From 184b9d9bd869e804304d841489dc0d7a97019901 Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Wed, 15 Jan 2025 13:49:29 -0300
Subject: [PATCH 1/6] ocrspace init

---
 .../actions/get-ocr-result/get-ocr-result.mjs |  26 ++
 .../actions/process-image/process-image.mjs   |  53 +++
 .../actions/process-pdf/process-pdf.mjs       |  54 +++
 components/ocrspace/ocrspace.app.mjs          | 356 +++++++++++++++++-
 .../new-file-uploaded/new-file-uploaded.mjs   | 111 ++++++
 .../new-ocr-job-completed-instant.mjs         |  67 ++++
 6 files changed, 665 insertions(+), 2 deletions(-)
 create mode 100644 components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
 create mode 100644 components/ocrspace/actions/process-image/process-image.mjs
 create mode 100644 components/ocrspace/actions/process-pdf/process-pdf.mjs
 create mode 100644 components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
 create mode 100644 components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs

diff --git a/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs b/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
new file mode 100644
index 0000000000000..eab70083fffb3
--- /dev/null
+++ b/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
@@ -0,0 +1,26 @@
+import ocrspace from "../../ocrspace.app.mjs";
+import { axios } from "@pipedream/platform";
+
+export default {
+  key: "ocrspace-get-ocr-result",
+  name: "Get OCR Result",
+  description: "Retrieves the processed OCR result for a specific job ID. [See the documentation](https://ocr.space/ocrapi)",
+  version: "0.0.{{ts}}",
+  type: "action",
+  props: {
+    ocrspace,
+    jobId: {
+      propDefinition: [
+        ocrspace,
+        "jobId",
+      ],
+    },
+  },
+  async run({ $ }) {
+    const result = await this.ocrspace.retrieveOcrResult({
+      jobId: this.jobId,
+    });
+    $.export("$summary", `Retrieved OCR result for job ID ${this.jobId}`);
+    return result;
+  },
+};
diff --git a/components/ocrspace/actions/process-image/process-image.mjs b/components/ocrspace/actions/process-image/process-image.mjs
new file mode 100644
index 0000000000000..e0499c680626d
--- /dev/null
+++ b/components/ocrspace/actions/process-image/process-image.mjs
@@ -0,0 +1,53 @@
+import ocrspace from "../../ocrspace.app.mjs";
+import { axios } from "@pipedream/platform";
+
+export default {
+  key: "ocrspace-process-image",
+  name: "Process Image",
+  description: "Submits an image file for OCR processing using OCR.space. [See the documentation](https://ocr.space/ocrapi)",
+  version: "0.0.{{ts}}",
+  type: "action",
+  props: {
+    ocrspace: {
+      type: "app",
+      app: "ocrspace",
+    },
+    imageUrl: {
+      propDefinition: [
+        "ocrspace",
+        "imageUrl",
+      ],
+    },
+    imageFile: {
+      propDefinition: [
+        "ocrspace",
+        "imageFile",
+      ],
+    },
+    imageLanguage: {
+      propDefinition: [
+        "ocrspace",
+        "imageLanguage",
+      ],
+      optional: true,
+    },
+  },
+  async run({ $ }) {
+    if (!this.imageUrl && !this.imageFile) {
+      throw new Error("Either Image File URL or Image File Upload must be provided.");
+    }
+
+    const response = await this.ocrspace.submitImage({
+      imageUrl: this.imageUrl,
+      imageFile: this.imageFile,
+      imageLanguage: this.imageLanguage,
+    });
+
+    const summary = response.JobId
+      ? `Image submitted for OCR processing. Job ID: ${response.JobId}`
+      : "Image submitted for OCR processing.";
+
+    $.export("$summary", summary);
+    return response;
+  },
+};
diff --git a/components/ocrspace/actions/process-pdf/process-pdf.mjs b/components/ocrspace/actions/process-pdf/process-pdf.mjs
new file mode 100644
index 0000000000000..50ec5141305f6
--- /dev/null
+++ b/components/ocrspace/actions/process-pdf/process-pdf.mjs
@@ -0,0 +1,54 @@
+import ocrspace from "../../ocrspace.app.mjs";
+import { axios } from "@pipedream/platform";
+
+export default {
+  key: "ocrspace-process-pdf",
+  name: "Process PDF for OCR",
+  description: "Submit a PDF for OCR processing. [See the documentation]()",
+  version: "0.0.{{ts}}",
+  type: "action",
+  props: {
+    ocrspace: {
+      type: "app",
+      app: "ocrspace",
+    },
+    pdfUrl: {
+      propDefinition: [
+        ocrspace,
+        "pdfUrl",
+      ],
+    },
+    pdfFile: {
+      propDefinition: [
+        ocrspace,
+        "pdfFile",
+      ],
+    },
+    pdfLanguage: {
+      propDefinition: [
+        ocrspace,
+        "pdfLanguage",
+      ],
+      optional: true,
+    },
+    pdfPages: {
+      propDefinition: [
+        ocrspace,
+        "pdfPages",
+      ],
+      optional: true,
+    },
+  },
+  async run({ $ }) {
+    const response = await this.ocrspace.submitPdf({
+      pdfUrl: this.pdfUrl,
+      pdfFile: this.pdfFile,
+      pdfLanguage: this.pdfLanguage,
+      pdfPages: this.pdfPages,
+    });
+
+    const jobId = response.JobId || response.jobId || "N/A";
+    $.export("$summary", `Submitted PDF for OCR processing. Job ID: ${jobId}`);
+    return response;
+  },
+};
diff --git a/components/ocrspace/ocrspace.app.mjs b/components/ocrspace/ocrspace.app.mjs
index 915aa5782b482..ae49aa9d24882 100644
--- a/components/ocrspace/ocrspace.app.mjs
+++ b/components/ocrspace/ocrspace.app.mjs
@@ -1,9 +1,361 @@
+import { axios } from "@pipedream/platform";
+import FormData from "form-data";
+
 export default {
   type: "app",
   app: "ocrspace",
-  propDefinitions: {},
+  version: "0.0.{{ts}}",
+  propDefinitions: {
+    webhookUrl: {
+      type: "string",
+      label: "Webhook URL",
+      description: "The URL to receive notifications of completed OCR processing.",
+    },
+    monitoredFolder: {
+      type: "string",
+      label: "Monitored Folder",
+      description: "Optional folder to monitor for new file uploads.",
+      optional: true,
+    },
+    processingQueue: {
+      type: "string",
+      label: "Processing Queue",
+      description: "Optional processing queue to monitor for new file uploads.",
+      optional: true,
+    },
+    imageUrl: {
+      type: "string",
+      label: "Image File URL",
+      description: "The URL of the image file to submit for OCR processing.",
+      optional: true,
+    },
+    imageFile: {
+      type: "file",
+      label: "Image File Upload",
+      description: "The image file to submit for OCR processing.",
+      optional: true,
+    },
+    imageLanguage: {
+      type: "string",
+      label: "Image Language",
+      description: "Optional language setting for image OCR processing.",
+      optional: true,
+      options: [
+        {
+          label: "Arabic",
+          value: "ara",
+        },
+        {
+          label: "Bulgarian",
+          value: "bul",
+        },
+        {
+          label: "Chinese (Simplified)",
+          value: "chs",
+        },
+        {
+          label: "Chinese (Traditional)",
+          value: "cht",
+        },
+        {
+          label: "Croatian",
+          value: "hrv",
+        },
+        {
+          label: "Czech",
+          value: "cze",
+        },
+        {
+          label: "Danish",
+          value: "dan",
+        },
+        {
+          label: "Dutch",
+          value: "dut",
+        },
+        {
+          label: "English",
+          value: "eng",
+        },
+        {
+          label: "Finnish",
+          value: "fin",
+        },
+        {
+          label: "French",
+          value: "fre",
+        },
+        {
+          label: "German",
+          value: "ger",
+        },
+        {
+          label: "Greek",
+          value: "gre",
+        },
+        {
+          label: "Hungarian",
+          value: "hun",
+        },
+        {
+          label: "Korean",
+          value: "kor",
+        },
+        {
+          label: "Italian",
+          value: "ita",
+        },
+        {
+          label: "Japanese",
+          value: "jpn",
+        },
+        {
+          label: "Polish",
+          value: "pol",
+        },
+        {
+          label: "Portuguese",
+          value: "por",
+        },
+        {
+          label: "Russian",
+          value: "rus",
+        },
+        {
+          label: "Slovenian",
+          value: "slv",
+        },
+        {
+          label: "Spanish",
+          value: "spa",
+        },
+        {
+          label: "Swedish",
+          value: "swe",
+        },
+        {
+          label: "Turkish",
+          value: "tur",
+        },
+      ],
+    },
+    pdfUrl: {
+      type: "string",
+      label: "PDF File URL",
+      description: "The URL of the PDF file to submit for OCR processing.",
+      optional: true,
+    },
+    pdfFile: {
+      type: "file",
+      label: "PDF File Upload",
+      description: "The PDF file to submit for OCR processing.",
+      optional: true,
+    },
+    pdfLanguage: {
+      type: "string",
+      label: "PDF Language",
+      description: "Optional language setting for PDF OCR processing.",
+      optional: true,
+      options: [
+        {
+          label: "Arabic",
+          value: "ara",
+        },
+        {
+          label: "Bulgarian",
+          value: "bul",
+        },
+        {
+          label: "Chinese (Simplified)",
+          value: "chs",
+        },
+        {
+          label: "Chinese (Traditional)",
+          value: "cht",
+        },
+        {
+          label: "Croatian",
+          value: "hrv",
+        },
+        {
+          label: "Czech",
+          value: "cze",
+        },
+        {
+          label: "Danish",
+          value: "dan",
+        },
+        {
+          label: "Dutch",
+          value: "dut",
+        },
+        {
+          label: "English",
+          value: "eng",
+        },
+        {
+          label: "Finnish",
+          value: "fin",
+        },
+        {
+          label: "French",
+          value: "fre",
+        },
+        {
+          label: "German",
+          value: "ger",
+        },
+        {
+          label: "Greek",
+          value: "gre",
+        },
+        {
+          label: "Hungarian",
+          value: "hun",
+        },
+        {
+          label: "Korean",
+          value: "kor",
+        },
+        {
+          label: "Italian",
+          value: "ita",
+        },
+        {
+          label: "Japanese",
+          value: "jpn",
+        },
+        {
+          label: "Polish",
+          value: "pol",
+        },
+        {
+          label: "Portuguese",
+          value: "por",
+        },
+        {
+          label: "Russian",
+          value: "rus",
+        },
+        {
+          label: "Slovenian",
+          value: "slv",
+        },
+        {
+          label: "Spanish",
+          value: "spa",
+        },
+        {
+          label: "Swedish",
+          value: "swe",
+        },
+        {
+          label: "Turkish",
+          value: "tur",
+        },
+      ],
+    },
+    pdfPages: {
+      type: "string",
+      label: "PDF Pages",
+      description: "Optional specific pages to process in the PDF file.",
+      optional: true,
+    },
+    jobId: {
+      type: "string",
+      label: "Job ID",
+      description: "The Job ID to retrieve the processed OCR result.",
+    },
+  },
   methods: {
-    // this.$auth contains connected account data
+    _baseUrl() {
+      return "https://api.ocr.space";
+    },
+    async _makeRequest(opts = {}) {
+      const {
+        $, method = "POST", path = "/parse/image", headers, ...otherOpts
+      } = opts;
+      return axios($, {
+        ...otherOpts,
+        method,
+        url: this._baseUrl() + path,
+        headers: {
+          ...headers,
+          apikey: this.$auth.apikey,
+        },
+      });
+    },
+    async submitImage(opts = {}) {
+      const {
+        imageUrl, imageFile, imageLanguage,
+      } = opts;
+      const formData = new FormData();
+      if (imageUrl) {
+        formData.append("url", imageUrl);
+      }
+      if (imageFile) {
+        formData.append("file", imageFile);
+      }
+      formData.append("isOverlayRequired", false);
+      if (imageLanguage) {
+        formData.append("language", imageLanguage);
+      }
+      return this._makeRequest({
+        path: "/parse/image",
+        method: "POST",
+        data: formData,
+        headers: formData.getHeaders(),
+      });
+    },
+    async submitPdf(opts = {}) {
+      const {
+        pdfUrl, pdfFile, pdfLanguage, pdfPages,
+      } = opts;
+      const formData = new FormData();
+      if (pdfUrl) {
+        formData.append("url", pdfUrl);
+      }
+      if (pdfFile) {
+        formData.append("file", pdfFile);
+      }
+      formData.append("isOverlayRequired", false);
+      if (pdfLanguage) {
+        formData.append("language", pdfLanguage);
+      }
+      if (pdfPages) {
+        formData.append("pages", pdfPages);
+      }
+      return this._makeRequest({
+        path: "/parse/image",
+        method: "POST",
+        data: formData,
+        headers: formData.getHeaders(),
+      });
+    },
+    async retrieveOcrResult(opts = {}) {
+      const { jobId } = opts;
+      const path = `/parse/image/${jobId}`;
+      return this._makeRequest({
+        path,
+        method: "GET",
+      });
+    },
+    async emitOcrJobCompleted(data) {
+      const { webhookUrl } = this;
+      await axios(this, {
+        method: "POST",
+        url: webhookUrl,
+        data,
+      });
+    },
+    async emitNewFileUploaded(data) {
+      const { webhookUrl } = this;
+      await axios(this, {
+        method: "POST",
+        url: webhookUrl,
+        data,
+      });
+    },
     authKeys() {
       console.log(Object.keys(this.$auth));
     },
diff --git a/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs b/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
new file mode 100644
index 0000000000000..72d297ec5ee32
--- /dev/null
+++ b/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
@@ -0,0 +1,111 @@
+import {
+  axios, DEFAULT_POLLING_SOURCE_TIMER_INTERVAL,
+} from "@pipedream/platform";
+import ocrspace from "../../ocrspace.app.mjs";
+
+export default {
+  key: "ocrspace-new-file-uploaded",
+  name: "New OCR File Uploaded",
+  description: "Emit a new event when a new file is uploaded for OCR processing. [See the documentation]()",
+  version: "0.0.{{ts}}",
+  type: "source",
+  dedupe: "unique",
+  props: {
+    ocrspace,
+    db: "$.service.db",
+    timer: {
+      type: "$.interface.timer",
+      default: {
+        intervalSeconds: DEFAULT_POLLING_SOURCE_TIMER_INTERVAL,
+      },
+    },
+    monitoredFolder: {
+      type: "string",
+      label: "Monitored Folder",
+      description: "Optional folder to monitor for new file uploads.",
+      optional: true,
+    },
+    processingQueue: {
+      type: "string",
+      label: "Processing Queue",
+      description: "Optional processing queue to monitor for new file uploads.",
+      optional: true,
+    },
+  },
+  methods: {
+    async fetchUploadedFiles() {
+      const params = {};
+      if (this.monitoredFolder) {
+        params.folder = this.monitoredFolder;
+      }
+      if (this.processingQueue) {
+        params.queue = this.processingQueue;
+      }
+      const uploadedFiles = await this.ocrspace._makeRequest({
+        path: "/list/uploads",
+        method: "GET",
+        params,
+      });
+      return uploadedFiles;
+    },
+  },
+  hooks: {
+    async deploy() {
+      const files = await this.fetchUploadedFiles();
+      // Sort files by timestamp descending
+      files.sort((a, b) => new Date(b.timestamp) - new Date(a.timestamp));
+      const recentFiles = files.slice(0, 50).reverse();
+      for (const file of recentFiles) {
+        const id = file.id || new Date(file.timestamp).getTime();
+        const summary = `New file uploaded: ${file.name}`;
+        const ts = file.timestamp
+          ? new Date(file.timestamp).getTime()
+          : Date.now();
+        this.$emit(file, {
+          id,
+          summary,
+          ts,
+        });
+      }
+      if (files.length > 0) {
+        const latestFile = files[0];
+        const lastFileId = latestFile.id || new Date(latestFile.timestamp).getTime();
+        await this.db.set("lastFileId", lastFileId);
+      }
+    },
+    async activate() {
+      // No action needed on activate
+    },
+    async deactivate() {
+      // No action needed on deactivate
+    },
+  },
+  async run() {
+    const files = await this.fetchUploadedFiles();
+    const lastFileId = (await this.db.get("lastFileId")) || 0;
+    const newFiles = files.filter(
+      (file) =>
+        (file.id && file.id > lastFileId) ||
+        (file.timestamp && new Date(file.timestamp).getTime() > lastFileId),
+    );
+    // Sort newFiles by timestamp ascending
+    newFiles.sort((a, b) => new Date(a.timestamp) - new Date(b.timestamp));
+    for (const file of newFiles) {
+      const id = file.id || new Date(file.timestamp).getTime();
+      const summary = `New file uploaded: ${file.name}`;
+      const ts = file.timestamp
+        ? new Date(file.timestamp).getTime()
+        : Date.now();
+      this.$emit(file, {
+        id,
+        summary,
+        ts,
+      });
+    }
+    if (files.length > 0) {
+      const latestFile = files[0];
+      const lastFileId = latestFile.id || new Date(latestFile.timestamp).getTime();
+      await this.db.set("lastFileId", lastFileId);
+    }
+  },
+};
diff --git a/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs b/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs
new file mode 100644
index 0000000000000..cfed5441c340f
--- /dev/null
+++ b/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs
@@ -0,0 +1,67 @@
+import ocrspace from "../../ocrspace.app.mjs";
+import { axios } from "@pipedream/platform";
+
+export default {
+  key: "ocrspace-new-ocr-job-completed-instant",
+  name: "New OCR Job Completed",
+  description: "Emit new event when an OCR job is completed. [See the documentation](https://ocr.space/ocrapi)",
+  version: "0.0.{{ts}}",
+  type: "source",
+  dedupe: "unique",
+  props: {
+    ocrspace: {
+      type: "app",
+      app: "ocrspace",
+    },
+    http: {
+      type: "$.interface.http",
+      customResponse: true,
+    },
+    db: "$.service.db",
+  },
+  hooks: {
+    async activate() {
+      console.log(`Please configure your OCRSpace webhook URL to: ${this.http.endpoint}`);
+    },
+    async deactivate() {
+      console.log("Webhook deactivated. Please remove the webhook URL from your OCRSpace settings.");
+    },
+    async deploy() {
+      const lastJobId = await this.db.get("lastJobId");
+      if (lastJobId) {
+        try {
+          const job = await this.ocrspace.retrieveOcrResult({
+            jobId: lastJobId,
+          });
+          this.$emit(job, {
+            id: job.jobId || lastJobId,
+            summary: `OCR job completed: ${job.jobId || lastJobId}`,
+            ts: Date.now(),
+          });
+        } catch (error) {
+          console.error(`Failed to retrieve OCR job with ID ${lastJobId}:`, error);
+        }
+      }
+    },
+  },
+  async run(event) {
+    const job = event;
+    const jobId = job.jobId || job.id || null;
+    const summary = jobId
+      ? `OCR job completed: ${jobId}`
+      : "OCR job completed";
+    const ts = job.completedAt
+      ? Date.parse(job.completedAt)
+      : Date.now();
+
+    this.$emit(job, {
+      id: jobId || ts.toString(),
+      summary,
+      ts,
+    });
+
+    if (jobId) {
+      await this.db.set("lastJobId", jobId);
+    }
+  },
+};

From c92f9e558862cc35bd0b4a800230e6819fbeace7 Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Thu, 16 Jan 2025 12:17:31 -0300
Subject: [PATCH 2/6] [Components] ocrspace #15148 Actions  - Process Image  -
 Process PDF

---
 .../ocrspace/actions/common/process-base.mjs  |  70 ++++
 .../actions/get-ocr-result/get-ocr-result.mjs |  26 --
 .../actions/process-image/process-image.mjs   |  51 +--
 .../actions/process-pdf/process-pdf.mjs       |  56 +--
 components/ocrspace/common/constants.mjs      | 117 ++++++
 components/ocrspace/common/utils.mjs          |  31 ++
 components/ocrspace/ocrspace.app.mjs          | 371 +++---------------
 components/ocrspace/package.json              |  18 +
 .../new-file-uploaded/new-file-uploaded.mjs   | 111 ------
 .../new-ocr-job-completed-instant.mjs         |  67 ----
 10 files changed, 313 insertions(+), 605 deletions(-)
 create mode 100644 components/ocrspace/actions/common/process-base.mjs
 delete mode 100644 components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
 create mode 100644 components/ocrspace/common/constants.mjs
 create mode 100644 components/ocrspace/common/utils.mjs
 create mode 100644 components/ocrspace/package.json
 delete mode 100644 components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
 delete mode 100644 components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs

diff --git a/components/ocrspace/actions/common/process-base.mjs b/components/ocrspace/actions/common/process-base.mjs
new file mode 100644
index 0000000000000..adc8f0cfffb4c
--- /dev/null
+++ b/components/ocrspace/actions/common/process-base.mjs
@@ -0,0 +1,70 @@
+import FormData from "form-data";
+import { getUrlOrFile } from "../../common/utils.mjs";
+import ocrspace from "../../ocrspace.app.mjs";
+
+export default {
+  props: {
+    ocrspace,
+    language: {
+      propDefinition: [
+        ocrspace,
+        "language",
+      ],
+    },
+    isOverlayRequired: {
+      propDefinition: [
+        ocrspace,
+        "isOverlayRequired",
+      ],
+    },
+    detectOrientation: {
+      propDefinition: [
+        ocrspace,
+        "detectOrientation",
+      ],
+    },
+    scale: {
+      propDefinition: [
+        ocrspace,
+        "scale",
+      ],
+    },
+    isTable: {
+      propDefinition: [
+        ocrspace,
+        "isTable",
+      ],
+    },
+    ocrEngine: {
+      propDefinition: [
+        ocrspace,
+        "ocrEngine",
+      ],
+    },
+  },
+  async run({ $ }) {
+    const data = new FormData();
+    const {
+      url, file,
+    } = getUrlOrFile(this.file);
+
+    if (url) data.append("url", url);
+    if (file) data.append("file", file);
+    if (this.imageLanguage) data.append("language", this.imageLanguage);
+    if (this.isOverlayRequired) data.append("isOverlayRequired", `${this.isOverlayRequired}`);
+    if (this.filetype) data.append("filetype", this.filetype);
+    if (this.detectOrientation) data.append("detectOrientation", `${this.detectOrientation}`);
+    if (this.scale) data.append("scale", `${this.scale}`);
+    if (this.isTable) data.append("isTable", `${this.isTable}`);
+    if (this.ocrEngine) data.append("OCREngine", this.ocrEngine);
+
+    const response = await this.ocrspace.processImage({
+      $,
+      data,
+      headers: data.getHeaders(),
+    });
+
+    $.export("$summary", this.getSummary());
+    return response;
+  },
+};
diff --git a/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs b/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
deleted file mode 100644
index eab70083fffb3..0000000000000
--- a/components/ocrspace/actions/get-ocr-result/get-ocr-result.mjs
+++ /dev/null
@@ -1,26 +0,0 @@
-import ocrspace from "../../ocrspace.app.mjs";
-import { axios } from "@pipedream/platform";
-
-export default {
-  key: "ocrspace-get-ocr-result",
-  name: "Get OCR Result",
-  description: "Retrieves the processed OCR result for a specific job ID. [See the documentation](https://ocr.space/ocrapi)",
-  version: "0.0.{{ts}}",
-  type: "action",
-  props: {
-    ocrspace,
-    jobId: {
-      propDefinition: [
-        ocrspace,
-        "jobId",
-      ],
-    },
-  },
-  async run({ $ }) {
-    const result = await this.ocrspace.retrieveOcrResult({
-      jobId: this.jobId,
-    });
-    $.export("$summary", `Retrieved OCR result for job ID ${this.jobId}`);
-    return result;
-  },
-};
diff --git a/components/ocrspace/actions/process-image/process-image.mjs b/components/ocrspace/actions/process-image/process-image.mjs
index e0499c680626d..dace35e1e2ca1 100644
--- a/components/ocrspace/actions/process-image/process-image.mjs
+++ b/components/ocrspace/actions/process-image/process-image.mjs
@@ -1,53 +1,30 @@
-import ocrspace from "../../ocrspace.app.mjs";
-import { axios } from "@pipedream/platform";
+import common from "../common/process-base.mjs";
 
 export default {
+  ...common,
   key: "ocrspace-process-image",
   name: "Process Image",
   description: "Submits an image file for OCR processing using OCR.space. [See the documentation](https://ocr.space/ocrapi)",
-  version: "0.0.{{ts}}",
+  version: "0.0.1",
   type: "action",
   props: {
-    ocrspace: {
-      type: "app",
-      app: "ocrspace",
-    },
-    imageUrl: {
-      propDefinition: [
-        "ocrspace",
-        "imageUrl",
-      ],
-    },
-    imageFile: {
+    ...common.props,
+    file: {
       propDefinition: [
-        "ocrspace",
-        "imageFile",
+        common.props.ocrspace,
+        "file",
       ],
     },
-    imageLanguage: {
+    filetype: {
       propDefinition: [
-        "ocrspace",
-        "imageLanguage",
+        common.props.ocrspace,
+        "filetype",
       ],
-      optional: true,
     },
   },
-  async run({ $ }) {
-    if (!this.imageUrl && !this.imageFile) {
-      throw new Error("Either Image File URL or Image File Upload must be provided.");
-    }
-
-    const response = await this.ocrspace.submitImage({
-      imageUrl: this.imageUrl,
-      imageFile: this.imageFile,
-      imageLanguage: this.imageLanguage,
-    });
-
-    const summary = response.JobId
-      ? `Image submitted for OCR processing. Job ID: ${response.JobId}`
-      : "Image submitted for OCR processing.";
-
-    $.export("$summary", summary);
-    return response;
+  methods: {
+    getSummary() {
+      return "Image submitted for OCR processing.";
+    },
   },
 };
diff --git a/components/ocrspace/actions/process-pdf/process-pdf.mjs b/components/ocrspace/actions/process-pdf/process-pdf.mjs
index 50ec5141305f6..4669979849a62 100644
--- a/components/ocrspace/actions/process-pdf/process-pdf.mjs
+++ b/components/ocrspace/actions/process-pdf/process-pdf.mjs
@@ -1,54 +1,26 @@
-import ocrspace from "../../ocrspace.app.mjs";
-import { axios } from "@pipedream/platform";
+import common from "../common/process-base.mjs";
 
 export default {
+  ...common,
   key: "ocrspace-process-pdf",
   name: "Process PDF for OCR",
-  description: "Submit a PDF for OCR processing. [See the documentation]()",
-  version: "0.0.{{ts}}",
+  description: "Submit a PDF for OCR processing. [See the documentation](https://ocr.space/ocrapi)",
+  version: "0.0.1",
   type: "action",
   props: {
-    ocrspace: {
-      type: "app",
-      app: "ocrspace",
-    },
-    pdfUrl: {
-      propDefinition: [
-        ocrspace,
-        "pdfUrl",
-      ],
-    },
-    pdfFile: {
-      propDefinition: [
-        ocrspace,
-        "pdfFile",
-      ],
-    },
-    pdfLanguage: {
-      propDefinition: [
-        ocrspace,
-        "pdfLanguage",
-      ],
-      optional: true,
-    },
-    pdfPages: {
+    ...common.props,
+    file: {
       propDefinition: [
-        ocrspace,
-        "pdfPages",
+        common.props.ocrspace,
+        "file",
       ],
-      optional: true,
+      label: "PDF File",
+      description: "The URL of the PDF file or the path to the file saved to the `/tmp` directory  (e.g. `/tmp/example.jpg`)  to process. [See the documentation](https://pipedream.com/docs/workflows/steps/code/nodejs/working-with-files/#the-tmp-directory).",
     },
   },
-  async run({ $ }) {
-    const response = await this.ocrspace.submitPdf({
-      pdfUrl: this.pdfUrl,
-      pdfFile: this.pdfFile,
-      pdfLanguage: this.pdfLanguage,
-      pdfPages: this.pdfPages,
-    });
-
-    const jobId = response.JobId || response.jobId || "N/A";
-    $.export("$summary", `Submitted PDF for OCR processing. Job ID: ${jobId}`);
-    return response;
+  methods: {
+    getSummary() {
+      return "Submitted PDF for OCR processing.";
+    },
   },
 };
diff --git a/components/ocrspace/common/constants.mjs b/components/ocrspace/common/constants.mjs
new file mode 100644
index 0000000000000..ea51ce5ad0337
--- /dev/null
+++ b/components/ocrspace/common/constants.mjs
@@ -0,0 +1,117 @@
+export const LANGUAGE_OPTIONS = [
+  {
+    label: "Arabic",
+    value: "ara",
+  },
+  {
+    label: "Bulgarian",
+    value: "bul",
+  },
+  {
+    label: "Chinese (Simplified)",
+    value: "chs",
+  },
+  {
+    label: "Chinese (Traditional)",
+    value: "cht",
+  },
+  {
+    label: "Croatian",
+    value: "hrv",
+  },
+  {
+    label: "Czech",
+    value: "cze",
+  },
+  {
+    label: "Danish",
+    value: "dan",
+  },
+  {
+    label: "Dutch",
+    value: "dut",
+  },
+  {
+    label: "English",
+    value: "eng",
+  },
+  {
+    label: "Finnish",
+    value: "fin",
+  },
+  {
+    label: "French",
+    value: "fre",
+  },
+  {
+    label: "German",
+    value: "ger",
+  },
+  {
+    label: "Greek",
+    value: "gre",
+  },
+  {
+    label: "Hungarian",
+    value: "hun",
+  },
+  {
+    label: "Korean",
+    value: "kor",
+  },
+  {
+    label: "Italian",
+    value: "ita",
+  },
+  {
+    label: "Japanese",
+    value: "jpn",
+  },
+  {
+    label: "Polish",
+    value: "pol",
+  },
+  {
+    label: "Portuguese",
+    value: "por",
+  },
+  {
+    label: "Russian",
+    value: "rus",
+  },
+  {
+    label: "Slovenian",
+    value: "slv",
+  },
+  {
+    label: "Spanish",
+    value: "spa",
+  },
+  {
+    label: "Swedish",
+    value: "swe",
+  },
+  {
+    label: "Turkish",
+    value: "tur",
+  },
+];
+
+export const IMAGE_FILETYPE_OPTIONS = [
+  "GIF",
+  "PNG",
+  "JPG",
+  "TIF",
+  "BMP",
+];
+
+export const OCR_ENGINE_OPTIONS = [
+  {
+    label: "OCR Engine 1",
+    value: "1",
+  },
+  {
+    label: "OCR Engine 2",
+    value: "2",
+  },
+];
diff --git a/components/ocrspace/common/utils.mjs b/components/ocrspace/common/utils.mjs
new file mode 100644
index 0000000000000..b9920ab13c563
--- /dev/null
+++ b/components/ocrspace/common/utils.mjs
@@ -0,0 +1,31 @@
+import fs from "fs";
+
+export const isValidUrl = (urlString) => {
+  var urlPattern = new RegExp("^(https?:\\/\\/)?" + // validate protocol
+"((([a-z\\d]([a-z\\d-]*[a-z\\d])*)\\.)+[a-z]{2,}|" + // validate domain name
+"((\\d{1,3}\\.){3}\\d{1,3}))" + // validate OR ip (v4) address
+"(\\:\\d+)?(\\/[-a-z\\d%_.~+]*)*" + // validate port and path
+"(\\?[;&a-z\\d%_.~+=-]*)?" + // validate query string
+"(\\#[-a-z\\d_]*)?$", "i"); // validate fragment locator
+  return !!urlPattern.test(urlString);
+};
+
+export const checkTmp = (filename) => {
+  if (filename.indexOf("/tmp") === -1) {
+    return `/tmp/${filename}`;
+  }
+  return filename;
+};
+
+export const getUrlOrFile = (url) => {
+  if (!isValidUrl(url)) {
+    const data = fs.readFileSync(checkTmp(url));
+    const base64Image = Buffer.from(data, "binary").toString("base64");
+    return {
+      file: `data:image/jpeg;base64,${base64Image}`,
+    };
+  }
+  return {
+    url,
+  };
+};
diff --git a/components/ocrspace/ocrspace.app.mjs b/components/ocrspace/ocrspace.app.mjs
index ae49aa9d24882..ae96d7db00db7 100644
--- a/components/ocrspace/ocrspace.app.mjs
+++ b/components/ocrspace/ocrspace.app.mjs
@@ -1,363 +1,90 @@
 import { axios } from "@pipedream/platform";
-import FormData from "form-data";
+import {
+  IMAGE_FILETYPE_OPTIONS,
+  LANGUAGE_OPTIONS,
+  OCR_ENGINE_OPTIONS,
+} from "./common/constants.mjs";
 
 export default {
   type: "app",
   app: "ocrspace",
-  version: "0.0.{{ts}}",
   propDefinitions: {
-    webhookUrl: {
+    file: {
       type: "string",
-      label: "Webhook URL",
-      description: "The URL to receive notifications of completed OCR processing.",
+      label: "Image",
+      description: "The URL of the image or the path to the file saved to the `/tmp` directory  (e.g. `/tmp/example.jpg`)  to process. [See the documentation](https://pipedream.com/docs/workflows/steps/code/nodejs/working-with-files/#the-tmp-directory).",
     },
-    monitoredFolder: {
+    language: {
       type: "string",
-      label: "Monitored Folder",
-      description: "Optional folder to monitor for new file uploads.",
+      label: "Language",
+      description: "Language setting for image OCR processing.",
+      options: LANGUAGE_OPTIONS,
       optional: true,
     },
-    processingQueue: {
-      type: "string",
-      label: "Processing Queue",
-      description: "Optional processing queue to monitor for new file uploads.",
+    isOverlayRequired: {
+      type: "boolean",
+      label: "Is Overlay Required",
+      description: "If true, returns the coordinates of the bounding boxes for each word. If false, the OCR'ed text is returned only as a text block (this makes the JSON reponse smaller). Overlay data can be used, for example, to show [text over the image](https://ocr.space/english).",
       optional: true,
     },
-    imageUrl: {
+    filetype: {
       type: "string",
-      label: "Image File URL",
-      description: "The URL of the image file to submit for OCR processing.",
+      label: "File Type",
+      description: "Overwrites the automatic file type detection based on content-type. Supported image file formats are png, jpg (jpeg), gif, tif (tiff) and bmp. For document ocr, the api supports the Adobe PDF format. Multi-page TIFF files are supported.",
+      options: IMAGE_FILETYPE_OPTIONS,
       optional: true,
     },
-    imageFile: {
-      type: "file",
-      label: "Image File Upload",
-      description: "The image file to submit for OCR processing.",
+    detectOrientation: {
+      type: "boolean",
+      label: "Detect Orientation",
+      description: "If set to true, the api autorotates the image correctly and sets the TextOrientation parameter in the JSON response. If the image is not rotated, then TextOrientation=0, otherwise it is the degree of the rotation, e. g. \"270\".",
       optional: true,
     },
-    imageLanguage: {
-      type: "string",
-      label: "Image Language",
-      description: "Optional language setting for image OCR processing.",
+    scale: {
+      type: "boolean",
+      label: "Scale",
+      description: "If set to true, the api does some internal upscaling. This can improve the OCR result significantly, especially for low-resolution PDF scans. Note that the front page demo uses scale=true, but the API uses scale=false by default. See also this OCR forum post.",
       optional: true,
-      options: [
-        {
-          label: "Arabic",
-          value: "ara",
-        },
-        {
-          label: "Bulgarian",
-          value: "bul",
-        },
-        {
-          label: "Chinese (Simplified)",
-          value: "chs",
-        },
-        {
-          label: "Chinese (Traditional)",
-          value: "cht",
-        },
-        {
-          label: "Croatian",
-          value: "hrv",
-        },
-        {
-          label: "Czech",
-          value: "cze",
-        },
-        {
-          label: "Danish",
-          value: "dan",
-        },
-        {
-          label: "Dutch",
-          value: "dut",
-        },
-        {
-          label: "English",
-          value: "eng",
-        },
-        {
-          label: "Finnish",
-          value: "fin",
-        },
-        {
-          label: "French",
-          value: "fre",
-        },
-        {
-          label: "German",
-          value: "ger",
-        },
-        {
-          label: "Greek",
-          value: "gre",
-        },
-        {
-          label: "Hungarian",
-          value: "hun",
-        },
-        {
-          label: "Korean",
-          value: "kor",
-        },
-        {
-          label: "Italian",
-          value: "ita",
-        },
-        {
-          label: "Japanese",
-          value: "jpn",
-        },
-        {
-          label: "Polish",
-          value: "pol",
-        },
-        {
-          label: "Portuguese",
-          value: "por",
-        },
-        {
-          label: "Russian",
-          value: "rus",
-        },
-        {
-          label: "Slovenian",
-          value: "slv",
-        },
-        {
-          label: "Spanish",
-          value: "spa",
-        },
-        {
-          label: "Swedish",
-          value: "swe",
-        },
-        {
-          label: "Turkish",
-          value: "tur",
-        },
-      ],
     },
-    pdfUrl: {
-      type: "string",
-      label: "PDF File URL",
-      description: "The URL of the PDF file to submit for OCR processing.",
+    isTable: {
+      type: "boolean",
+      label: "Is Table",
+      description: "If set to true, the OCR logic makes sure that the parsed text result is always returned line by line. This switch is recommended for [table OCR](https://ocr.space/tablerecognition), [receipt OCR](https://ocr.space/receiptscanning), invoice processing and all other type of input documents that have a table like structure.",
       optional: true,
     },
-    pdfFile: {
-      type: "file",
-      label: "PDF File Upload",
-      description: "The PDF file to submit for OCR processing.",
-      optional: true,
-    },
-    pdfLanguage: {
+    ocrEngine: {
       type: "string",
-      label: "PDF Language",
-      description: "Optional language setting for PDF OCR processing.",
+      label: "OCR Engine",
+      description: "Engine 1 is default. [See OCR Engines](https://ocr.space/OCRAPI#ocrengine).",
+      options: OCR_ENGINE_OPTIONS,
       optional: true,
-      options: [
-        {
-          label: "Arabic",
-          value: "ara",
-        },
-        {
-          label: "Bulgarian",
-          value: "bul",
-        },
-        {
-          label: "Chinese (Simplified)",
-          value: "chs",
-        },
-        {
-          label: "Chinese (Traditional)",
-          value: "cht",
-        },
-        {
-          label: "Croatian",
-          value: "hrv",
-        },
-        {
-          label: "Czech",
-          value: "cze",
-        },
-        {
-          label: "Danish",
-          value: "dan",
-        },
-        {
-          label: "Dutch",
-          value: "dut",
-        },
-        {
-          label: "English",
-          value: "eng",
-        },
-        {
-          label: "Finnish",
-          value: "fin",
-        },
-        {
-          label: "French",
-          value: "fre",
-        },
-        {
-          label: "German",
-          value: "ger",
-        },
-        {
-          label: "Greek",
-          value: "gre",
-        },
-        {
-          label: "Hungarian",
-          value: "hun",
-        },
-        {
-          label: "Korean",
-          value: "kor",
-        },
-        {
-          label: "Italian",
-          value: "ita",
-        },
-        {
-          label: "Japanese",
-          value: "jpn",
-        },
-        {
-          label: "Polish",
-          value: "pol",
-        },
-        {
-          label: "Portuguese",
-          value: "por",
-        },
-        {
-          label: "Russian",
-          value: "rus",
-        },
-        {
-          label: "Slovenian",
-          value: "slv",
-        },
-        {
-          label: "Spanish",
-          value: "spa",
-        },
-        {
-          label: "Swedish",
-          value: "swe",
-        },
-        {
-          label: "Turkish",
-          value: "tur",
-        },
-      ],
-    },
-    pdfPages: {
-      type: "string",
-      label: "PDF Pages",
-      description: "Optional specific pages to process in the PDF file.",
-      optional: true,
-    },
-    jobId: {
-      type: "string",
-      label: "Job ID",
-      description: "The Job ID to retrieve the processed OCR result.",
     },
   },
   methods: {
     _baseUrl() {
       return "https://api.ocr.space";
     },
-    async _makeRequest(opts = {}) {
-      const {
-        $, method = "POST", path = "/parse/image", headers, ...otherOpts
-      } = opts;
+    _headers(headers = {}) {
+      return {
+        "apikey": this.$auth.apikey,
+        ...headers,
+      };
+    },
+    _makeRequest({
+      $ = this, path, headers, ...opts
+    }) {
       return axios($, {
-        ...otherOpts,
-        method,
         url: this._baseUrl() + path,
-        headers: {
-          ...headers,
-          apikey: this.$auth.apikey,
-        },
+        headers: this._headers(headers),
+        ...opts,
       });
     },
-    async submitImage(opts = {}) {
-      const {
-        imageUrl, imageFile, imageLanguage,
-      } = opts;
-      const formData = new FormData();
-      if (imageUrl) {
-        formData.append("url", imageUrl);
-      }
-      if (imageFile) {
-        formData.append("file", imageFile);
-      }
-      formData.append("isOverlayRequired", false);
-      if (imageLanguage) {
-        formData.append("language", imageLanguage);
-      }
+    processImage(opts = {}) {
       return this._makeRequest({
-        path: "/parse/image",
         method: "POST",
-        data: formData,
-        headers: formData.getHeaders(),
-      });
-    },
-    async submitPdf(opts = {}) {
-      const {
-        pdfUrl, pdfFile, pdfLanguage, pdfPages,
-      } = opts;
-      const formData = new FormData();
-      if (pdfUrl) {
-        formData.append("url", pdfUrl);
-      }
-      if (pdfFile) {
-        formData.append("file", pdfFile);
-      }
-      formData.append("isOverlayRequired", false);
-      if (pdfLanguage) {
-        formData.append("language", pdfLanguage);
-      }
-      if (pdfPages) {
-        formData.append("pages", pdfPages);
-      }
-      return this._makeRequest({
         path: "/parse/image",
-        method: "POST",
-        data: formData,
-        headers: formData.getHeaders(),
+        ...opts,
       });
     },
-    async retrieveOcrResult(opts = {}) {
-      const { jobId } = opts;
-      const path = `/parse/image/${jobId}`;
-      return this._makeRequest({
-        path,
-        method: "GET",
-      });
-    },
-    async emitOcrJobCompleted(data) {
-      const { webhookUrl } = this;
-      await axios(this, {
-        method: "POST",
-        url: webhookUrl,
-        data,
-      });
-    },
-    async emitNewFileUploaded(data) {
-      const { webhookUrl } = this;
-      await axios(this, {
-        method: "POST",
-        url: webhookUrl,
-        data,
-      });
-    },
-    authKeys() {
-      console.log(Object.keys(this.$auth));
-    },
   },
 };
diff --git a/components/ocrspace/package.json b/components/ocrspace/package.json
new file mode 100644
index 0000000000000..7c2dc40c7d722
--- /dev/null
+++ b/components/ocrspace/package.json
@@ -0,0 +1,18 @@
+{
+  "name": "@pipedream/ocrspace",
+  "version": "0.1.0",
+  "description": "Pipedream OCRSpace Components",
+  "main": "ocrspace.app.mjs",
+  "keywords": [
+    "pipedream",
+    "ocrspace"
+  ],
+  "homepage": "https://pipedream.com/apps/ocrspace",
+  "author": "Pipedream <support@pipedream.com> (https://pipedream.com/)",
+  "publishConfig": {
+    "access": "public"
+  },
+  "dependencies": {
+    "@pipedream/platform": "^3.0.3"
+  }
+}
diff --git a/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs b/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
deleted file mode 100644
index 72d297ec5ee32..0000000000000
--- a/components/ocrspace/sources/new-file-uploaded/new-file-uploaded.mjs
+++ /dev/null
@@ -1,111 +0,0 @@
-import {
-  axios, DEFAULT_POLLING_SOURCE_TIMER_INTERVAL,
-} from "@pipedream/platform";
-import ocrspace from "../../ocrspace.app.mjs";
-
-export default {
-  key: "ocrspace-new-file-uploaded",
-  name: "New OCR File Uploaded",
-  description: "Emit a new event when a new file is uploaded for OCR processing. [See the documentation]()",
-  version: "0.0.{{ts}}",
-  type: "source",
-  dedupe: "unique",
-  props: {
-    ocrspace,
-    db: "$.service.db",
-    timer: {
-      type: "$.interface.timer",
-      default: {
-        intervalSeconds: DEFAULT_POLLING_SOURCE_TIMER_INTERVAL,
-      },
-    },
-    monitoredFolder: {
-      type: "string",
-      label: "Monitored Folder",
-      description: "Optional folder to monitor for new file uploads.",
-      optional: true,
-    },
-    processingQueue: {
-      type: "string",
-      label: "Processing Queue",
-      description: "Optional processing queue to monitor for new file uploads.",
-      optional: true,
-    },
-  },
-  methods: {
-    async fetchUploadedFiles() {
-      const params = {};
-      if (this.monitoredFolder) {
-        params.folder = this.monitoredFolder;
-      }
-      if (this.processingQueue) {
-        params.queue = this.processingQueue;
-      }
-      const uploadedFiles = await this.ocrspace._makeRequest({
-        path: "/list/uploads",
-        method: "GET",
-        params,
-      });
-      return uploadedFiles;
-    },
-  },
-  hooks: {
-    async deploy() {
-      const files = await this.fetchUploadedFiles();
-      // Sort files by timestamp descending
-      files.sort((a, b) => new Date(b.timestamp) - new Date(a.timestamp));
-      const recentFiles = files.slice(0, 50).reverse();
-      for (const file of recentFiles) {
-        const id = file.id || new Date(file.timestamp).getTime();
-        const summary = `New file uploaded: ${file.name}`;
-        const ts = file.timestamp
-          ? new Date(file.timestamp).getTime()
-          : Date.now();
-        this.$emit(file, {
-          id,
-          summary,
-          ts,
-        });
-      }
-      if (files.length > 0) {
-        const latestFile = files[0];
-        const lastFileId = latestFile.id || new Date(latestFile.timestamp).getTime();
-        await this.db.set("lastFileId", lastFileId);
-      }
-    },
-    async activate() {
-      // No action needed on activate
-    },
-    async deactivate() {
-      // No action needed on deactivate
-    },
-  },
-  async run() {
-    const files = await this.fetchUploadedFiles();
-    const lastFileId = (await this.db.get("lastFileId")) || 0;
-    const newFiles = files.filter(
-      (file) =>
-        (file.id && file.id > lastFileId) ||
-        (file.timestamp && new Date(file.timestamp).getTime() > lastFileId),
-    );
-    // Sort newFiles by timestamp ascending
-    newFiles.sort((a, b) => new Date(a.timestamp) - new Date(b.timestamp));
-    for (const file of newFiles) {
-      const id = file.id || new Date(file.timestamp).getTime();
-      const summary = `New file uploaded: ${file.name}`;
-      const ts = file.timestamp
-        ? new Date(file.timestamp).getTime()
-        : Date.now();
-      this.$emit(file, {
-        id,
-        summary,
-        ts,
-      });
-    }
-    if (files.length > 0) {
-      const latestFile = files[0];
-      const lastFileId = latestFile.id || new Date(latestFile.timestamp).getTime();
-      await this.db.set("lastFileId", lastFileId);
-    }
-  },
-};
diff --git a/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs b/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs
deleted file mode 100644
index cfed5441c340f..0000000000000
--- a/components/ocrspace/sources/new-ocr-job-completed-instant/new-ocr-job-completed-instant.mjs
+++ /dev/null
@@ -1,67 +0,0 @@
-import ocrspace from "../../ocrspace.app.mjs";
-import { axios } from "@pipedream/platform";
-
-export default {
-  key: "ocrspace-new-ocr-job-completed-instant",
-  name: "New OCR Job Completed",
-  description: "Emit new event when an OCR job is completed. [See the documentation](https://ocr.space/ocrapi)",
-  version: "0.0.{{ts}}",
-  type: "source",
-  dedupe: "unique",
-  props: {
-    ocrspace: {
-      type: "app",
-      app: "ocrspace",
-    },
-    http: {
-      type: "$.interface.http",
-      customResponse: true,
-    },
-    db: "$.service.db",
-  },
-  hooks: {
-    async activate() {
-      console.log(`Please configure your OCRSpace webhook URL to: ${this.http.endpoint}`);
-    },
-    async deactivate() {
-      console.log("Webhook deactivated. Please remove the webhook URL from your OCRSpace settings.");
-    },
-    async deploy() {
-      const lastJobId = await this.db.get("lastJobId");
-      if (lastJobId) {
-        try {
-          const job = await this.ocrspace.retrieveOcrResult({
-            jobId: lastJobId,
-          });
-          this.$emit(job, {
-            id: job.jobId || lastJobId,
-            summary: `OCR job completed: ${job.jobId || lastJobId}`,
-            ts: Date.now(),
-          });
-        } catch (error) {
-          console.error(`Failed to retrieve OCR job with ID ${lastJobId}:`, error);
-        }
-      }
-    },
-  },
-  async run(event) {
-    const job = event;
-    const jobId = job.jobId || job.id || null;
-    const summary = jobId
-      ? `OCR job completed: ${jobId}`
-      : "OCR job completed";
-    const ts = job.completedAt
-      ? Date.parse(job.completedAt)
-      : Date.now();
-
-    this.$emit(job, {
-      id: jobId || ts.toString(),
-      summary,
-      ts,
-    });
-
-    if (jobId) {
-      await this.db.set("lastJobId", jobId);
-    }
-  },
-};

From ceb186ae48d57e99e9c5f27650d25b9375314db7 Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Thu, 16 Jan 2025 12:30:10 -0300
Subject: [PATCH 3/6] pnpm update

---
 pnpm-lock.yaml | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index ca391384a140c..6bff53a27f362 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -7099,6 +7099,12 @@ importers:
 
   components/ocr_web_service: {}
 
+  components/ocrspace:
+    dependencies:
+      '@pipedream/platform':
+        specifier: ^3.0.3
+        version: 3.0.3
+
   components/octoparse: {}
 
   components/octopus_deploy: {}
@@ -9592,8 +9598,7 @@ importers:
 
   components/showpad: {}
 
-  components/shutterstock:
-    specifiers: {}
+  components/shutterstock: {}
 
   components/sidetracker: {}
 
@@ -31013,6 +31018,8 @@ snapshots:
       '@putout/operator-filesystem': 5.0.0(putout@36.13.1(eslint@8.57.1)(typescript@5.6.3))
       '@putout/operator-json': 2.2.0
       putout: 36.13.1(eslint@8.57.1)(typescript@5.6.3)
+    transitivePeerDependencies:
+      - supports-color
 
   '@putout/operator-regexp@1.0.0(putout@36.13.1(eslint@8.57.1)(typescript@5.6.3))':
     dependencies:

From 5f17db7043793f63963ceb392fbddc6c652bc249 Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Thu, 16 Jan 2025 15:47:44 -0300
Subject: [PATCH 4/6] some adjusts

---
 components/ocrspace/actions/process-pdf/process-pdf.mjs | 2 +-
 components/ocrspace/common/utils.mjs                    | 7 +++++--
 components/ocrspace/package.json                        | 3 ++-
 3 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/components/ocrspace/actions/process-pdf/process-pdf.mjs b/components/ocrspace/actions/process-pdf/process-pdf.mjs
index 4669979849a62..cda0790f5c620 100644
--- a/components/ocrspace/actions/process-pdf/process-pdf.mjs
+++ b/components/ocrspace/actions/process-pdf/process-pdf.mjs
@@ -15,7 +15,7 @@ export default {
         "file",
       ],
       label: "PDF File",
-      description: "The URL of the PDF file or the path to the file saved to the `/tmp` directory  (e.g. `/tmp/example.jpg`)  to process. [See the documentation](https://pipedream.com/docs/workflows/steps/code/nodejs/working-with-files/#the-tmp-directory).",
+      description: "The URL of the PDF file or the path to the file saved to the `/tmp` directory  (e.g. `/tmp/example.pdf`)  to process. [See the documentation](https://pipedream.com/docs/workflows/steps/code/nodejs/working-with-files/#the-tmp-directory).",
     },
   },
   methods: {
diff --git a/components/ocrspace/common/utils.mjs b/components/ocrspace/common/utils.mjs
index b9920ab13c563..3384dcc350320 100644
--- a/components/ocrspace/common/utils.mjs
+++ b/components/ocrspace/common/utils.mjs
@@ -1,4 +1,5 @@
 import fs from "fs";
+import mime from "mime";
 
 export const isValidUrl = (urlString) => {
   var urlPattern = new RegExp("^(https?:\\/\\/)?" + // validate protocol
@@ -19,10 +20,12 @@ export const checkTmp = (filename) => {
 
 export const getUrlOrFile = (url) => {
   if (!isValidUrl(url)) {
-    const data = fs.readFileSync(checkTmp(url));
+    const filePath = checkTmp(url);
+    const data = fs.readFileSync(filePath);
+    const mimeType = mime.getType(filePath);
     const base64Image = Buffer.from(data, "binary").toString("base64");
     return {
-      file: `data:image/jpeg;base64,${base64Image}`,
+      file: `data:${mimeType};base64,${base64Image}`,
     };
   }
   return {
diff --git a/components/ocrspace/package.json b/components/ocrspace/package.json
index 7c2dc40c7d722..a0ab1ae330c10 100644
--- a/components/ocrspace/package.json
+++ b/components/ocrspace/package.json
@@ -13,6 +13,7 @@
     "access": "public"
   },
   "dependencies": {
-    "@pipedream/platform": "^3.0.3"
+    "@pipedream/platform": "^3.0.3",
+    "mime": "^4.0.6"
   }
 }

From 9545167a0daa87056f86029be6fd7521c2fb26ae Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Thu, 16 Jan 2025 15:51:26 -0300
Subject: [PATCH 5/6] pnpm update

---
 pnpm-lock.yaml | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index 6bff53a27f362..b484508c2f1d9 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -7104,6 +7104,9 @@ importers:
       '@pipedream/platform':
         specifier: ^3.0.3
         version: 3.0.3
+      mime:
+        specifier: ^4.0.6
+        version: 4.0.6
 
   components/octoparse: {}
 
@@ -22379,6 +22382,11 @@ packages:
     engines: {node: '>=16'}
     hasBin: true
 
+  mime@4.0.6:
+    resolution: {integrity: sha512-4rGt7rvQHBbaSOF9POGkk1ocRP16Md1x36Xma8sz8h8/vfCUI2OtEIeCqe4Ofes853x4xDoPiFLIT47J5fI/7A==}
+    engines: {node: '>=16'}
+    hasBin: true
+
   mimer@2.0.2:
     resolution: {integrity: sha512-izxvjsB7Ur5HrTbPu6VKTrzxSMBFBqyZQc6dWlZNQ4/wAvf886fD4lrjtFd8IQ8/WmZKdxKjUtqFFNaj3hQ52g==}
     engines: {node: '>= 12'}
@@ -39679,6 +39687,8 @@ snapshots:
 
   mime@4.0.4: {}
 
+  mime@4.0.6: {}
+
   mimer@2.0.2: {}
 
   mimic-fn@2.1.0: {}

From 39ec3bc903a9c9196bc840f185c2dbca2a9d87d1 Mon Sep 17 00:00:00 2001
From: Luan Cazarine <luanhc@gmail.com>
Date: Fri, 17 Jan 2025 11:57:35 -0300
Subject: [PATCH 6/6] fix file field name

---
 components/ocrspace/actions/common/process-base.mjs | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/components/ocrspace/actions/common/process-base.mjs b/components/ocrspace/actions/common/process-base.mjs
index adc8f0cfffb4c..2a8fbc60500d1 100644
--- a/components/ocrspace/actions/common/process-base.mjs
+++ b/components/ocrspace/actions/common/process-base.mjs
@@ -1,3 +1,4 @@
+import { ConfigurationError } from "@pipedream/platform";
 import FormData from "form-data";
 import { getUrlOrFile } from "../../common/utils.mjs";
 import ocrspace from "../../ocrspace.app.mjs";
@@ -49,7 +50,7 @@ export default {
     } = getUrlOrFile(this.file);
 
     if (url) data.append("url", url);
-    if (file) data.append("file", file);
+    if (file) data.append("base64Image", file);
     if (this.imageLanguage) data.append("language", this.imageLanguage);
     if (this.isOverlayRequired) data.append("isOverlayRequired", `${this.isOverlayRequired}`);
     if (this.filetype) data.append("filetype", this.filetype);
@@ -65,6 +66,11 @@ export default {
     });
 
     $.export("$summary", this.getSummary());
+
+    if (response.ErrorMessage) {
+      throw new ConfigurationError(response.ErrorMessage[0]);
+    }
+
     return response;
   },
 };