28:147f0472ed01
30:9ffbeab0056e
|
2009-09-08 |
Paul Boddie |
changeset
files
shortlog
graph
|
Introduced a Document class which is instantiated in order to present data to
the IndexWriter in a more efficient way through the add_document method, thus
rendering the commit_document method obsolete.
Reduced the data flushing threshold to a more reasonable size.
Fixed maximum offset definitions in cases where datasets are empty. |
|
|
iixr.py test.py
|
|
27:89404463dda2
|
2009-09-08 |
Paul Boddie |
changeset
files
shortlog
graph
|
Optimised the write_number method slightly.
Changed the flushing mechanism to use documents rather than fields and positions
as the unit of information.
Added a commit_document method to IndexWriter in order to support some kind of
automatic flushing. |
|
|
iixr.py
|
|
26:251e042dada7
27:89404463dda2 29:dec7d2ac1cc9
|
2009-09-06 |
Paul Boddie |
changeset
files
shortlog
graph
|
Added a set_fields method to IndexWriter.
Added a find_terms method to IndexReader. |
|
|
iixr.py
|
|
25:37a06a948a52
|
2009-09-06 |
Paul Boddie |
changeset
files
shortlog
graph
|
Added term retrieval by prefix.
Added a conversion of field values to Unicode in the add_field method. |
|
|
iixr.py test.py
|
|
24:ffec42ec943a
|
2009-09-05 |
Paul Boddie |
changeset
files
shortlog
graph
|
Fixed comment.
Added common merge method. |
|
|
iixr.py
|
|
23:db86d8090718
|
2009-09-04 |
Paul Boddie |
changeset
files
shortlog
graph
|
Fixed old, incorrect reference. |
|
|
iixr.py
|
|
22:4c3c6201310e
|
2009-09-03 |
Paul Boddie |
changeset
files
shortlog
graph
|
Fixed position writing to restart document numbering for each section of the
position file.
Fixed position dictionary iteration to reset the position iterator so that the
document numbering of a newly encountered section is properly interpreted.
Fixed position dictionary iteration across index entries, ensuring that entries
at the start of sections, recorded in index entries themselves, are handled
correctly.
Removed unnecessary reader reset operations where iterators will be created with
reset state anyway.
Added a document frequency method to IndexReader.
Added result caching to the position dictionary iterator in order to preserve
record data for documents which were visited unintentionally. |
|
|
iixr.py test.py
|
|
21:af39faebc734
|
2009-09-03 |
Paul Boddie |
changeset
files
shortlog
graph
|
Added navigation to specific documents in the position dictionary iterator.
Fixed merging to preserve existing merged partitions. |
|
|
iixr.py test.py
|
|
20:e8e80bfe5b5d
|
2009-09-02 |
Paul Boddie |
changeset
files
shortlog
graph
|
Fixed position dictionary writing to write the correct document number in the
index for the specified position file offset.
Fixed index iterator usage in the position dictionary iterator.
Introduced document indexing interval parameters into the API.
Fixed the test program to remove old test files, to close the correct files, and
to use the appropriate names (fixing copying errors). |
|
|
iixr.py test.py
|
|
19:1cf3b82959f3
|
2009-09-02 |
Paul Boddie |
changeset
files
shortlog
graph
|
Attempted to introduce position dictionaries with extra term record details
providing document frequency information.
Attempted to introduce file descriptor duplication in order to support
concurrent iterators. |
|
|
iixr.py test.py
|
|