log query.cpp @ 770:c54bc2ffbf92 tip

age author description
Sat, 15 Oct 2011 17:28:07 +0000 mas01mc Adding Kullback-Leibler divergence as alternate distance function
Fri, 26 Nov 2010 08:15:08 +0000 mas01mc Minor, somewhat cosmetic but currently wasteful, memory arithmetic in qpointers allocation mkc_lsh_update
Thu, 25 Nov 2010 13:42:40 +0000 mas01mc Adding mkc_lsh_update branch, trunk candidate with improved LSH: merged trunk 1095 and branch multiprobe_lsh fewerQueryDatumReads
Fri, 05 Mar 2010 16:01:06 +0000 mas01cr First implementation of separate query / instance hopsizes
Fri, 05 Mar 2010 16:00:53 +0000 mas01cr Extend results structure to include a key for the query.
Thu, 28 Jan 2010 10:23:42 +0000 mas01cr Better treatment of refine.hopsize
Fri, 28 Aug 2009 17:14:06 +0000 mas01cr MORE MEMORY SANITY
Wed, 12 Aug 2009 14:57:11 +0000 mas01cr Test for audiodb_retrieve_datum()
Tue, 11 Aug 2009 21:42:49 +0000 mas01cr Take advantage of those new handy _or_goto_error macros
Tue, 11 Aug 2009 21:42:13 +0000 mas01cr Header file rearrangement
Tue, 14 Jul 2009 15:35:36 +0000 mas01cr API for retrieving a track's data.
Sun, 08 Feb 2009 22:32:33 +0000 mas01mc Rework new audiodb_track_id_datum_offset() and audiodb_insert_create_datum_offset() so that all offsets are in num_vectors. Internals of data storage are left the the methods to implement. If we are going to export these methods to the API, how do we expose the cache ? It needs to be passed back to the user and back in by the user on each call for caching to be re-entrant. No (intentional) functional changes this version. multiprobeLSH
Sun, 08 Feb 2009 15:53:57 +0000 mas01mc Expanded interface to audiodb_insert_create_datum() and audiodb_track_id_datum() to include _offset versions of both. Default cache is NULL, but offset and size parameters required. multiprobeLSH
Sun, 08 Feb 2009 05:06:01 +0000 mas01mc Fixed non-NULL memory initializaton error. We require NULL fields for correct free() behaviour. multiprobeLSH
Sat, 07 Feb 2009 16:59:31 +0000 mas01mc Added file caching for sparse datum reads. This required making a new type called adb_fd_cache_t and modifying the read methods audiodb_track_id_datum() and audiodb_insert_create_datum() to use the cache struct if one is provided by the user. multiprobeLSH
Sat, 07 Feb 2009 01:20:05 +0000 mas01mc Optimized the query_loop_points inner loop for memcpy and I/O efficiency. Uses sparse seeks and reads to perform scattered reads across data set. Current version does not cache fid between open calls to the same trackID. multiprobeLSH
Sat, 31 Jan 2009 16:25:30 +0000 mas01mc Removed redundant memcpy calls from query_loop_queue(). This changed the order of some memory allocation and de-allocation calls for loaded database data. multiprobeLSH
Fri, 30 Jan 2009 19:44:51 +0000 mas01mc Completed audiodb_datum_qpointers_partial(), fixed bug that was scribbling over pre-computed l2norms when points arrived in original less<PointPair> order, switched to greater<PointPair> order and all is fixed. multiprobeLSH
Fri, 30 Jan 2009 14:58:35 +0000 mas01mc Fixed a problem with inserting into exact_evaluation_queue, duplicate pairs were being inserted when querying multiple hashtables. That is now fixed by adding a unique constraint using set<PointPair> inserts and testing. This should speed things up quite a bit (like, 10-fold with default num tables). multiprobeLSH
Thu, 29 Jan 2009 13:44:47 +0000 mas01mc Added audiodb_datum_qpointers_partial() to make exact evaluation of LSH results more efficient. This routine calculates partial sums only for those vectors needed. multiprobeLSH
Wed, 28 Jan 2009 18:55:46 +0000 mas01mc Added more debugging information: report numPairs returned from LSH bucket retrieval (pre-distance computation), entry/exit from audiodb_query_spec_qpointers(...) multiprobeLSH
Thu, 15 Jan 2009 13:57:33 +0000 mas01cr Code rearrangements to tease apart library code from C++ audioDB code.
Sat, 10 Jan 2009 16:47:57 +0000 mas01cr Merge api-inversion branch (-r656:771, but I don't expect to return to
Sat, 10 Jan 2009 13:14:05 +0000 mas01mc added free() call for data_buffer at end of query_loop_points() memory-leaks
Sat, 10 Jan 2009 13:12:52 +0000 mas01mc removed defunct EXACT_MULT from query_loop_points() memory-leaks
Fri, 09 Jan 2009 22:12:57 +0000 mas01mc modified point-bits coding for indexing (BREAKS BACKWARDS COMPATIBILITY). Plugged file-descriptor leak in index.cpp
Fri, 09 Jan 2009 18:05:32 +0000 mas01mc Fixed memory leaks, added WS --no_unit_norming, and removed capping of LSH_N_POINT_BITS to 15 bits, instead allow any number of bits to encode points, remaining bits encode tracks
Tue, 06 Jan 2009 16:27:01 +0000 mas01cr const declarations for some API arguments. api-inversion
Tue, 06 Jan 2009 15:25:39 +0000 mas01cr Clean up some compiler warnings. api-inversion
Wed, 31 Dec 2008 15:44:16 +0000 mas01cr Move audioDB::query over to audioDB.cpp api-inversion
Wed, 31 Dec 2008 15:44:12 +0000 mas01cr YAY! api-inversion
Wed, 31 Dec 2008 12:25:22 +0000 mas01cr Almost there! api-inversion
Tue, 30 Dec 2008 23:56:57 +0000 mas01cr No more audioDB::index_allocate and audioDB::index_init_query api-inversion
Tue, 30 Dec 2008 15:38:59 +0000 mas01cr audioDB::query_loop and audioDB::query_loop_points are no more. api-inversion
Tue, 30 Dec 2008 15:38:55 +0000 mas01cr Simplify audioDB::query_loop_points. api-inversion
Tue, 30 Dec 2008 15:38:50 +0000 mas01cr Factor out some bits of audiodb_query_spec_qpointers() api-inversion
Tue, 30 Dec 2008 10:36:01 +0000 mas01cr Convert index-finding logic to C functions. api-inversion
Sun, 28 Dec 2008 22:43:50 +0000 mas01cr No more global shingle vector of vectors. api-inversion
Sun, 28 Dec 2008 18:44:08 +0000 mas01cr Start using the query state structure. api-inversion
Wed, 24 Dec 2008 10:57:23 +0000 mas01cr Begin pushing adb_t *adb into index_query_loop and query_loop_points api-inversion
Wed, 24 Dec 2008 10:57:18 +0000 mas01cr No more indata api-inversion
Wed, 24 Dec 2008 10:57:14 +0000 mas01cr Almost finish with audioDB::query_loop. api-inversion
Wed, 24 Dec 2008 10:57:09 +0000 mas01cr No more audioDB::error in audioDB::query_loop api-inversion
Wed, 24 Dec 2008 10:57:05 +0000 mas01cr Miscellaneous trivial de-audioDB::ifications api-inversion
Wed, 24 Dec 2008 10:57:01 +0000 mas01cr No more trackOffsetTable api-inversion
Wed, 24 Dec 2008 10:56:56 +0000 mas01cr use audiodb_key_index() in audiodb_query_spec_qpointers() api-inversion
Wed, 24 Dec 2008 10:56:53 +0000 mas01cr audiodb_set_up_dbpointers() is a better name than audiodb_set_up_db(). api-inversion
Wed, 24 Dec 2008 10:56:49 +0000 mas01cr LESS VERBOSITY api-inversion
Wed, 24 Dec 2008 10:56:45 +0000 mas01cr Incrementally zapping bits of code. api-inversion
Wed, 24 Dec 2008 10:56:41 +0000 mas01cr Better memory discipline api-inversion
Wed, 24 Dec 2008 10:56:37 +0000 mas01cr No more audioDB::set_up_query{,_from_key} api-inversion
Wed, 24 Dec 2008 10:56:33 +0000 mas01cr New (internal) function audiodb_query_spec_qpointers() api-inversion
Wed, 24 Dec 2008 10:56:16 +0000 mas01cr audioDB::initialize_arrays no longer depends on anything in the C++ class. api-inversion
Wed, 24 Dec 2008 10:56:12 +0000 mas01cr audioDB::set_up_db can use an adb_qpointers_internal_t... api-inversion
Wed, 24 Dec 2008 10:56:07 +0000 mas01cr Hey, who let audioDB::set_up_query into the room? api-inversion
Wed, 24 Dec 2008 10:56:01 +0000 mas01cr really zap all unnecessary sequenceLength from query.cpp api-inversion
Wed, 24 Dec 2008 10:55:56 +0000 mas01cr begin pushing an adb_query_spec_t * all the way through query. api-inversion
Wed, 24 Dec 2008 10:55:52 +0000 mas01cr adb_t-ize most of audioDB::set_up_db. api-inversion
Wed, 24 Dec 2008 10:55:48 +0000 mas01cr More on the query rationalization api-inversion
Wed, 24 Dec 2008 10:55:40 +0000 mas01cr Punishment gluttony. api-inversion
Wed, 24 Dec 2008 10:55:36 +0000 mas01cr Make the keys std::set in adb_t a std::map instead. api-inversion
Wed, 24 Dec 2008 10:55:24 +0000 mas01cr Move sequence_foo() functions out of audioDB:: namespace... api-inversion
Wed, 24 Dec 2008 10:55:20 +0000 mas01cr Deal with audioDB::UnitNorm api-inversion
Wed, 24 Dec 2008 10:55:16 +0000 mas01cr Begin pushing information through. api-inversion
Wed, 24 Dec 2008 10:55:12 +0000 mas01cr Accumulators everywhere! api-inversion
Wed, 24 Dec 2008 10:55:08 +0000 mas01cr Cheap hackery abounds. api-inversion
Wed, 24 Dec 2008 10:55:03 +0000 mas01cr Baby steps... api-inversion
Fri, 05 Dec 2008 22:32:49 +0000 mas01cr invert audioDB::insert / audiodb_insert(). api-inversion
Wed, 12 Nov 2008 15:23:29 +0000 mas01cr check return values of read() and write(), erroring out on an error gcc-4.3-cleanups
Tue, 07 Oct 2008 18:41:38 +0000 mas01mc make error message reflect boundary condition when queryPoint >= numVectors-sequenceLength+1
Sun, 21 Sep 2008 09:33:57 +0000 mas01mc fixed error when queryPoint was one-past-the-end of the query track. Added test for this case.
Sat, 30 Aug 2008 14:33:24 +0000 mas01mc Added some extra logging to alert the user to which query is called
Thu, 21 Aug 2008 21:28:33 +0000 mas01mc Merge of branches/large_adb -r 514:524 onto the trunk. No conflicts. Added LARGE_ADB support. Turn on with --ntracks 20001 or greater. Use --adb_feature_root to locate feature files at QUERY time. A bug fix in LSH indexing that was incorrectly thresholding large numbers of shingles.
Thu, 21 Aug 2008 19:16:21 +0000 mas01mc Added --adb_feature_root=path filename prefix for locating LARGE_ADB features with relative paths at QUERY time. Also added convenience argument --adb_root=path prefix for -d database command option. large_adb
Wed, 20 Aug 2008 13:50:58 +0000 mas01mc working LARGE_ADB support. Activiate at creation time with -N --ntracks 20001 or greater, or with --datasize 1356 or greater, or both. LARGE_ADB blocks non-indexed QUERY. large_adb
Tue, 19 Aug 2008 20:27:15 +0000 mas01mc Added indexing support for O2_FLAG_LARGE_ADB. Tested on indexed query by features. No indexed query-by-key yet. No --lsh_exact yet. large_adb
Tue, 12 Aug 2008 14:25:51 +0000 mas01mc Fixed memory leak in query_loop_points(). Was causing memalloc failure on multiple WS queries with lsh_exact flag on
Tue, 12 Aug 2008 01:21:44 +0000 mas01mc working SIIGRAPH08 version. Fixed powerTable mmap memory leak in WS calls (only showed up in big databases). Implements radius queries over WS with new wsdl file
Thu, 07 Aug 2008 01:53:38 +0000 mas01mc Added server-side loading of persistent index (LSH hashtables) via --load_index -d dbName -R radius -l sequenceLength. Queries using these parameters will lookup the memory-resident hashtable instead of loading one from disk.
Tue, 29 Jul 2008 22:01:17 +0000 mas01mc Completed merge of lshlib back to current version of the trunk.
Tue, 01 Jul 2008 09:12:40 +0000 mas01cr Merge trunk changes -r326:386 into audiodb-debian branch. audiodb-debian
Mon, 09 Jun 2008 19:20:39 +0000 mas01mc Added new query type "onetoonensequence" for matching each query sequence to the single closest match in the database.
Wed, 16 Apr 2008 09:59:43 +0000 mas01cr Commit patch deprecating the --size argument, replacing it with
Mon, 14 Apr 2008 15:36:29 +0000 map01bf merged in the trunk updates to the adding-emd branch, also added the emd.c and emd.h support files. Actually starting the feature integration now adding-emd
Mon, 18 Feb 2008 10:47:03 +0000 mas01mc Added trackSequenceQueryRadNNReporter: orders retrieved tracks by one-to-one point counts per track and
Sun, 17 Feb 2008 14:56:02 +0000 mas01mc Added new query type: nsequence : this reports the n-nearest neighbours for each track in a sequence averaging query.
Fri, 14 Dec 2007 14:41:37 +0000 mas01cr Merge trunk changes -r290:313 into audiodb-debian branch. audiodb-debian
Thu, 13 Dec 2007 14:28:54 +0000 mas01cr OK, time for trackSequenceQueryNN to die.
Thu, 13 Dec 2007 14:23:32 +0000 mas01cr Merge refactoring branch.
Wed, 12 Dec 2007 16:43:04 +0000 mas01cr VERB_LOG macro into audioDB.h refactoring
Wed, 12 Dec 2007 11:10:48 +0000 mas01cr Template classize pointQueryReporter. refactoring
Wed, 12 Dec 2007 10:58:24 +0000 mas01cr OAOO for the trackPointQuery and trackSequenceQueryNN Reporter classes. refactoring
Wed, 12 Dec 2007 10:22:31 +0000 mas01cr Only one call to trackSequenceQueryNN() refactoring
Mon, 10 Dec 2007 20:01:43 +0000 mas01cr Proof of Reporter concept: restore pointQuery and trackPointQuery refactoring
Mon, 10 Dec 2007 18:25:55 +0000 mas01cr Code defensively in sequence_sum() refactoring
Thu, 06 Dec 2007 14:41:07 +0000 mas01cr And now, the big one: two into one will go. refactoring
Wed, 05 Dec 2007 12:47:33 +0000 mas01cr Move usingQueryPoint logic into set_up_query() refactoring
Wed, 05 Dec 2007 12:14:00 +0000 mas01cr Fix for tests/0035 refactoring
Wed, 05 Dec 2007 11:09:45 +0000 mas01cr Move query/db times setup into set_up_query() and set_up_db() refactoring
Tue, 04 Dec 2007 16:48:23 +0000 mas01cr Compute durations more correctly. refactoring
Tue, 04 Dec 2007 12:47:49 +0000 mas01cr set_up_db() analogue to set_up_query() refactoring
Tue, 04 Dec 2007 10:23:28 +0000 mas01cr Comment fixes. refactoring
Tue, 04 Dec 2007 10:17:30 +0000 mas01cr Move some more code into set_up_query() refactoring
Tue, 04 Dec 2007 09:40:55 +0000 mas01cr set_up_query() refactoring
Fri, 30 Nov 2007 10:30:01 +0000 mas01cr Trivial cleanups of UnitNorm() refactoring
Fri, 30 Nov 2007 10:19:26 +0000 mas01cr delete_arrays() function. refactoring
Thu, 29 Nov 2007 10:18:04 +0000 mas01cr Make reading the data in be its own function refactoring
Wed, 28 Nov 2007 17:46:02 +0000 mas01cr New initialize_arrays() helper routine for remaining queries. refactoring
Wed, 28 Nov 2007 17:22:42 +0000 mas01cr Move some code around a little. refactoring
Wed, 28 Nov 2007 17:04:09 +0000 mas01cr Delete pointQuery() and trackPointQuery() refactoring
Wed, 28 Nov 2007 15:10:28 +0000 mas01cr Filewise refactor. refactoring