Repository navigation
Conversation
Six notes from the agents working beside the four lanes, each answering one question for one lane owner: what each phone surface can measure, how an iPhone without LiDAR measures the placement quantities, how the capture knows it has enough before the homeowner leaves, challenger reviews of the lane A and lane C plans (with solver test cases), and a scoring protocol for comparing capture methods against a tape survey. Thresholds without a public source appear as rules.yaml parameters. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
|
Note Review skipped Codemetrics is installed on this repository, but it hasn't been added to your team yet. To enable AI code reviews for this repo, go to your dashboard and add house-scanning to your team's repositories. |
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
|
Warning Review limit reachedNext included review available in 55 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Advanced Run ID: 📒 Files selected for processing (6)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
A pass on a borderline survey is a missed review, not an unsafe pass, so unsafe always means the survey shows the spot breaks a rule. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Code <noreply@anthropic.com>
The ground-truth section described a meters-based file mapped to Site Model v0.1, a contract that was removed before the setup merged. It now defers to the formats in experiments/scoring/README.md (feet, survey ids, sha256 capture and rules ids) and names the Measure Lab importer for the AR row. The recorder is Measure Lab until the lane A app records sessions, and the projection gate names the capture-contract probe. Survey outcomes use the harness's strict rule, and decisions made without their measurement are counted separately. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Code <noreply@anthropic.com>
|
@coderabbitai review |
|
Its gaps (capture ending at the wall ends instead of when each check has evidence, no per-spot evidence tracking, LiDAR in the acceptance case) are handled by the iOS app's coverage map and gap loop and by the decision that LiDAR is optional. No other note or branch links to it. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
|
Closing without merging. These notes were written for the four lanes, which the client and server team split replaced. They link to docs that #14 deleted, and their content has since been built: the scoring protocol into #4, no-LiDAR capture into #7, first-try capture into #10, and the solver review into #11's tests. The branch stays. #4's one citation of the scoring protocol now points to a permanent copy. |
* Sync b1-detection breadth work from the Project store (no result file yet) * Sync b1-detection breadth work from the Project store (no result file yet) * Sync b1-detection breadth work from the Project store (result file B1-detection.md)
1 participant
What changed
docs/research/gets six notes and an index. Each answers one question for the person who owns that part of the plan, and builds ondocs/00throughdocs/05rather than replacing them.t3-capture-surfaces.md(everyone): what each phone surface can measure. AR on iPhone needs native code, which an App Clip can supply without a full install; Android Chrome's WebXR has tracking but its export is untested; iPhone Safari has no WebXR AR. It ends with the path to Android after the hackathon.t3-no-lidar-capture.md(lanes A and B): how an iPhone without LiDAR measures the wall line, positions along the wall, the facing gap and overhead clearance, using ground-contact taps, a vertical plane built from them, and two-view triangulation. Accuracy figures are hypotheses, with a one-hour tape test to settle them.t3-first-try-capture.md(lane A): the evidence each check needs, the finish rule, close-up gates for the meter label and panel, the homeowner's screens, and a first trial with four staged cases. It proposes ending the walk at one fully evidenced spot. The plan anddocs/05keep the walk-to-both-ends rule, so that proposal needs the team's decision.t3-lane-c-review.md(lane C): what the solver specification must settle before it approves or rejects a site, plus test cases inrules.yamlparameter terms.t3-scoring-protocol.md(lane D): how to compare capture methods on one to three houses against an independent tape survey, and what that evidence can and cannot show.The notes propose; the lane owners decide.
Why
Most homeowners have no LiDAR, and the plan's answer for them was to leave the facing gap and headroom UNSURE. These notes test whether ordinary phones can do better, and review the lane plans before code hardens around them. Two experiments build on them: the measurement rig in
experiments/measure-laband the scoring harness in #4.Base's material stays out
Following
AGENTS.md, no note quotes or summarizes what Base gave the team. Numbers come only from Base's public help page, the codes (IRC R328, NEC 110.26 and 706.15) and the team's own docs. Every other threshold is a namedrules.yamlparameter marked for Base to set. Solver test cases that depend on Base-specific policy stay in the team's private notes.Verification
git diff --checkpasses.Checklist
🤖 Generated with Claude Code