gbloader.cpp

来自「ncbi源码」· C++ 代码 · 共 1,566 行 · 第 1/4 页

* Added PCH ncbi_pch.hpp
*
* Revision 1.105  2004/03/16 15:47:29  vasilche
* Added CBioseq_set_Handle and set of EditHandles
*
* Revision 1.104  2004/02/17 21:17:53  vasilche
* Fixed 'non-const reference to temporary' warning.
*
* Revision 1.103  2004/02/04 20:59:46  ucko
* Fix the PubSeq entry point name to something that actually exists....
*
* Revision 1.102  2004/02/04 17:47:40  kuznets
* Fixed naming of entry points
*
* Revision 1.101  2004/01/22 20:10:35  vasilche
* 1. Splitted ID2 specs to two parts.
* ID2 now specifies only protocol.
* Specification of ID2 split data is moved to seqsplit ASN module.
* For now they are still reside in one resulting library as before - libid2.
* As the result split specific headers are now in objects/seqsplit.
* 2. Moved ID2 and ID1 specific code out of object manager.
* Protocol is processed by corresponding readers.
* ID2 split parsing is processed by ncbi_xreader library - used by all readers.
* 3. Updated OBJMGR_LIBS correspondingly.
*
* Revision 1.100  2004/01/13 21:52:06  vasilche
* Resurrected new version.
*
* Revision 1.4  2004/01/13 16:55:55  vasilche
* CReader, CSeqref and some more classes moved from xobjmgr to separate lib.
* Headers moved from include/objmgr to include/objtools/data_loaders/genbank.
*
* Revision 1.3  2003/12/30 22:14:41  vasilche
* Updated genbank loader and readers plugins.
*
* Revision 1.98  2003/12/30 19:51:24  vasilche
* Implemented CGBDataLoader::GetSatSatkey() method.
*
* Revision 1.97  2003/12/30 17:43:17  vasilche
* Fixed warning about unused variable.
*
* Revision 1.96  2003/12/30 16:41:19  vasilche
* Removed warning about unused variable.
*
* Revision 1.95  2003/12/19 19:49:21  vasilche
* Use direct Seq-id -> sat/satkey resolution without intermediate gi.
*
* Revision 1.94  2003/12/03 15:14:05  kuznets
* CReader management re-written to use plugin manager
*
* Revision 1.93  2003/12/02 23:17:34  vasilche
* Fixed exception in ID1 reader when invalid Seq-id is supplied.
*
* Revision 1.92  2003/12/01 23:42:29  vasilche
* Temporary fix for segfault in genbank data loader in multithreaded applications.
*
* Revision 1.91  2003/11/26 17:55:58  vasilche
* Implemented ID2 split in ID1 cache.
* Fixed loading of splitted annotations.
*
* Revision 1.90  2003/10/27 19:28:02  vasilche
* Removed debug message.
*
* Revision 1.89  2003/10/27 18:50:49  vasilche
* Detect 'private' blobs in ID1 reader.
* Avoid reconnecting after ID1 server replied with error packet.
*
* Revision 1.88  2003/10/27 15:05:41  vasilche
* Added correct recovery of cached ID1 loader if gi->sat/satkey cache is invalid.
* Added recognition of ID1 error codes: private, etc.
* Some formatting of old code.
*
* Revision 1.87  2003/10/22 16:12:37  vasilche
* Added CLoaderException::eNoConnection.
* Added check for 'fail' state of ID1 connection stream.
* CLoaderException::eNoConnection will be rethrown from CGBLoader.
*
* Revision 1.86  2003/10/07 13:43:23  vasilche
* Added proper handling of named Seq-annots.
* Added feature search from named Seq-annots.
* Added configurable adaptive annotation search (default: gene, cds, mrna).
* Fixed selection of blobs for loading from GenBank.
* Added debug checks to CSeq_id_Mapper for easier finding lost CSeq_id_Handles.
* Fixed leaked split chunks annotation stubs.
* Moved some classes definitions in separate *.cpp files.
*
* Revision 1.85  2003/09/30 16:22:02  vasilche
* Updated internal object manager classes to be able to load ID2 data.
* SNP blobs are loaded as ID2 split blobs - readers convert them automatically.
* Scope caches results of requests for data to data loaders.
* Optimized CSeq_id_Handle for gis.
* Optimized bioseq lookup in scope.
* Reduced object allocations in annotation iterators.
* CScope is allowed to be destroyed before other objects using this scope are
* deleted (feature iterators, bioseq handles etc).
* Optimized lookup for matching Seq-ids in CSeq_id_Mapper.
* Added 'adaptive' option to objmgr_demo application.
*
* Revision 1.84  2003/09/05 17:29:40  grichenk
* Structurized Object Manager exceptions
*
* Revision 1.83  2003/08/27 14:25:22  vasilche
* Simplified CCmpTSE class.
*
* Revision 1.82  2003/08/14 20:05:19  vasilche
* Simple SNP features are stored as table internally.
* They are recreated when needed using CFeat_CI.
*
* Revision 1.81  2003/07/24 19:28:09  vasilche
* Implemented SNP split for ID1 loader.
*
* Revision 1.80  2003/07/22 22:01:43  vasilche
* Removed use of HAVE_LIBDL.
*
* Revision 1.79  2003/07/17 22:51:31  vasilche
* Fixed unused variables warnings.
*
* Revision 1.78  2003/07/17 20:07:56  vasilche
* Reduced memory usage by feature indexes.
* SNP data is loaded separately through PUBSEQ_OS.
* String compression for SNP data.
*
* Revision 1.77  2003/06/24 14:25:18  vasilche
* Removed obsolete CTSE_Guard class.
* Used separate mutexes for bioseq and annot maps.
*
* Revision 1.76  2003/06/11 14:54:06  vasilche
* Fixed wrong error message in CGBDataLoader destructor when
* some data requests were failed.
*
* Revision 1.75  2003/06/10 19:01:07  vasilche
* Fixed loader methods string.
*
* Revision 1.74  2003/06/10 15:25:33  vasilche
* Changed wrong _ASSERT to _VERIFY
*
* Revision 1.73  2003/06/02 16:06:37  dicuccio
* Rearranged src/objects/ subtree.  This includes the following shifts:
*     - src/objects/asn2asn --> arc/app/asn2asn
*     - src/objects/testmedline --> src/objects/ncbimime/test
*     - src/objects/objmgr --> src/objmgr
*     - src/objects/util --> src/objmgr/util
*     - src/objects/alnmgr --> src/objtools/alnmgr
*     - src/objects/flat --> src/objtools/flat
*     - src/objects/validator --> src/objtools/validator
*     - src/objects/cddalignview --> src/objtools/cddalignview
* In addition, libseq now includes six of the objects/seq... libs, and libmmdb
* replaces the three libmmdb? libs.
*
* Revision 1.72  2003/05/20 21:13:02  vasilche
* Fixed ambiguity on MSVC.
*
* Revision 1.71  2003/05/20 18:27:29  vasilche
* Fixed ambiguity on MSVC.
*
* Revision 1.70  2003/05/20 16:18:42  vasilche
* Fixed compilation errors on GCC.
*
* Revision 1.69  2003/05/20 15:44:37  vasilche
* Fixed interaction of CDataSource and CDataLoader in multithreaded app.
* Fixed some warnings on WorkShop.
* Added workaround for memory leak on WorkShop.
*
* Revision 1.68  2003/05/13 20:27:05  vasilche
* Added lost SAT SATKEY info in error message.
*
* Revision 1.67  2003/05/13 20:21:10  vasilche
* *** empty log message ***
*
* Revision 1.66  2003/05/13 20:14:40  vasilche
* Catching exceptions and reconnection were moved from readers to genbank loader.
*
* Revision 1.65  2003/05/13 18:32:29  vasilche
* Fixed use of GBLOG_POST() macro.
*
* Revision 1.64  2003/05/12 19:18:29  vasilche
* Fixed locking of object manager classes in multi-threaded application.
*
* Revision 1.63  2003/05/12 18:26:08  vasilche
* Removed buggy _ASSERT() on reconnection.
*
* Revision 1.62  2003/05/06 16:52:28  vasilche
* Try to reconnect to genbank on any exception.
*
* Revision 1.61  2003/04/29 19:51:13  vasilche
* Fixed interaction of Data Loader garbage collector and TSE locking mechanism.
* Made some typedefs more consistent.
*
* Revision 1.60  2003/04/24 16:12:38  vasilche
* Object manager internal structures are splitted more straightforward.
* Removed excessive header dependencies.
*
* Revision 1.59  2003/04/18 17:38:01  grichenk
* Use GENBANK_LOADER_METHOD env. variable to specify GB readers.
* Default is "PUBSEQOS:ID1".
*
* Revision 1.58  2003/04/15 16:32:29  dicuccio
* Added include for I_DriverContext from DBAPI library - avoids concerning
* warning about deletion of unknwon type.
*
* Revision 1.57  2003/04/15 15:30:15  vasilche
* Added include <memory> when needed.
* Removed buggy buffer in printing methods.
* Removed unnecessary include of stream_util.hpp.
*
* Revision 1.56  2003/04/15 14:24:08  vasilche
* Changed CReader interface to not to use fake streams.
*
* Revision 1.55  2003/03/21 19:22:51  grichenk
* Redesigned TSE locking, replaced CTSE_Lock with CRef<CTSE_Info>.
*
* Revision 1.54  2003/03/11 15:51:06  kuznets
* iterate -> ITERATE
*
* Revision 1.53  2003/03/05 20:54:41  vasilche
* Commented out wrong assert().
*
* Revision 1.52  2003/03/03 20:34:51  vasilche
* Added NCBI_THREADS macro - it's opposite to NCBI_NO_THREADS.
* Avoid using _REENTRANT macro - use NCBI_THREADS instead.
*
* Revision 1.51  2003/03/01 23:07:42  kimelman
* bugfix: MTsafe
*
* Revision 1.50  2003/03/01 22:27:57  kimelman
* performance fixes
*
* Revision 1.49  2003/02/27 21:58:26  vasilche
* Fixed performance of Object Manager's garbage collector.
*
* Revision 1.48  2003/02/26 18:03:31  vasilche
* Added some error check.
* Fixed formatting.
*
* Revision 1.47  2003/02/24 18:57:22  vasilche
* Make feature gathering in one linear pass using CSeqMap iterator.
* Do not use feture index by sub locations.
* Sort features at the end of gathering in one vector.
* Extracted some internal structures and classes in separate header.
* Delay creation of mapped features.
*
* Revision 1.46  2003/02/05 17:59:17  dicuccio
* Moved formerly private headers into include/objects/objmgr/impl
*
* Revision 1.45  2003/01/22 20:11:54  vasilche
* Merged functionality of CSeqMapResolved_CI to CSeqMap_CI.
* CSeqMap_CI now supports resolution and iteration over sequence range.
* Added several caches to CScope.
* Optimized CSeqVector().
* Added serveral variants of CBioseqHandle::GetSeqVector().
* Tried to optimize annotations iterator (not much success).
* Rewritten CHandleRange and CHandleRangeMap classes to avoid sorting of list.
*
* Revision 1.44  2002/12/26 20:53:02  dicuccio
* Moved tse_info.hpp -> include/ tree.  Minor tweaks to relieve compiler
* warnings in MSVC.
*
* Revision 1.43  2002/11/08 19:43:35  grichenk
* CConstRef<> constructor made explicit
*
* Revision 1.42  2002/11/04 21:29:12  grichenk
* Fixed usage of const CRef<> and CRef<> constructor
*
* Revision 1.41  2002/09/09 16:12:43  dicuccio
* Fixed minor typo ("conneciton" -> "connection").
*
* Revision 1.40  2002/07/24 20:30:27  ucko
* Move CTimerGuard out to file scope to fix a MIPSpro compiler core dump.
*
* Revision 1.39  2002/07/22 22:53:24  kimelman
* exception handling fixed: 2level mutexing moved to Guard class + added
* handling of confidential data.
*
* Revision 1.38  2002/06/04 17:18:33  kimelman
* memory cleanup :  new/delete/Cref rearrangements
*
* Revision 1.37  2002/05/14 20:06:26  grichenk
* Improved CTSE_Info locking by CDataSource and CDataLoader
*
* Revision 1.36  2002/05/10 16:44:32  kimelman
* tuning to allow pubseq enable build
*
* Revision 1.35  2002/05/08 22:23:48  kimelman
* MT fixes
*
* Revision 1.34  2002/05/06 03:28:47  vakatov
* OM/OM1 renaming
*
* Revision 1.33  2002/05/03 21:28:10  ucko
* Introduce T(Signed)SeqPos.
*
* Revision 1.32  2002/04/30 18:56:54  gouriano
* added multithreading-related initialization
*
* Revision 1.31  2002/04/28 03:36:47  vakatov
* Temporarily turn off R1.30(b) unconditionally -- until it is buildable
*
* Revision 1.30  2002/04/26 16:32:23  kimelman
* a) turn on GC b) turn on PubSeq where sybase is available
*
* Revision 1.29  2002/04/12 22:54:28  kimelman
* pubseq_reader auto call commented per Denis request
*
* Revision 1.28  2002/04/12 21:10:33  kimelman
* traps for coredumps
*
* Revision 1.27  2002/04/11 18:45:35  ucko
* Pull in extra headers to make KCC happy.
*
* Revision 1.26  2002/04/10 22:47:56  kimelman
* added pubseq_reader as default one
*
* Revision 1.25  2002/04/09 19:04:23  kimelman
* make gcc happy
*
* Revision 1.24  2002/04/09 18:48:15  kimelman
* portability bugfixes: to compile on IRIX, sparc gcc
*
* Revision 1.23  2002/04/05 23:47:18  kimelman
* playing around tests
*
* Revision 1.22  2002/04/04 01:35:35  kimelman
* more MT tests
*
* Revision 1.21  2002/04/02 17:27:00  gouriano
* bugfix: skip test for yet unregistered data
*
* Revision 1.20  2002/04/02 16:27:20  gouriano
* memory leak
*
* Revision 1.19  2002/04/02 16:02:31  kimelman
* MT testing
*
* Revision 1.18  2002/03/30 19:37:06  kimelman
* gbloader MT test
*
* Revision 1.17  2002/03/29 02:47:04  kimelman
* gbloader: MT scalability fixes
*
* Revision 1.16  2002/03/27 20:23:50  butanaev
* Added connection pool.
*
* Revision 1.15  2002/03/26 23:31:08  gouriano
* memory leaks and garbage collector fix
*
* Revision 1.14  2002/03/26 15:39:24  kimelman
* GC fixes
*
* Revision 1.13  2002/03/25 17:49:12  kimelman
* ID1 failure handling
*
* Revision 1.12  2002/03/25 15:44:46  kimelman
* proper logging and exception handling
*
* Revision 1.11  2002/03/22 18:56:05  kimelman
* GC list fix
*
* Revision 1.10  2002/03/22 18:51:18  kimelman
* stream WS skipping fix
*
* Revision 1.9  2002/03/22 18:15:47  grichenk
* Unset "skipws" flag in binary stream
*
* Revision 1.8  2002/03/21 23:16:32  kimelman
* GC bugfixes
*
* Revision 1.7  2002/03/21 21:39:48  grichenk
* garbage collector bugfix
*
* Revision 1.6  2002/03/21 19:14:53  kimelman
* GB related bugfixes
*
* Revision 1.5  2002/03/21 01:34:53  kimelman
* gbloader related bugfixes
*
* Revision 1.4  2002/03/20 21:24:59  gouriano
* *** empty log message ***
*
* Revision 1.3  2002/03/20 19:06:30  kimelman
* bugfixes
*
* Revision 1.2  2002/03/20 17:03:24  gouriano
* minor changes to make it compilable on MS Windows
*
* Revision 1.1  2002/03/20 04:50:13  kimelman
* GB loader added
*
* ===========================================================================
*/