nimblemachine / poly-ir-toolkit

Automatically exported from code.google.com/p/poly-ir-toolkit

Repository from Github https://github.comnimblemachine/poly-ir-toolkit

nimblemachine/poly-ir-toolkit Issues

Store the exact layering parameters in the index meta file for reference.
Updated 10 years ago
Clean up the way we pass the filenames of the various on disk structures that make up an index.
Closed 10 years ago1
src/ir_toolkit.cc:617:34: error: 'fstat' was not declared in this scope
Updated 10 years ago
Benchmark inlining of NextGEQ.
Updated 10 years ago
WAND and MaxScore don't need a layered index.
Updated 10 years ago
Indices should store docID offset in meta file. Index reader automatically applies offset to the first decoded docID of each list.
Updated 10 years ago1
Document map needs to store document numbers for WARC files (used by ClueWeb).
Updated 10 years ago
Single word and OR style query performance.
Updated 10 years ago3
Speed up decoding of positions.
Updated 10 years ago1
Block header decoding improvement.
Updated 10 years ago
Lay out members to adhere to cache line sizes during list traversal.
Updated 10 years ago3
Integrate external index with the rest of your index slices naming convention.
Closed 10 years ago6
Implement query processing across indices.
Updated 10 years ago
Load layering info into dynamic array for lexicon.
Closed 10 years ago1
Remove NextGEQOld() function.
Closed 10 years ago1
During indexing, all docIDs will be assigned in a monotonically increasing order. No document reordering during indexing.
Closed 10 years ago1
Integrate TREC/WARC document identifiers into document map.
Closed 10 years ago2
Don't limit the number of frequencies.
Closed 10 years ago1
Lower bound on chunk size incorrect when not indexing positions.
Closed 10 years ago2
DocID remapping must also work on the document map.
Closed 10 years ago2