Skip to content

Rank datasets guide matches for natural phrases - #16

Merged
galjos merged 2 commits into
mainfrom
fix/datasets-guide-rank-terms
Sep 6, 2026
Merged

Rank datasets guide matches for natural phrases#16
galjos merged 2 commits into
mainfrom
fix/datasets-guide-rank-terms

Conversation

@galjos

@galjos galjos commented Sep 2, 2026

Copy link
Copy Markdown
Owner

Summary

  • datasets guide ranks catalogue hits by how many query terms match instead of requiring every word.
  • Filler words reuse the traffic search stopword list so phrases like roadworks on a state road still find traffic events.
  • datasets search keeps the stricter all-terms filter.
  • Closes datasets guide returns nothing for a naturally-phrased question #14.

Test plan

  • go test ./internal/commands/ -run 'DatasetsGuide|RankDatasets'
  • Spot-check: odh datasets guide "roadworks on a state road" includes mobility.traffic-events
  • Spot-check: odh datasets guide "ev charging availability nearby" still includes mobility.charging

Natural phrases like "roadworks on a state road" returned nothing while
the bare topic still matched, so agents read the guide as "no data".
Score by term hits, drop zero-score rows, and ignore traffic stopwords.
Refs #14.
@galjos
galjos force-pushed the fix/datasets-guide-rank-terms branch 2 times, most recently from 3c55e98 to fcc0857 Compare September 4, 2026 08:14
@galjos
galjos merged commit 117c25c into main Sep 6, 2026
2 checks passed
@galjos
galjos deleted the fix/datasets-guide-rank-terms branch September 6, 2026 13:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

datasets guide returns nothing for a naturally-phrased question

1 participant