feat: add percentage

[dewabisma]

No description provided.

[n13]

Reviewer model: GPT-6.1 Sol

REQUEST_CHANGES — the new OG freshness window defeats caching for normal tracker snapshots.

• P2: Preserve cache reuse when the tracker observation is older than five minutes — src/lib/qtc-market.ts:18. The live tracker returned updatedAt=2026-10-14T12:45:06.010Z; at 12:55 UTC that snapshot was about ten minutes old, and its HTTP response still advertised roughly 50 minutes of cache lifetime. The OG handler stores this observation time as X-Observed-At and uses it for both image freshness and browser TTL. With the new five-minute threshold, even a newly rendered image is immediately stale: two consecutive handler requests using that snapshot performed two tracker fetches and two image renders, both returning Cache-Control: public, max-age=0. This repeatedly renders unchanged data and increases request latency and Worker load. Restore the one-hour window, or implement bounded revalidation that reuses the image when the tracker observation has not changed. Cover consecutive requests with an observation older than five minutes in a handler regression test.

Validation at b0cee07c665c91003bf4743c3dbc68744d9abdc4:

• bun test src/lib/qtc-market.test.ts: 13 passed.
• bun run build: passed, including TypeScript.
• Prettier check on all four changed files and git diff --check: passed.
• Handler reproduction with the captured live snapshot, an in-memory cache, and a stubbed image renderer confirmed the finding.
• All current GitHub checks passed; the CI “Build and Test” job currently runs the build with its test step disabled.

[cloudflare-workers-and-pages]

Deploying explorer with    Cloudflare Pages

Latest commit:	67a1a49
Status:	✅  Deploy successful!
Preview URL:	https://1e678333.explorer-kpt.pages.dev
Branch Preview URL:	https://beast-update-with-percentage.explorer-kpt.pages.dev

View logs

[n13]

Reviewer model: GPT-6.1 Sol

APPROVE — no blocking findings at 67a1a496591f3d39b46c9101e4182270e4e79ac8.

Reviewed the complete diff and affected callers. The shared FromAddressCell applies the minting label consistently across the transfer, unified-transaction, and extrinsic-transfer tables; both account-type label callers supply the address.

Rechecked the earlier OG-cache finding against the current upstream behavior. At 14:08:22 UTC, the tracker observation was 247 seconds old and advertised 52 seconds of remaining cache lifetime, consistent with a five-minute window. The earlier one-hour upstream-cache mismatch is no longer present in these live responses. With the captured payload at a one-minute observation age, two consecutive handler requests used one tracker fetch and one image render, both with a positive browser TTL. The exact-head Cloudflare preview also returned HTTP 200, a valid PNG, and max-age=179.

Validation:

• bun test on the market, minting-display, From-cell, address-classification, and account-filter suites: 58 passed, 0 failed.
• bun run build: passed, including TypeScript.
• Prettier check on all 15 changed files and git diff --check: passed.
• Chromium check of the production price display at 375, 640, 680, 768, and 1024 pixels: price and percentage remained visible.
• All current GitHub checks passed. The CI test step remains disabled; the focused tests above were run locally.