mirror of
https://github.com/MBombeck/HealthLog.git
synced 2026-11-08 20:42:43 +00:00
Marc-André Bombeck
df6d20b696
On a host whose CPU cannot run the page renderer (x64 without AVX2, arm64 musl without LSE/RCpc), every AI read of a PDF for a provider that only takes images failed with "Scanning PDFs needs a Claude vision provider", which says nothing about the cause. Most medical PDFs are produced digitally and carry every word in their text layer, which pdf.js reads without rendering. `prepareVisionInput` now answers a text read (`mode: "text"`) when the renderer is unavailable on this host, the provider cannot take the PDF whole and the PDF has a usable text layer. Summary, suggest and the background summary send that text through their existing text mode; "Read with AI" and the index backfill write it as the local text index it is, with no provider call; the automatic index leaves it to its local path. `planVisionReads` does the same for the extraction and the lab scan: text reads by page group, with anchored provenance, and the pages without a text layer reported as `extractionIncomplete` on a document and as `unreadPages` on a lab scan. A supplement label PDF is read from its text layer too. A scan on such a host now answers `documents.inbound.pdfRendererUnavailable` and `labs.ocr.pdfRendererUnavailable` (422): this server's processor cannot run the page renderer, so a scanned PDF can only be read by a provider that takes PDF files directly, or as photos. The `pdfNeedsAnthropic` codes are no longer emitted, since with a working renderer every provider reads a PDF; the recorded index outcome keeps its `pdf-needs-anthropic` value with corrected copy. The extraction path no longer folds a render failure of one document into the capability error: like `prepareVisionInput` since #776 it reports "The PDF pages couldn't be rendered for scanning." (`extractFailed`), and the lab scan answers `labs.ocr.extractFailed`. Hosts that render, and providers that take PDFs whole, are unchanged. Refs #1311
src/
Application source. HealthLog is a Next.js 16 App Router project (output: "standalone", React 19, TypeScript strict). The path alias @/* resolves to this directory. Everything a request touches lives under here; the layout below follows the request flow rather than alphabetical order.
Request flow
client → src/proxy.ts → src/app/<route>/route.ts → src/lib/api-handler.ts → handler body → src/lib/api-response.ts → client
proxy.ts— the HTTP edge (Next.js 16's renamedmiddleware). Public-path allowlist, demo-mode mutation block, server-side onboarding redirect, worker-only refusal,x-request-id+ CSP nonce generation, every security header.app/— App Router pages (RSC by default) and API routes. Every API route wraps its handler inapiHandler. Seeapp/README.md.lib/— the server/runtime toolkit:apiHandler, auth, crypto, rollups, AI, notifications, pg-boss jobs, integrations, rate limiting. Seelib/README.md.components/— React components (shadcn/ui new-york, kebab-case files). Seecomponents/README.md.
Other directories
hooks/— shared React hooks (use-prefix, kebab-case).types/— shared TypeScript declarations.generated/prisma/— the generated Prisma client (do not edit; regenerated byprisma generate). Imported as@/generated/prisma.instrumentation.ts— Next.js startup hook; boots the pg-boss worker registry and boot-time backfill queues.__tests__/— cross-cutting unit tests; most tests live in__tests__/folders next to the code they cover.
Conventions
See ../CLAUDE.md for the full set. The load-bearing ones: RSC by default ("use client" only when a hook/state/browser API needs it); every API route wraps apiHandler and parses its body with Zod safeParse; userId is always narrowed from session or Bearer, never accepted as a body field; TanStack Query keys come only from src/lib/query-keys/; outbound HTTP goes through src/lib/safe-fetch.ts.