import { randomUUID } from "node:crypto"; import { mkdir, rm } from "node:fs/promises"; import { tmpdir } from "node:os"; import { join } from "node:path"; import { extractText } from "@snapotter/ai"; import { getBundleForTool, TOOL_BUNDLE_MAP } from "@snapotter/shared"; import type { FastifyInstance, FastifyReply, FastifyRequest } from "fastify"; import { z } from "zod"; import { autoOrient } from "../../lib/auto-orient.js"; import { formatZodErrors } from "../../lib/errors.js"; import { isToolInstalled } from "../../lib/feature-status.js"; import { validateImageBuffer } from "../../lib/file-validation.js"; import { sanitizeFilename } from "../../lib/filename.js"; import { decodeToSharpCompat, needsCliDecode } from "../../lib/format-decoders.js"; import { decodeHeic } from "../../lib/heic-converter.js"; import { updateSingleFileProgress } from "../progress.js"; const settingsSchema = z.object({ quality: z.enum(["fast", "balanced", "best"]).default("balanced"), language: z.enum(["auto", "en", "de", "fr", "es", "zh", "ja", "ko"]).default("auto"), enhance: z.boolean().default(true), // Backward compat: old "engine" param still accepted engine: z.enum(["tesseract", "paddleocr"]).optional(), }); /** * OCR / text extraction route. * Returns JSON with extracted text rather than an image. */ export function registerOcr(app: FastifyInstance) { app.post("/api/v1/tools/ocr", async (request: FastifyRequest, reply: FastifyReply) => { const toolId = "ocr"; if (!isToolInstalled(toolId)) { const bundle = getBundleForTool(toolId); return reply.status(501).send({ error: "Feature not installed", code: "FEATURE_NOT_INSTALLED", feature: TOOL_BUNDLE_MAP[toolId], featureName: bundle?.name ?? toolId, estimatedSize: bundle?.estimatedSize ?? "unknown", }); } let fileBuffer: Buffer | null = null; let filename = "image"; let settingsRaw: string | null = null; let clientJobId: string | null = null; try { const parts = request.parts(); for await (const part of parts) { if (part.type === "file") { const chunks: Buffer[] = []; for await (const chunk of part.file) { chunks.push(chunk); } fileBuffer = Buffer.concat(chunks); filename = sanitizeFilename(part.filename ?? "image"); } else if (part.fieldname === "settings") { settingsRaw = part.value as string; } else if (part.fieldname === "clientJobId") { const raw = part.value as string; if (typeof raw === "string" && raw.length > 0 && raw.length <= 128) { clientJobId = raw; } } } } catch (err) { return reply.status(400).send({ error: "Failed to parse multipart request", details: err instanceof Error ? err.message : String(err), }); } if (!fileBuffer || fileBuffer.length === 0) { return reply.status(400).send({ error: "No image file provided" }); } const validation = await validateImageBuffer(fileBuffer, filename); if (!validation.valid) { return reply.status(400).send({ error: `Invalid image: ${validation.reason}` }); } let scratchDir = ""; try { // Decode HEIC/HEIF input via system decoder if (validation.format === "heif") { fileBuffer = await decodeHeic(fileBuffer); } // Decode CLI-decoded formats (RAW, TGA, PSD, EXR, HDR, etc.) if (needsCliDecode(validation.format)) { fileBuffer = await decodeToSharpCompat(fileBuffer, validation.format); } // Auto-orient to fix EXIF rotation before OCR fileBuffer = await autoOrient(fileBuffer); let settings: z.infer; try { const parsed = settingsRaw ? JSON.parse(settingsRaw) : {}; const result = settingsSchema.safeParse(parsed); if (!result.success) { return reply .status(400) .send({ error: "Invalid settings", details: formatZodErrors(result.error.issues) }); } settings = result.data; } catch { return reply.status(400).send({ error: "Settings must be valid JSON" }); } // Backward compat: map old engine param to quality let quality = settings.quality; if (settings.engine && !settingsRaw?.includes('"quality"')) { quality = settings.engine === "tesseract" ? "fast" : "balanced"; } request.log.info( { toolId: "ocr", imageSize: fileBuffer.length, quality, language: settings.language, }, "Starting OCR", ); const jobId = randomUUID(); scratchDir = join(tmpdir(), "snapotter-scratch", jobId); await mkdir(scratchDir, { recursive: true }); const jobIdForProgress = clientJobId; const onProgress = jobIdForProgress ? (percent: number, stage: string) => { updateSingleFileProgress({ jobId: jobIdForProgress, phase: "processing", stage, percent, }); } : undefined; // Fallback chain: best -> balanced -> fast // PaddleOCR can crash with segfault on some platforms, so we retry // with a lower quality tier at the Node.js level. const fallbackChain: Array<"fast" | "balanced" | "best"> = quality === "best" ? ["best", "balanced", "fast"] : quality === "balanced" ? ["balanced", "fast"] : ["fast"]; let lastError: unknown; for (const tier of fallbackChain) { try { const result = await extractText( fileBuffer, scratchDir, { quality: tier, language: settings.language, enhance: settings.enhance, }, onProgress, ); // If a higher-quality tier returns empty text but didn't crash, // fall back to the next tier rather than returning nothing. if (!result.text && tier !== fallbackChain[fallbackChain.length - 1]) { request.log.warn( { toolId: "ocr", quality: tier, engine: result.engine }, `OCR ${tier} returned empty text, falling back to next tier`, ); if (onProgress) onProgress(15, "Retrying..."); continue; } if (clientJobId) { updateSingleFileProgress({ jobId: clientJobId, phase: "complete", percent: 100, }); } const expectedEngine = tier === "fast" ? "tesseract" : tier === "balanced" ? "paddleocr-v5" : "paddleocr-vl"; if (result.engine && result.engine !== expectedEngine) { request.log.warn( { toolId: "ocr", requested: tier, expected: expectedEngine, actual: result.engine }, `OCR engine fallback: requested ${tier} (${expectedEngine}) but used ${result.engine}`, ); } return reply.send({ jobId, filename, text: result.text, engine: result.engine, }); } catch (err) { lastError = err; const msg = (err instanceof Error ? err.message : String(err)).toLowerCase(); // If the Python process crashed (segfault, dispatcher exit), try next tier if ( msg.includes("exited unexpectedly") || msg.includes("exited with code") || msg.includes("segmentation fault") || msg.includes("process crashed") ) { request.log.warn( { toolId: "ocr", quality: tier, err }, `OCR ${tier} crashed, falling back`, ); if (onProgress) onProgress(15, "Retrying..."); continue; } // Non-crash errors (validation, timeout) should not retry throw err; } } // All tiers failed throw lastError; } catch (err) { request.log.error({ err, toolId: "ocr" }, "OCR failed"); return reply.status(422).send({ error: "OCR failed", details: err instanceof Error ? err.message : "Unknown error", }); } finally { if (scratchDir) await rm(scratchDir, { recursive: true, force: true }).catch(() => {}); } }); }