feat: import pdfs & images, improve clip from URL, direct to discord,…

… new languages enabled (#1515)

packages/backend/src/services/email/welcome.ts

@@ -15,7 +15,7 @@ export const sendWelcome = async (to: string[], ccTo: string[]) => {
 
     <br /><br />
 
-    Please feel free to contact me if you have questions, concerns or comments at <a href="mailto:[email protected]?subject=RecipeSage%20Support">[email protected]</a>.
+    Please feel free to contact me if you have questions, concerns or comments via <a href="https://discord.gg/yCfzBft">Discord</a>.
 
     <br />
 
@@ -30,7 +30,7 @@ export const sendWelcome = async (to: string[], ccTo: string[]) => {
 
     You can access the RecipeSage user guide for more information about using the application: https://docs.recipesage.com
 
-    Please feel free to contact me if you have questions, concerns or comments at [email protected].
+    Please feel free to contact me if you have questions, concerns or comments via Discord https://discord.gg/yCfzBft.
 
 
     ${signaturePlain}

packages/express/src/routes/import/job/cookmate.ts

@@ -18,7 +18,10 @@ import { JobMeta, prisma } from "@recipesage/prisma";
 import * as Sentry from "@sentry/node";
 import * as xmljs from "xml-js";
 import { cleanLabelTitle, JOB_RESULT_CODES } from "@recipesage/util/shared";
-import { deletePathsSilent } from "@recipesage/util/server/general";
+import {
+  deletePathsSilent,
+  getImportJobResultCode,
+} from "@recipesage/util/server/general";
 import { z } from "zod";
 
 const schema = {
@@ -161,6 +164,10 @@ export const cookmateHandler = defineHandler(
         });
       }
 
+      if (standardizedRecipeImportInput.length === 0) {
+        throw new Error("No recipes");
+      }
+
       await prisma.job.update({
         where: {
           id: job.id,
@@ -213,19 +220,23 @@ export const cookmateHandler = defineHandler(
           e instanceof BadRequestError &&
           e.message === "Bad cookmate file format";
 
+        const isNoRecipesError =
+          e instanceof Error && e.message === "No recipes";
+
         await prisma.job.update({
           where: {
             id: job.id,
           },
           data: {
             status: JobStatus.FAIL,
-            resultCode: isBadFormatError
-              ? JOB_RESULT_CODES.badFile
-              : JOB_RESULT_CODES.unknown,
+            resultCode: getImportJobResultCode({
+              isBadFormat: isBadFormatError,
+              isNoRecipes: isNoRecipesError,
+            }),
           },
         });
 
-        if (!isBadFormatError) {
+        if (!isBadFormatError && !isNoRecipesError) {
           Sentry.captureException(e, {
             extra: {
               jobId: job.id,

packages/express/src/routes/import/job/csv.ts

@@ -18,7 +18,10 @@ import {
   cleanLabelTitle,
   JOB_RESULT_CODES,
 } from "@recipesage/util/shared";
-import { deletePathsSilent } from "@recipesage/util/server/general";
+import {
+  deletePathsSilent,
+  getImportJobResultCode,
+} from "@recipesage/util/server/general";
 import { z } from "zod";
 import { parse } from "csv-parse";
 
@@ -194,6 +197,10 @@ export const csvHandler = defineHandler(
 
       await done;
 
+      if (standardizedRecipeImportInput.length === 0) {
+        throw new Error("No recipes");
+      }
+
       await prisma.job.update({
         where: {
           id: job.id,
@@ -246,19 +253,23 @@ export const csvHandler = defineHandler(
           e instanceof Error &&
           e.message === "end of central directory record signature not found";
 
+        const isNoRecipesError =
+          e instanceof Error && e.message === "No recipes";
+
         await prisma.job.update({
           where: {
             id: job.id,
           },
           data: {
             status: JobStatus.FAIL,
-            resultCode: isBadFormatError
-              ? JOB_RESULT_CODES.badFile
-              : JOB_RESULT_CODES.unknown,
+            resultCode: getImportJobResultCode({
+              isBadFormat: isBadFormatError,
+              isNoRecipes: isNoRecipesError,
+            }),
           },
         });
 
-        if (!isBadFormatError) {
+        if (!isBadFormatError && !isNoRecipesError) {
           Sentry.captureException(e, {
             extra: {
               jobId: job.id,

packages/express/src/routes/import/job/fdxz.ts

@@ -11,7 +11,10 @@ import * as Sentry from "@sentry/node";
 import { userHasCapability } from "@recipesage/util/server/capabilities";
 import { z } from "zod";
 import { spawn } from "child_process";
-import { deletePathsSilent } from "@recipesage/util/server/general";
+import {
+  deletePathsSilent,
+  getImportJobResultCode,
+} from "@recipesage/util/server/general";
 import {
   Capabilities,
   cleanLabelTitle,
@@ -155,9 +158,9 @@ export const fdxzHandler = defineHandler(
           },
           data: {
             status: JobStatus.FAIL,
-            resultCode: isBadFormatError
-              ? JOB_RESULT_CODES.badFile
-              : JOB_RESULT_CODES.unknown,
+            resultCode: getImportJobResultCode({
+              isBadFormat: isBadFormatError,
+            }),
           },
         });
 

packages/express/src/routes/import/job/images.ts

@@ -0,0 +1,200 @@
+import { BadRequestError } from "../../../errors";
+import {
+  AuthenticationEnforcement,
+  defineHandler,
+} from "../../../defineHandler";
+import * as multer from "multer";
+import * as fs from "fs/promises";
+import * as extract from "extract-zip";
+import * as path from "path";
+import { indexRecipes } from "@recipesage/util/server/search";
+import { JobStatus, JobType } from "@prisma/client";
+import {
+  importStandardizedRecipes,
+  StandardizedRecipeImportEntry,
+} from "@recipesage/util/server/db";
+import { ocrImagesToRecipe } from "@recipesage/util/server/ml";
+import { JobMeta, prisma } from "@recipesage/prisma";
+import * as Sentry from "@sentry/node";
+import {
+  deletePathsSilent,
+  getImportJobResultCode,
+} from "@recipesage/util/server/general";
+import { cleanLabelTitle, JOB_RESULT_CODES } from "@recipesage/util/shared";
+import { z } from "zod";
+
+const schema = {
+  query: z.object({
+    labels: z.string().optional(),
+  }),
+};
+
+export const imagesHandler = defineHandler(
+  {
+    schema,
+    authentication: AuthenticationEnforcement.Required,
+    beforeHandlers: [
+      multer({
+        dest: "/tmp/import/",
+      }).single("file"),
+    ],
+  },
+  async (req, res) => {
+    const userLabels =
+      req.query.labels?.split(",").map((label) => cleanLabelTitle(label)) || [];
+
+    const cleanupPaths: string[] = [];
+
+    const file = req.file;
+    if (!file) {
+      throw new BadRequestError(
+        "Request must include multipart file under the 'file' field",
+      );
+    }
+
+    const job = await prisma.job.create({
+      data: {
+        userId: res.locals.session.userId,
+        type: JobType.IMPORT,
+        status: JobStatus.RUN,
+        progress: 1,
+        meta: {
+          importType: "images",
+          importLabels: userLabels,
+        } satisfies JobMeta,
+      },
+    });
+
+    // We complete this work outside of the scope of the request
+    const start = async () => {
+      const zipPath = file.path;
+      cleanupPaths.push(zipPath);
+      const extractPath = zipPath + "-extract";
+      cleanupPaths.push(extractPath);
+
+      await extract(zipPath, { dir: extractPath });
+
+      const fileNames = await fs.readdir(extractPath);
+
+      const standardizedRecipeImportInput: StandardizedRecipeImportEntry[] = [];
+      for (const fileName of fileNames) {
+        const filePath = path.join(extractPath, fileName);
+
+        if (
+          !filePath.endsWith(".jpg") &&
+          !filePath.endsWith(".jpeg") &&
+          !filePath.endsWith(".png")
+        ) {
+          continue;
+        }
+
+        const recipeImageBuffer = await fs.readFile(filePath);
+        const recipeImageBase64 = await fs.readFile(filePath, "base64");
+        const images = [];
+        images.push(recipeImageBase64);
+
+        const recipe = await ocrImagesToRecipe([recipeImageBuffer]);
+        if (!recipe) {
+          continue;
+        }
+
+        standardizedRecipeImportInput.push({
+          ...recipe,
+          images,
+          labels: userLabels,
+        });
+      }
+
+      if (standardizedRecipeImportInput.length === 0) {
+        throw new Error("No recipes");
+      }
+
+      await prisma.job.update({
+        where: {
+          id: job.id,
+        },
+        data: {
+          progress: 50,
+        },
+      });
+
+      const createdRecipeIds = await importStandardizedRecipes(
+        res.locals.session.userId,
+        standardizedRecipeImportInput,
+      );
+
+      const recipesToIndex = await prisma.recipe.findMany({
+        where: {
+          id: {
+            in: createdRecipeIds,
+          },
+          userId: res.locals.session.userId,
+        },
+      });
+
+      await prisma.job.update({
+        where: {
+          id: job.id,
+        },
+        data: {
+          progress: 75,
+        },
+      });
+
+      await indexRecipes(recipesToIndex);
+
+      await prisma.job.update({
+        where: {
+          id: job.id,
+        },
+        data: {
+          status: JobStatus.SUCCESS,
+          resultCode: JOB_RESULT_CODES.success,
+          progress: 100,
+        },
+      });
+    };
+
+    start()
+      .catch(async (e) => {
+        const isBadZipError =
+          e instanceof Error &&
+          e.message === "end of central directory record signature not found";
+
+        const isNoRecipesError =
+          e instanceof Error && e.message === "No recipes";
+
+        await prisma.job.update({
+          where: {
+            id: job.id,
+          },
+          data: {
+            status: JobStatus.FAIL,
+            resultCode: getImportJobResultCode({
+              isBadFormat: isBadZipError,
+              isNoRecipes: isNoRecipesError,
+            }),
+          },
+        });
+
+        if (!isBadZipError && !isNoRecipesError) {
+          Sentry.captureException(e, {
+            extra: {
+              jobId: job.id,
+            },
+          });
+          console.error(e);
+        }
+      })
+      .finally(async () => {
+        await deletePathsSilent(cleanupPaths);
+      });
+
+    return {
+      statusCode: 201,
+      data: {
+        jobId: job.id,
+      },
+    };
+  },
+);

packages/express/src/routes/import/job/index.ts

@@ -9,6 +9,8 @@ import { pepperplateHandler } from "./pepperplate";
 import { recipekeeperHandler } from "./recipekeeper";
 import { urlsHandler } from "./urls";
 import { csvHandler } from "./csv";
+import { pdfsHandler } from "./pdfs";
+import { imagesHandler } from "./images";
 
 const router = express.Router();
 
@@ -22,5 +24,7 @@ router.post("/recipekeeper", ...recipekeeperHandler);
 router.post("/textfiles", ...textfilesHandler);
 router.post("/urls", ...urlsHandler);
 router.post("/csv", ...csvHandler);
+router.post("/pdfs", ...pdfsHandler);
+router.post("/images", ...imagesHandler);
 
 export { router as importJobRouter };