| Wed, 12 Nov 2008 14:01:09 +0000 |
mas01cr |
More parentheses in ALIGN_FOO_UP() macros.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 14:01:06 +0000 |
mas01cr |
Initialize UseApiError to 0 (false) in audioDB constructors.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:14 +0000 |
mas01cr |
Add some braces to avoid ambiguous `else's.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:13 +0000 |
mas01cr |
Fix a buffer-overrun detected (statically!) by gcc
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:12 +0000 |
mas01cr |
one more bit of const correctness.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:10 +0000 |
mas01cr |
YET MORE CONST-CORRECTNESS.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:09 +0000 |
mas01cr |
$(LIBRARY) makefile target doesn't need GSOAP_INCLUDE.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 13:05:08 +0000 |
mas01cr |
apparently main()'s argc is an int, not an unsigned. Who knew?
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 12:02:49 +0000 |
mas01cr |
Branch for cleaning up all the gcc-4.3 warnings and errors.
gcc-4.3-cleanups
|
| Wed, 12 Nov 2008 10:21:06 +0000 |
mas01ik |
First version of the API, committed to the main trunk. Thanks Christophe, for all the help!
|
| Fri, 07 Nov 2008 12:48:26 +0000 |
mas01mc |
Added serialized feature query Web Service, server-side only, example Python script: QueryADB.py -v
pre_api
|
| Fri, 07 Nov 2008 12:18:43 +0000 |
mas01mc |
serialized features query complete and tested
serializedQuery
|
| Thu, 06 Nov 2008 22:57:39 +0000 |
mas01mc |
serialized query now with power.
serializedQuery
|
| Thu, 06 Nov 2008 16:07:57 +0000 |
mas01mc |
completed serialized shingle query, with some initial testing using QueryADB.py -v
serializedQuery
|
| Thu, 06 Nov 2008 08:54:40 +0000 |
mas01mc |
serialized shingle query added to WS, compiles but not tested
serializedQuery
|
| Mon, 27 Oct 2008 11:09:34 +0000 |
mas01mc |
Adding QueryADB.py python library for audioDB query over SOAP
serializedQuery
|
| Mon, 27 Oct 2008 11:08:13 +0000 |
mas01mc |
New branch for serialized queries over SOAP
serializedQuery
|
| Wed, 08 Oct 2008 17:23:15 +0000 |
mas01cr |
remove -h flag for sequencehop, restoring buildability with
|
| Wed, 08 Oct 2008 13:55:16 +0000 |
mas01cr |
Fix for building cmdline.c with required architecture flags on OS X.
|
| Wed, 08 Oct 2008 11:06:11 +0000 |
mas01cr |
Quieten newer compilers with MORE CONST CORRECTNESS.
|
| Tue, 07 Oct 2008 21:03:26 +0000 |
mas01mc |
unionized punning of the sbucket.snext field into {subucket* ptr, unsigned numBuckets}snext; so that list-head code is 64-bit friendly.
|
| Tue, 07 Oct 2008 20:08:22 +0000 |
mas01mc |
added portable C++ reinterpret_cast<Uns32T> instead of C-style cast to fix 64-bit compile error in query.cpp. This is in advance of re-writing bucket class to include a union{} of pointer and counter fields.
|
| 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.
|
| Wed, 10 Sep 2008 18:55:16 +0000 |
mas01mc |
Added LSH_CORE_ARRAY structure for hashtables instead of linked lists. Maintained Backwards Compatibiliity with indexes build for linked list format. Added tests for indexing and merging. Tested backwards compatibility OK.\n\n The purpose of the LSH_CORE_ARRAY data structure is greater space efficiency and L1/2 cache usage. Essential for multiple indexes with multiple hashtables in RAM
|
| Fri, 05 Sep 2008 20:01:58 +0000 |
mas01mc |
Added new INDEX tests: LARGE_ADB, PATH PREFIX, cmdline and WS. Removed redundant feature PREFIX in audioDB.cpp. Moved tests/0050 to tests/0041 so INDEX tests are all in 0040-0049 range.
|
| Fri, 05 Sep 2008 16:40:41 +0000 |
mas01mc |
fixed startServer() reporting of non-locatable database or index. Achieved by setting isServer=1 AFTER checking database and loading INDEX.
|
| Fri, 05 Sep 2008 14:35:44 +0000 |
mas01mc |
gengetopt does not support multiple dependencies, so added test for database_given for INDEX command. Still cannot figure out why sometimes error(...) throws char* exception and SIG_ABORT
|
| Fri, 05 Sep 2008 14:16:21 +0000 |
mas01mc |
converted read/write into fread/fwrite for LSH hashtable serialize and unserialize. INDEXING is now faster.
|
| Tue, 02 Sep 2008 17:10:59 +0000 |
mas01mc |
Added test for --LISZT over command-line and WS, plus --lisztOffset and --lisztLength
|
| Tue, 02 Sep 2008 16:16:59 +0000 |
mas01mc |
Added command-line and WS methods to (LI)st key-(S)trings and si(Z)es of (T)racks --LISZT --lisztOffset offset --lisztLength len
|
| Mon, 01 Sep 2008 15:35:05 +0000 |
mas01cr |
Ugly hack upon ugly hack:
|
| Sun, 31 Aug 2008 08:45:22 +0000 |
mas01mc |
Added new tests for indexed query over WS. Fixed potential NULL string dereference in audioDB.cpp
|
| Sat, 30 Aug 2008 23:35:13 +0000 |
mas01mc |
fixed another potential NULL string de-reference in soap.cpp
|
| Sat, 30 Aug 2008 15:47:53 +0000 |
mas01mc |
fixed null string reference in WS adb__sequenceQueryByKey()
|
| Sat, 30 Aug 2008 14:33:24 +0000 |
mas01mc |
Added some extra logging to alert the user to which query is called
|
| Sat, 30 Aug 2008 13:41:50 +0000 |
mas01mc |
Added WS support for Radius search by featureFileName via adb__sequenceQueryByKey(). Added relative path support for non-large ADB.
|
| Fri, 29 Aug 2008 20:16:10 +0000 |
mas01mc |
Added INDEX and LSH query test
|
| Fri, 29 Aug 2008 18:26:21 +0000 |
mas01mc |
reverted tests/0020/run-test.sh
|
| Fri, 29 Aug 2008 16:39:53 +0000 |
mas01mc |
Fixed indexing error in non-large ADB introduced with large ADB.
|
| 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 21:02:14 +0000 |
mas01mc |
Fixed an indexing bug where rest of track was ignored after first shingle with power below threshold. Put default O2_LSH_POINT_BITS back to 14 (16384 points), can be altered at compile time with CFLAGS+=-DO2_LSH_POINT_BITS n
large_adb
|
| Thu, 21 Aug 2008 19:58:55 +0000 |
mas01mc |
Added --adb_root and --adb_feature_root support to server-side Web Services instance. Client does not have to know location of database and features on the server when server performs path substitutions with these arguments.
large_adb
|
| 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, 19 Aug 2008 15:50:26 +0000 |
mas01mc |
Ammended batchinsert and status to memory map featureFilenameTables, powerFileNameTable and timesFileNameTable for O2_FLAG_LARGE_ADB
large_adb
|
| Tue, 19 Aug 2008 14:38:39 +0000 |
mas01mc |
Added large_adb Support for batchinsert and status commands
large_adb
|
| Tue, 19 Aug 2008 14:27:21 +0000 |
mas01mc |
Adding large_adb support (up to 1M tracks)
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
|
| Sat, 09 Aug 2008 15:46:42 +0000 |
mas01mc |
Check for index-exists on WS --load_index and exit if it does not. Revert unmap_tables because cleanup() did it fine, but there is a memory leak in ws_query calls.
|
| Sat, 09 Aug 2008 15:17:01 +0000 |
mas01mc |
unmap tables on ws_query exit.
|
| Sat, 09 Aug 2008 15:04:16 +0000 |
mas01mc |
Added sequencehop on Indexing to optionally reduce size of large indexes. Added cleanup() call on WS query exit to fix MMAP memory leak
|
| Fri, 08 Aug 2008 14:01:33 +0000 |
mas01mc |
Added Web Services support for --exhaustive and --lsh_exact flags in Radius sequence searches.
|
| Thu, 07 Aug 2008 17:05:22 +0000 |
mas01mc |
Fixed persistent index bug where indexName identifier was an un protected string pointer rather than a copy of the string. Hence the index name was getting deleted and the test for persistent index was failing.
|
| 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.
|
| Wed, 06 Aug 2008 21:23:14 +0000 |
mas01mc |
Added reporters radius search via Web Services. The ordering for radius search is degenerate if usingQueryPoint, because the query point can only be counted once. This behaviour is changed by specifying -n 1 to emulate the trackAveragingReporter ordering behaviour
|
| Tue, 05 Aug 2008 22:40:38 +0000 |
mas01mc |
changed FORMAT2 index serialization so that token bits don't scribble over point index for >=32768 tracks
|
| Tue, 05 Aug 2008 15:34:10 +0000 |
mas01cr |
Guard against too-short tracks. Fixes a segfault on attempting to index
|
| Tue, 05 Aug 2008 13:38:36 +0000 |
mas01cr |
Wow, so many bugs in one short stanza of code. Really fix batchinsert
|
| Mon, 04 Aug 2008 10:01:47 +0000 |
mas01cr |
Whoops. My fault: time() takes an argument.
|
| Mon, 04 Aug 2008 10:00:34 +0000 |
mas01cr |
Implement O((M+N) log(M+N)) duplicate key detection in batchinsert,
|
| Fri, 01 Aug 2008 16:21:51 +0000 |
mas01mc |
added more informative status display for flags
|
| Fri, 01 Aug 2008 15:48:40 +0000 |
mas01tc |
Corrected error message in audioDB line 332
|
| Fri, 01 Aug 2008 15:45:10 +0000 |
mas01mc |
fixed recently-introduced strict point ordering bug when LSH index is initially empty
|
| Fri, 01 Aug 2008 15:33:00 +0000 |
mas01tc |
Fixed bug in audioDB.cpp line 331:
|
| Fri, 01 Aug 2008 15:13:21 +0000 |
mas01mc |
set the default flag value to 'off' for --lsh_on_disk and made lsh_in_core the compliment of this value
|
| Fri, 01 Aug 2008 15:04:31 +0000 |
mas01mc |
fixed file size allocation for FORMAT2 files. Made LSH index size() in bytes an unsigned long long. Changed the name of lsh_inCore flag to lsh_on_disk (to reverse the sense of the 'flag').
|
| Thu, 31 Jul 2008 19:26:04 +0000 |
mas01mc |
added cap on number of vectors per track to index: this cap is set relative to the maximum pointID available in the dual <trackID,pointID> indexing scheme that is used to identify points and tracks using a single 32-bit unsigned integer that is passed to lshlib.
|
| Wed, 30 Jul 2008 18:00:50 +0000 |
mas01mc |
fixed UNIT_TEST_LSH.cpp to work with refactored lshlib and fixed default LSH file FORMAT on serialization.
|
| Wed, 30 Jul 2008 15:22:22 +0000 |
mas01mc |
Refactored LSH interface to separate hashfunctions and parameters from insertion/retrieval/serialization
|
| Tue, 29 Jul 2008 22:01:17 +0000 |
mas01mc |
Completed merge of lshlib back to current version of the trunk.
|
| Tue, 22 Jul 2008 20:09:31 +0000 |
mas01mc |
Adding to tests/pointset_test OCTAVE function to plot LSH retrieval characteristic from test output
|
| Tue, 22 Jul 2008 18:04:04 +0000 |
mas01mc |
slight modification of LSH point_set testing
|
| Thu, 17 Jul 2008 14:43:42 +0000 |
mas01mc |
abstracted RADII as list at top of script
|
| Thu, 17 Jul 2008 10:06:42 +0000 |
mas01cr |
Don't take away the -X shortcut for sampling: make it easier to merge in
|
| Wed, 16 Jul 2008 18:28:28 +0000 |
mas01mc |
extensive improvement to LSH point-set testing and LSH parameter estimation in Octave
|
| Tue, 15 Jul 2008 20:46:25 +0000 |
mas01mc |
generated points test, requires compile of audioDB with lshlib and -DQUERY_WITHOUT_NORM. Compares LSHLIB and EXACT retrieval of points on spherical shells at several radii
|
| Mon, 14 Jul 2008 21:50:47 +0000 |
mas01mc |
test for lshlib with point-set generator independent of normalization so that statistics for (P1,P2,R,cR)-sensitivity to k,L and R can be gathered.
|
| Mon, 07 Jul 2008 08:57:06 +0000 |
mas01cr |
Really finish with the sampling branch, this time merging all of it,
|
| Thu, 03 Jul 2008 08:50:57 +0000 |
mas01cr |
Trivial fix for some compilers trying to be smart and noting a path
|
| Wed, 02 Jul 2008 17:14:10 +0000 |
map01bf |
modifed the makefile to correctly link to gsl on a mac
|
| Wed, 02 Jul 2008 14:20:28 +0000 |
mas01cr |
Some tidying up post-merge of sampling branch:
|
| Wed, 02 Jul 2008 14:07:10 +0000 |
mas01cr |
Merge sampling branch (-r361:405, though I hope that the branch is now
|
| Wed, 02 Jul 2008 13:53:23 +0000 |
mas01cr |
make the RNG a part of the audioDB object.
sampling
|
| Tue, 01 Jul 2008 22:17:33 +0000 |
mas01cr |
Use the GNU Scientific Library implementation of MT19937 for generating
sampling
|
| Tue, 01 Jul 2008 09:12:40 +0000 |
mas01cr |
Merge trunk changes -r326:386 into audiodb-debian branch.
audiodb-debian
|
| Tue, 01 Jul 2008 09:00:29 +0000 |
mas01cr |
Restore static decoration to yfun and yinv.
sampling
|
| Mon, 30 Jun 2008 17:38:59 +0000 |
mas01mc |
fixed bug in Radius reporters where *worst* n tracks were being reporter, needed std::greater<Radresult> in priority_queue
|
| Mon, 23 Jun 2008 09:49:22 +0000 |
mas01cr |
Let the user specify the number of samples to take on the command-line.
sampling
|
| Tue, 17 Jun 2008 12:27:01 +0000 |
mas01cr |
Make yinv and yfun not static, so that they can be called from gdb.
sampling
|
| Tue, 17 Jun 2008 11:43:40 +0000 |
mas01cr |
Add utility for quickly computing x_threshold value.
sampling
|
| Mon, 16 Jun 2008 17:17:11 +0000 |
mas01cr |
Small cleanups.
sampling
|
| Mon, 16 Jun 2008 11:59:43 +0000 |
mas01cr |
Parametrize nsamples (though not on the command-line yet)
sampling
|
| Mon, 16 Jun 2008 11:57:25 +0000 |
mas01cr |
Whoops. Fix EXTREMELY EMBARRASSING bug in distance computation.
sampling
|
| Mon, 16 Jun 2008 11:15:15 +0000 |
mas01cr |
Write y / yinv functions (using the GNU Scientific Library); use them
sampling
|
| Mon, 16 Jun 2008 11:14:21 +0000 |
mas01cr |
Actually use the sequenceLength -l parameter passed on the command-line
sampling
|
| Sat, 14 Jun 2008 17:13:26 +0000 |
mas01cr |
Add initial sampling of database distances. Zillions of FIXME comments
sampling
|
| Sat, 14 Jun 2008 17:12:04 +0000 |
mas01cr |
Create a branch for work-in-progress sampling statistic generation (for
sampling
|
| Tue, 10 Jun 2008 16:22:52 +0000 |
mas01mc |
Fixed trackSequenceQueryNNReporter and trackSequenceQueryRadNNReporter so that order of point reporting is
0.8.9.anteindex
|
| 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.
|
| Mon, 19 May 2008 16:25:12 +0000 |
mas01cr |
Remove, I think, memory leaks from audioDB::batchinsert().
|
| Mon, 19 May 2008 16:05:39 +0000 |
mas01cr |
Fix the segfault in the -Q nsequence reporter. The order is still the
|
| Mon, 19 May 2008 15:59:20 +0000 |
mas01cr |
Add a test file for -Q nsequence. Currently fails hilariously.
|
| Fri, 25 Apr 2008 17:40:19 +0000 |
map01bf |
added cmdline stubs for distance switch and skeleton methods for EMD
|
| Fri, 25 Apr 2008 17:40:19 +0000 |
map01bf |
added cmdline stubs for distance switch and skeleton methods for EMD
adding-emd
|
| Mon, 21 Apr 2008 15:36:50 +0000 |
map01bf |
added a working osx makefile. user must change libsoap reference to reflect thei system.
|
| 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
|
| Sat, 12 Apr 2008 13:28:30 +0000 |
mas01mc |
Fixed side-effect error in trackSwquqnceRadNNReporter point reporter loop.
|
| Fri, 11 Apr 2008 13:01:31 +0000 |
mas01mc |
Fixed bug in TrackSequenceRadNNReporter report() method that overflowed when number of retrieved points < pointnn
|
