From 019b5d2f5ea0a78cb6c44be26b1eba60b2a4e88d Mon Sep 17 00:00:00 2001 From: Mohamed Bassem Date: Sun, 20 Oct 2024 21:06:58 +0000 Subject: feature: Add OCR support for images. Fixes #296 --- packages/shared/config.ts | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'packages') diff --git a/packages/shared/config.ts b/packages/shared/config.ts index 325d9ffa..41430566 100644 --- a/packages/shared/config.ts +++ b/packages/shared/config.ts @@ -25,6 +25,12 @@ const allEnv = z.object({ INFERENCE_TEXT_MODEL: z.string().default("gpt-4o-mini"), INFERENCE_IMAGE_MODEL: z.string().default("gpt-4o-mini"), INFERENCE_CONTEXT_LENGTH: z.coerce.number().default(2048), + OCR_CACHE_DIR: z.string().optional(), + OCR_LANGS: z + .string() + .default("eng") + .transform((val) => val.split(",")), + OCR_CONFIDENCE_THRESHOLD: z.coerce.number().default(50), CRAWLER_HEADLESS_BROWSER: stringBool("true"), BROWSER_WEB_URL: z.string().url().optional(), BROWSER_WEBSOCKET_URL: z.string().url().optional(), @@ -90,6 +96,11 @@ const serverConfigSchema = allEnv.transform((val) => { fullPageScreenshot: val.CRAWLER_FULL_PAGE_SCREENSHOT, fullPageArchive: val.CRAWLER_FULL_PAGE_ARCHIVE, }, + ocr: { + langs: val.OCR_LANGS, + cacheDir: val.OCR_CACHE_DIR, + confidenceThreshold: val.OCR_CONFIDENCE_THRESHOLD, + }, meilisearch: val.MEILI_ADDR ? { address: val.MEILI_ADDR, -- cgit v1.2.3-70-g09d2