CheckUser ( const Char_t *  groupname, const Char_t *  username = ""  )

virtual

Check in list of groups fUserGroups if the user name 'username' is part of the group 'groupname'. If 'username' is not given (default) we use current user info (gSystem->GetUserInof()->fUser).

Definition at line 431 of file KVDataSetManager.cpp.

GetAnalysisTaskAny()

KVDataAnalysisTask * KVDataSetManager::GetAnalysisTaskAny

(

const Char_t *

keywords

)

const

This method returns a pointer to the analysis task whose description (title) contains all of the whitespace-separated keywords (which may be regular expressions) given in the string "keywords". The comparison is case-insensitive. case-insensitive search for matches in list of all analysis tasks, based on 'title' attribute

Definition at line 622 of file KVDataSetManager.cpp.

GetAnalysisTaskList()

const KVSeqCollection * KVDataSetManager::GetAnalysisTaskList ( ) const

inlineprivate

Definition at line 58 of file KVDataSetManager.h.

GetAvailableDataSet()

KVDataSet * KVDataSetManager::GetAvailableDataSet ( Int_t  index ) const

virtual

Return pointer to available DataSet using index of available datasets Note this index begins at 1, and corresponds to the number printed next to the dataset when Print("available") is called

Definition at line 404 of file KVDataSetManager.cpp.

GetDataSet() [1/2]

KVDataSet * KVDataSetManager::GetDataSet

(

const Char_t *

name

)

Return pointer to DataSet using name.

Definition at line 391 of file KVDataSetManager.cpp.

GetDataSet() [2/2]

KVDataSet * KVDataSetManager::GetDataSet

(

Int_t

index

)

const

Return pointer to DataSet using index in list of all datasets, index>=0.

Definition at line 378 of file KVDataSetManager.cpp.

GetNavailable()

virtual Int_t KVDataSetManager::GetNavailable ( ) const

inlinevirtual

Definition at line 79 of file KVDataSetManager.h.

GetNtotal()

virtual Int_t KVDataSetManager::GetNtotal ( ) const

inlinevirtual

Definition at line 83 of file KVDataSetManager.h.

GetTask()

KVDataAnalysisTask * KVDataSetManager::GetTask ( const Char_t *  name )

virtual

Return pointer to named data analysis task.

Definition at line 419 of file KVDataSetManager.cpp.

GetUserGroups()

const KVNameValueList & KVDataSetManager::GetUserGroups ( )

inline

Definition at line 87 of file KVDataSetManager.h.

Init()

Bool_t KVDataSetManager::Init ( KVDataRepository *  dr = 0 )

virtual

Initialisation of dataset manager for the repository 'dr'. If dr=0x0 (default) then all known datasets are 'available', otherwise we check availability of datasets based on data present in repository. Initialise all possible data analysis tasks, then set list of possible tasks for each available dataset.

Returns kTRUE if all goes well. For remote data repositories we return kFALSE if no datasets are available (as the access to these is read-only)

Definition at line 65 of file KVDataSetManager.cpp.

NewDataSet()

KVDataSet * KVDataSetManager::NewDataSet ( )

privatevirtual

Creates and returns pointer to new data set object.

Definition at line 453 of file KVDataSetManager.cpp.

OpenAvailableDatasetsFile()

Bool_t KVDataSetManager::OpenAvailableDatasetsFile ( )

privatevirtual

Opens file KVBase::WorkingDirectory()/[repository name].available.datasets containing cached info on available datasets and associated subdirectories in data repository. Opens file for reading, & if all goes well returns kTRUE. Returns kFALSE in case of problems.

Reimplemented in KVRemoteDataSetManager.

Definition at line 492 of file KVDataSetManager.cpp.

Print()

void KVDataSetManager::Print ( Option_t *  opt = "" ) const

virtual

Print list of datasets If opt="" (default) all datasets are shown with full information if opt="available" only available datasets are shown, each with a number which can be used with GetAvailableDataSet(Int_t) in order to retrieve the corresponding dataset.

Definition at line 224 of file KVDataSetManager.cpp.

ReadAvailableDatasetsFile()

Bool_t KVDataSetManager::ReadAvailableDatasetsFile ( )

privatevirtual

Opens and reads file containing cached info on available datasets, and sets the availability of the concerned datasets. Returns kTRUE if all goes well. Returns kFALSE if no cache exists or if file cannot be opened.

Definition at line 511 of file KVDataSetManager.cpp.

ReadDataSetList()

Bool_t KVDataSetManager::ReadDataSetList ( )

privatevirtual

Initialise list of all known datasets from informations in $KVROOT/KVFIles/.kvrootrc (and user's .kvrootrc)

Definition at line 153 of file KVDataSetManager.cpp.

ReadTaskList()

Bool_t KVDataSetManager::ReadTaskList ( )

privatevirtual

Initialise list of all known analysis tasks from informations in $KVROOT/KVFIles/.kvrootrc (and user's .kvrootrc)

Definition at line 188 of file KVDataSetManager.cpp.

ReadUserGroups()

void KVDataSetManager::ReadUserGroups ( )

privatevirtual

Sets up list of user groups defining restricted access to certain datasets. Definition of different user groups is given in $KVROOT/KVFiles/.kvrootrc UserGroups env var contains whitespace-separated list of group names

Definition at line 120 of file KVDataSetManager.cpp.

Update()

void KVDataSetManager::Update ( )

virtual

Called when the physical state of the repository has changed i.e. a subdirectory for a new dataset or datatype has been added or removed. We update the available datatsets, datatypes and analysis tasks. check which datasets are available

Definition at line 467 of file KVDataSetManager.cpp.

Friends And Related Symbol Documentation

KVDataSet

friend class KVDataSet

friend

Definition at line 41 of file KVDataSetManager.h.

Member Data Documentation

fCacheAvailable

Bool_t KVDataSetManager::fCacheAvailable

private

kTRUE if caching is activated for parent repository

Definition at line 64 of file KVDataSetManager.h.

fCacheFileName

TString KVDataSetManager::fCacheFileName

private

name of cache file ( = [repository name].available.datasets)

Definition at line 66 of file KVDataSetManager.h.

fDatasets

std::ifstream KVDataSetManager::fDatasets

private

for reading cached repository available datasets file

Definition at line 44 of file KVDataSetManager.h.

fDataSets

KVUniqueNameList KVDataSetManager::fDataSets

private

list of datasets handled by manager

Definition at line 45 of file KVDataSetManager.h.

fIndex

std::vector<Int_t> KVDataSetManager::fIndex

private

array of indices of available datasets

Definition at line 48 of file KVDataSetManager.h.

fMaxCacheTime

UInt_t KVDataSetManager::fMaxCacheTime

private

maximum allowed age of cache file in seconds

Definition at line 65 of file KVDataSetManager.h.

fNavailable

Int_t KVDataSetManager::fNavailable

private

number of available datasets

Definition at line 47 of file KVDataSetManager.h.

fRepository

KVDataRepository* KVDataSetManager::fRepository

private

the repository for which data sets are handled

Definition at line 56 of file KVDataSetManager.h.

fTasks

KVUniqueNameList KVDataSetManager::fTasks

private

list of all known analysis tasks

Definition at line 46 of file KVDataSetManager.h.

fUserGroups

KVNameValueList KVDataSetManager::fUserGroups

private

list of user groups

Definition at line 49 of file KVDataSetManager.h.