factor out index table helper and define its contract

This commit is contained in:
Fischlurch 2015-01-04 13:23:57 +01:00
parent d0dcccbd1b
commit 80eec4132b
6 changed files with 211 additions and 260 deletions

View file

@ -39,39 +39,18 @@
#define LIB_DIFF_INDEX_TABLE_H
#include "lib/diff/list-diff.hpp"
#include "lib/format-string.hpp"
#include "lib/error.hpp"
#include <algorithm>
#include <vector>
#include <tuple>
#include <map>
namespace lib {
namespace diff{
#include "lib/test/run.hpp"
#include "lib/diff/list-diff.hpp"
#include "lib/iter-adapter.hpp"
#include "lib/itertools.hpp"
#include "lib/util.hpp"
#include <boost/noncopyable.hpp>
#include <string>
#include <vector>
using lib::append_all;
using util::unConst;
using util::isnil;
using std::string;
using std::vector;
using std::move;
using std::swap;
namespace lib {
namespace diff{
//#########################
using std::vector;
using std::map;
template<typename VAL>
class IndexTable
@ -111,172 +90,5 @@ namespace diff{
template<class SEQ>
class DiffDetector
: boost::noncopyable
{
using Val = typename SEQ::value_type;
using Idx = IndexTable<Val>;
Idx refIdx_;
SEQ const& currentData_;
using DiffStep = typename ListDiffLanguage<Val>::DiffStep;
/** @internal state frame for diff detection and generation. */
class DiffFrame;
public:
explicit
DiffDetector(SEQ const& refSeq)
: refIdx_(refSeq)
, currentData_(refSeq)
{ }
/** does the current state of the underlying sequence differ
* from the state embodied into the last reference snapshot taken?
* @remarks will possibly evaluate and iterate the whole sequence
*/
bool
isChanged() const
{
UNIMPLEMENTED("change detection");
}
/** Diff is a iterator to yield a sequence of DiffStep elements */
using Diff = lib::IterStateWrapper<DiffStep, DiffFrame>;
/** Diff generation core operation.
* Take a snapshot of the \em current state of the underlying sequence
* and establish a frame to find the differences to the previously captured
* \em old state. This possible difference evaluation is embodied into a #Diff
* iterator and handed over to the client, while the snapshot of the current state
* becomes the new reference point from now on.
* @return iterator to yield a sequence of DiffStep tokens, which describe the changes
* between the previous reference state and the current state of the sequence.
* @note takes a new snapshot to supersede the old one, i.e. updates the DiffDetector.
* @warning the returned iterator retains a reference to the current (new) snapshot.
* Any concurrent modification leads to undefined behaviour. You must not
* invoke #pullUpdate while another client still explores the result
* of an old evaluation.
*/
Diff
pullUpdate()
{
Idx mark (currentData_);
swap (mark, refIdx_); // mark now refers to old reference point
return Diff(DiffFrame(refIdx_, move(mark)));
}
};
/**
* A diff generation process is built on top of an "old" reference point
* and a "new" state of the underlying sequence. Within this reference frame,
* an demand-driven evaluation of the differences is handed out to the client
* as an iterator. While consuming this evaluation process, both the old and
* the new version of the sequence will be traversed once. In case of re-orderings,
* a nested forward lookup similar to insertion sort will look for matches in the
* old sequence, rendering the whole evaluation quadratic in worst-case.
*/
template<class SEQ>
class DiffDetector<SEQ>::DiffFrame
{
Idx old_;
Idx* new_;
size_t oldHead_=0,
newHead_=0;
static ListDiffLanguage<Val> token;
DiffStep currentStep_;
public:
DiffFrame(Idx& current, Idx&& refPoint)
: old_(refPoint)
, new_(&current)
, currentStep_(establishNextState())
{ }
/* === Iteration control API for IterStateWrapper== */
friend bool
checkPoint (DiffFrame const& frame)
{
return token.NIL != frame.currentStep_;
}
friend DiffStep&
yield (DiffFrame const& frame)
{
REQUIRE (checkPoint (frame));
return unConst(frame).currentStep_;
}
friend void
iterNext (DiffFrame & frame)
{
frame.establishNextState();
}
private:
DiffStep
establishNextState()
{
if (canPick())
{
consumeOld();
return token.pick (consumeNew());
}
if (canDelete())
return token.del (consumeOld());
if (canInsert())
return token.ins (consumeNew());
if (needFetch())
return token.find (consumeNew());
if (obsoleted())
return token.skip (consumeOld());
return token.NIL;
}
bool hasOld() const { return oldHead_ < old_.size(); }
bool hasNew() const { return newHead_ < new_->size(); }
bool canPick() const { return hasOld() && hasNew() && oldElm()==newElm(); }
bool canDelete() const { return hasOld() && !new_->contains(oldElm()); }
bool canInsert() const { return hasNew() && !old_.contains(newElm()); }
bool needFetch() const { return hasNew() && oldHead_ < old_.pos(newElm()); }
bool obsoleted() const { return hasOld() && newHead_ > new_->pos(oldElm()); }
Val const& oldElm() const { return old_.getElement (oldHead_); }
Val const& newElm() const { return new_->getElement (newHead_); }
Val const& consumeOld() { return old_.getElement (oldHead_++); }
Val const& consumeNew() { return new_->getElement (newHead_++); }
};
/** allocate static storage for the diff language token builder functions */
template<class SEQ>
ListDiffLanguage<typename DiffDetector<SEQ>::Val> DiffDetector<SEQ>::DiffFrame::token;
//#########################
}} // namespace lib::diff
#endif /*LIB_DIFF_INDEX_TABLE_H*/

View file

@ -59,76 +59,19 @@
#include "lib/diff/list-diff.hpp"
#include "lib/format-string.hpp"
#include <algorithm>
#include <vector>
#include <tuple>
namespace lib {
namespace diff{
#include "lib/test/run.hpp"
#include "lib/diff/list-diff.hpp"
#include "lib/diff/index-table.hpp"
#include "lib/iter-adapter.hpp"
#include "lib/itertools.hpp"
#include "lib/util.hpp"
#include <boost/noncopyable.hpp>
#include <string>
#include <vector>
using lib::append_all;
using util::unConst;
using util::isnil;
using std::string;
using std::vector;
using std::move;
using std::swap;
#include <utility>
namespace lib {
namespace diff{
//#########################
template<typename VAL>
class IndexTable
{
public:
template<class SEQ>
IndexTable(SEQ const& seq)
{
UNIMPLEMENTED("build index");
}
size_t
size() const
{
UNIMPLEMENTED("sequence size");
}
VAL const&
getElement (size_t i) const
{
UNIMPLEMENTED("indexed value access");
}
bool
contains (VAL const& elm) const
{
return size() == pos(elm);
}
size_t
pos (VAL const& elm) const
{
UNIMPLEMENTED("index lookup");
}
};
using util::unConst;
using std::move;
using std::swap;
@ -291,11 +234,6 @@ namespace diff{
ListDiffLanguage<typename DiffDetector<SEQ>::Val> DiffDetector<SEQ>::DiffFrame::token;
//#########################
}} // namespace lib::diff

View file

@ -72,6 +72,11 @@ return: 0
END
TEST "Diff: lookup table helper" DiffIndexTable_test <<END
return: 0
END
TEST "A Digxel (numeric component)" Digxel_test <<END
out-lit: empty____## +0.0 ##
out-lit: value____##-88.8 ##

View file

@ -0,0 +1,196 @@
/*
DiffIndexTable(Test) - simple sequence lookup table
Copyright (C) Lumiera.org
2015, Hermann Vosseler <Ichthyostega@web.de>
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License as
published by the Free Software Foundation; either version 2 of
the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
* *****************************************************/
#include "lib/test/run.hpp"
#include "lib/test/test-helper.hpp"
#include "lib/diff/index-table.hpp"
#include <utility>
#include <string>
#include <vector>
using std::string;
using std::vector;
using std::swap;
namespace lib {
namespace diff{
namespace test{
using lumiera::error::LUMIERA_ERROR_LOGIC;
namespace {//Test fixture....
using DataSeq = vector<string>;
using Index = IndexTable<string>;
#define TOK(id) id(STRINGIFY(id))
string TOK(a1), TOK(a2), TOK(a3), TOK(a4), TOK(a5);
string TOK(b1), TOK(b2), TOK(b3), TOK(b4);
}//(End)Test fixture
/*****************************************************************************//**
* @test Cover a simple lookup table helper, used to support diff generation.
* The IndexTable takes a snapshot of generic sequence data on construction.
* After that, it is immutable and supports lookup by element as key and
* membership check.
*
* @see IndexTable
* @see DiffListApplication_test
*/
class DiffIndexTable_test : public Test
{
virtual void
run (Arg)
{
simpleUsage();
verifySnapshot();
duplicateDetection();
copy_and_move();
}
void
simpleUsage()
{
DataSeq data({a5,a2,a1,a4,a3});
Index idx(data);
CHECK (5 == idx.size());
CHECK (idx.contains(a1));
CHECK (idx.contains(a2));
CHECK (idx.contains(a3));
CHECK (idx.contains(a4));
CHECK (idx.contains(a5));
CHECK (!idx.contains(b1));
CHECK (!idx.contains(b2));
CHECK (a5 == idx.getElement(0));
CHECK (a2 == idx.getElement(1));
CHECK (a1 == idx.getElement(2));
CHECK (a4 == idx.getElement(3));
CHECK (a3 == idx.getElement(4));
CHECK (0 == idx.pos(a5));
CHECK (1 == idx.pos(a2));
CHECK (2 == idx.pos(a1));
CHECK (3 == idx.pos(a4));
CHECK (4 == idx.pos(a3));
}
void
verifySnapshot()
{
DataSeq data({a5,a2,a1,a4,a3});
Index idx(data);
data.clear();
data.push_back(b1);
CHECK (5 == idx.size());
CHECK (idx.contains(a1));
CHECK (idx.contains(a2));
CHECK (idx.contains(a3));
CHECK (idx.contains(a4));
CHECK (idx.contains(a5));
CHECK (!idx.contains(b1));
CHECK (!idx.contains(b2));
CHECK (0 == idx.pos(a5));
CHECK (1 == idx.pos(a2));
CHECK (2 == idx.pos(a1));
CHECK (3 == idx.pos(a4));
CHECK (4 == idx.pos(a3));
}
void
duplicateDetection()
{
DataSeq data({a5,a2,a1,a4,a2,a3});
VERIFY_ERROR(LOGIC, Index idx(data));
}
void
copy_and_move()
{
DataSeq seqA({a5,a4,a1,a2,a3});
DataSeq seqB({b4,b3,b2,b1});
Index idxA(seqA);
Index idxB(seqB);
CHECK (5 == idxA.size());
CHECK (4 == idxB.size());
CHECK ( idxA.contains(a1));
CHECK (!idxA.contains(b1));
CHECK (!idxB.contains(a1));
CHECK ( idxB.contains(b1));
swap (idxA, idxB);
CHECK (!idxA.contains(a1));
CHECK ( idxA.contains(b1));
CHECK ( idxB.contains(a1));
CHECK (!idxB.contains(b1));
idxA = idxB;
CHECK (4 == idxA.size());
CHECK (4 == idxB.size());
CHECK (!idxA.contains(a1));
CHECK ( idxA.contains(b1));
CHECK (!idxB.contains(a1));
CHECK ( idxB.contains(b1));
}
};
/** Register this test class... */
LAUNCHER (DiffIndexTable_test, "unit common");
}}} // namespace lib::diff::test