Add ChunkMatch.BestLineMatch to return the best-scoring line #884
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jtibshirani
force-pushed
the
jtibs/best-line
branch
from
ce59d31 to
671bcb9
Compare
jtibshirani
commented
Jan 7, 2025
| @@ -149,7 +149,7 @@ func TestMatchSize(t *testing.T) { | |||
| size: 256, | |||
| }, { | |||
| v: ChunkMatch{}, | |||
| size: 112, | |||
| size: 120, | |||
There was a problem hiding this comment.
I ran the fieldalignment tool as this test suggests, and did not see a regression. Here is the output for api.go where ChunkMatch lives ... there is no mention of ChunkMatch or FileMatch:
/Users/jtibshirani/code/zoekt/api.go:232:16: struct with 136 pointer bytes could be 96
/Users/jtibshirani/code/zoekt/api.go:301:24: struct with 32 pointer bytes could be 8
/Users/jtibshirani/code/zoekt/api.go:503:19: struct with 216 pointer bytes could be 24
/Users/jtibshirani/code/zoekt/api.go:561:17: struct of size 224 could be 208
/Users/jtibshirani/code/zoekt/api.go:753:20: struct with 88 pointer bytes could be 56
/Users/jtibshirani/code/zoekt/api.go:833:27: struct with 16 pointer bytes could be 8
/Users/jtibshirani/code/zoekt/api.go:873:15: struct with 32 pointer bytes could be 16
/Users/jtibshirani/code/zoekt/api.go:929:20: struct of size 88 could be 80
stefanhengl
approved these changes
Jan 8, 2025
| @@ -336,7 +335,7 @@ func (p *contentProvider) fillContentChunkMatches(ms []*candidateMatch, numConte | |||
| chunks := chunkCandidates(ms, newlines, numContextLines) | |||
| chunkMatches := make([]ChunkMatch, 0, len(chunks)) | |||
| for _, chunk := range chunks { | |||
| score, debugScore, symbolInfo := p.candidateMatchScore(chunk.candidates, language, debug) | |||
| bestMatch, symbolInfo := p.candidateMatchScore(chunk.candidates, language, debug) | |||
| @@ -364,18 +363,24 @@ func (p *contentProvider) fillContentChunkMatches(ms []*candidateMatch, numConte | |||
| } | |||
| firstLineStart := newlines.lineStart(firstLineNumber) | |||
|
|
|||
| bestLineMatch := 0 | |||
There was a problem hiding this comment.
It is worth adding this to the debugScore?
There was a problem hiding this comment.
Good idea, I added it to the chunk match debug output. It looks like this:
score:5000.00 <- OverlapSymbol:4000.00, kind:Java:classes:1000.00, (line: 3)
Kind of weird but I couldn't come up with something better...
camdencheek
approved these changes
Jan 8, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds a new field
ChunkMatch.BestLineMatchwith the line number of top-scoring line in the chunk. This will let us address a long-standing issue with our new flexible keyword search, where chunk matches can become very large. Since our search results UX only shows the start of a chunk, the most relevant line may not even be visible. With this information on the best line match, we can adjust the search results UX to center the chunk on the most relevant line.Relates to SPLF-188