| Sat, 15 Oct 2011 17:28:07 +0000 |
mas01mc |
Adding Kullback-Leibler divergence as alternate distance function
|
| Fri, 26 Nov 2010 09:35:28 +0000 |
mas01mc |
A mostly cosmetic fix to query.cpp qpointers allocation, the byte-count vector_size was used a bit too liberally.
|
| 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
|
| Thu, 22 Apr 2010 21:03:47 +0000 |
mas01cr |
begin work on sampling API
|
| Fri, 05 Mar 2010 16:00:58 +0000 |
mas01cr |
rename struct adbfoo to struct adb_foo
|
| Fri, 05 Mar 2010 16:00:53 +0000 |
mas01cr |
Extend results structure to include a key for the query.
|
| Tue, 05 Jan 2010 14:09:43 +0000 |
mas01cr |
Add stdio.h include to audioDB-internals.h
|
| Fri, 28 Aug 2009 17:14:06 +0000 |
mas01cr |
MORE MEMORY SANITY
|
| Tue, 11 Aug 2009 21:42:44 +0000 |
mas01cr |
Make dump.cpp not use mmap()
|
| Tue, 11 Aug 2009 21:42:29 +0000 |
mas01cr |
Remove uses of mmap() from open.cpp
|
| Tue, 11 Aug 2009 21:42:24 +0000 |
mas01cr |
Address portability issues in database creation
|
| Tue, 11 Aug 2009 21:42:18 +0000 |
mas01cr |
Use _locking() to emulate fcntl() locks
|
| Tue, 11 Aug 2009 21:42:13 +0000 |
mas01cr |
Header file rearrangement
|
| Tue, 11 Aug 2009 21:42:07 +0000 |
mas01cr |
Use _commit() to emulate fsync()
|
| Tue, 14 Jul 2009 15:35:36 +0000 |
mas01cr |
API for retrieving a track's data.
|
| Thu, 18 Jun 2009 11:01:31 +0000 |
mas01cr |
Work around the non-universality of fdatasync()
|
| Fri, 20 Feb 2009 19:08:45 +0000 |
mas01mc |
INDEXING: fixed reverse-lookup roundoff error when data gets very large. Computation of offset was lacking an essential cast to off_t.
|
| Fri, 20 Feb 2009 19:03:17 +0000 |
mas01mc |
INDEXING: fixed reverse-lookup roundoff error when data gets very large. Computation of offset was lacking an essential cast to off_t.
multiprobeLSH
|
| 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
|
| 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
|
| Fri, 06 Feb 2009 21:08:35 +0000 |
mas01mc |
Added back single-query point I/O behaviour from pre API-inversion. Thus only those query vectors required are loaded when using query-from-file. Tests 0006 and 0007 fail silently with this change.
multiprobeLSH
|
| Wed, 04 Feb 2009 13:47:02 +0000 |
mas01cr |
Use upper_bound rather than lower_bound in audiodb_index_to_track_id()
|
| Wed, 04 Feb 2009 11:37:07 +0000 |
mas01cr |
Compile fix for previous revision
|
| Wed, 04 Feb 2009 10:45:57 +0000 |
mas01mc |
Removed LSH_N_POINT_BITS coding for LSH index. Now uses binay search via STL lower_bound to locate tracks and positions from global pointID searching over cumulative track lengths. MERGE of branches/multiprobeLSH -r 819:821 onto trunk. This is a non backward-compatible change; WARNING generated on attempt to use INDEXING with older audioDB databases. Only INDEXES are broken, not ADB instances.
|
| Mon, 02 Feb 2009 17:41:02 +0000 |
mas01mc |
DELETED LSH_N_POINT_BITS coding. Tracks and points are now identified with a binary search, std::lower_bound(), on *adb->track_offsets return from LSH retrieval.
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
|
