feat(textract): add documentation handling for textract

camunda · Jan 7, 2025 · 0b49829 · 0b49829
1 parent 3d20ed2
commit 0b49829
Show file tree

Hide file tree

Showing 10 changed files with 276 additions and 24 deletions.
diff --git a/connectors/aws/aws-textract/element-templates/aws-textract-outbound-connector.json b/connectors/aws/aws-textract/element-templates/aws-textract-outbound-connector.json
@@ -7,7 +7,7 @@
     "keywords" : [ "extract text", "extract data", "extract text from image", "extract data from image", "ocr" ]
   },
   "documentationRef" : "https://docs.camunda.io/docs/next/components/connectors/out-of-the-box-connectors/amazon-textract/",
-  "version" : 1,
+  "version" : 2,
   "category" : {
     "id" : "connectors",
     "name" : "Connectors"
@@ -151,6 +151,29 @@
       "name" : "Polling",
       "value" : "POLLING"
     } ]
+  }, {
+    "id" : "input.documentLocationType",
+    "label" : "Document location type",
+    "description" : "Document location",
+    "optional" : false,
+    "group" : "input",
+    "binding" : {
+      "name" : "input.documentLocationType",
+      "type" : "zeebe:input"
+    },
+    "condition" : {
+      "property" : "input.executionType",
+      "equals" : "SYNC",
+      "type" : "simple"
+    },
+    "type" : "Dropdown",
+    "choices" : [ {
+      "name" : "S3",
+      "value" : "S3"
+    }, {
+      "name" : "Uploaded document",
+      "value" : "UPLOADED"
+    } ]
   }, {
     "id" : "input.documentS3Bucket",
     "label" : "Document bucket",
@@ -165,6 +188,11 @@
       "name" : "input.documentS3Bucket",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.documentName",
@@ -180,6 +208,11 @@
       "name" : "input.documentName",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.documentVersion",
@@ -192,6 +225,11 @@
       "name" : "input.documentVersion",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.analyzeTables",
@@ -372,6 +410,25 @@
       "type" : "simple"
     },
     "type" : "String"
+  }, {
+    "id" : "input.document",
+    "label" : "Document",
+    "optional" : false,
+    "constraints" : {
+      "notEmpty" : true
+    },
+    "feel" : "required",
+    "group" : "input",
+    "binding" : {
+      "name" : "input.document",
+      "type" : "zeebe:input"
+    },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "UPLOADED",
+      "type" : "simple"
+    },
+    "type" : "String"
   }, {
     "id" : "resultVariable",
     "label" : "Result variable",

diff --git a/...ors/aws/aws-textract/element-templates/hybrid/aws-textract-outbound-connector-hybrid.json b/...ors/aws/aws-textract/element-templates/hybrid/aws-textract-outbound-connector-hybrid.json
@@ -7,7 +7,7 @@
     "keywords" : [ "extract text", "extract data", "extract text from image", "extract data from image", "ocr" ]
   },
   "documentationRef" : "https://docs.camunda.io/docs/next/components/connectors/out-of-the-box-connectors/amazon-textract/",
-  "version" : 1,
+  "version" : 2,
   "category" : {
     "id" : "connectors",
     "name" : "Connectors"
@@ -156,6 +156,29 @@
       "name" : "Polling",
       "value" : "POLLING"
     } ]
+  }, {
+    "id" : "input.documentLocationType",
+    "label" : "Document location type",
+    "description" : "Document location",
+    "optional" : false,
+    "group" : "input",
+    "binding" : {
+      "name" : "input.documentLocationType",
+      "type" : "zeebe:input"
+    },
+    "condition" : {
+      "property" : "input.executionType",
+      "equals" : "SYNC",
+      "type" : "simple"
+    },
+    "type" : "Dropdown",
+    "choices" : [ {
+      "name" : "S3",
+      "value" : "S3"
+    }, {
+      "name" : "Uploaded document",
+      "value" : "UPLOADED"
+    } ]
   }, {
     "id" : "input.documentS3Bucket",
     "label" : "Document bucket",
@@ -170,6 +193,11 @@
       "name" : "input.documentS3Bucket",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.documentName",
@@ -185,6 +213,11 @@
       "name" : "input.documentName",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.documentVersion",
@@ -197,6 +230,11 @@
       "name" : "input.documentVersion",
       "type" : "zeebe:input"
     },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "S3",
+      "type" : "simple"
+    },
     "type" : "String"
   }, {
     "id" : "input.analyzeTables",
@@ -377,6 +415,25 @@
       "type" : "simple"
     },
     "type" : "String"
+  }, {
+    "id" : "input.document",
+    "label" : "Document",
+    "optional" : false,
+    "constraints" : {
+      "notEmpty" : true
+    },
+    "feel" : "required",
+    "group" : "input",
+    "binding" : {
+      "name" : "input.document",
+      "type" : "zeebe:input"
+    },
+    "condition" : {
+      "property" : "input.documentLocationType",
+      "equals" : "UPLOADED",
+      "type" : "simple"
+    },
+    "type" : "String"
   }, {
     "id" : "resultVariable",
     "label" : "Result variable",

diff --git a/...s/aws-textract/src/main/java/io/camunda/connector/textract/TextractConnectorFunction.java b/...s/aws-textract/src/main/java/io/camunda/connector/textract/TextractConnectorFunction.java
@@ -34,7 +34,7 @@
               "ocr"
             }),
     inputDataClass = TextractRequest.class,
-    version = 1,
+    version = 2,
     propertyGroups = {
       @ElementTemplate.PropertyGroup(id = "authentication", label = "Authentication"),
       @ElementTemplate.PropertyGroup(id = "configuration", label = "Configuration"),

diff --git a/...s/aws-textract/src/main/java/io/camunda/connector/textract/caller/SyncTextractCaller.java b/...s/aws-textract/src/main/java/io/camunda/connector/textract/caller/SyncTextractCaller.java
@@ -11,6 +11,8 @@
 import com.amazonaws.services.textract.model.AnalyzeDocumentResult;
 import com.amazonaws.services.textract.model.Document;
 import io.camunda.connector.textract.model.TextractRequestData;
+import java.nio.ByteBuffer;
+import java.util.Objects;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -22,7 +24,8 @@ public class SyncTextractCaller implements TextractCaller<AnalyzeDocumentResult>
   public AnalyzeDocumentResult call(
       TextractRequestData requestData, AmazonTextract textractClient) {
     LOGGER.debug("Starting sync task for document analysis with request data: {}", requestData);
-    final Document document = new Document().withS3Object(prepareS3Obj(requestData));
+
+    final Document document = createDocument(requestData);
 
     final AnalyzeDocumentRequest analyzeDocumentRequest =
         new AnalyzeDocumentRequest()
@@ -31,4 +34,16 @@ public AnalyzeDocumentResult call(
 
     return textractClient.analyzeDocument(analyzeDocumentRequest);
   }
+
+  private Document createDocument(TextractRequestData requestData) {
+    final Document document = new Document();
+
+    if (Objects.isNull(requestData.document())) {
+      return document.withS3Object(prepareS3Obj(requestData));
+    }
+
+    byte[] docBytes = requestData.document().asByteArray();
+    document.withBytes(ByteBuffer.wrap(docBytes));
+    return document;
+  }
 }
diff --git a/.../aws-textract/src/main/java/io/camunda/connector/textract/model/DocumentLocationType.java b/.../aws-textract/src/main/java/io/camunda/connector/textract/model/DocumentLocationType.java
@@ -0,0 +1,12 @@
+/*
+ * Copyright Camunda Services GmbH and/or licensed to Camunda Services GmbH
+ * under one or more contributor license agreements. Licensed under a proprietary license.
+ * See the License.txt file for more information. You may not use this file
+ * except in compliance with the proprietary license.
+ */
+package io.camunda.connector.textract.model;
+
+public enum DocumentLocationType {
+  S3,
+  UPLOADED;
+}
diff --git a/...s/aws-textract/src/main/java/io/camunda/connector/textract/model/TextractRequestData.java b/...s/aws-textract/src/main/java/io/camunda/connector/textract/model/TextractRequestData.java
@@ -6,10 +6,11 @@
  */
 package io.camunda.connector.textract.model;
 
+import io.camunda.connector.generator.dsl.Property;
 import io.camunda.connector.generator.dsl.Property.FeelMode;
 import io.camunda.connector.generator.java.annotation.TemplateProperty;
+import io.camunda.document.Document;
 import jakarta.validation.constraints.AssertTrue;
-import jakarta.validation.constraints.NotBlank;
 import jakarta.validation.constraints.NotNull;
 import org.apache.commons.lang3.StringUtils;
 
@@ -28,23 +29,52 @@ public record TextractRequestData(
             description = "Endpoint inference type")
         @NotNull
         TextractExecutionType executionType,
+    @TemplateProperty(
+            group = "input",
+            label = "Document location type",
+            description = "Document location",
+            feel = FeelMode.disabled,
+            type = TemplateProperty.PropertyType.Dropdown,
+            choices = {
+              @TemplateProperty.DropdownPropertyChoice(value = "S3", label = "S3"),
+              @TemplateProperty.DropdownPropertyChoice(
+                  value = "UPLOADED",
+                  label = "Uploaded document")
+            },
+            condition =
+                @TemplateProperty.PropertyCondition(
+                    property = "input.executionType",
+                    equals = "SYNC"))
+        DocumentLocationType documentLocationType,
     @TemplateProperty(
             group = "input",
             label = "Document bucket",
-            description = "S3 bucket that contains document that needs to be processed")
-        @NotBlank
+            description = "S3 bucket that contains document that needs to be processed",
+            condition =
+                @TemplateProperty.PropertyCondition(
+                    property = "input.documentLocationType",
+                    equals = "S3"),
+            constraints = @TemplateProperty.PropertyConstraints(notEmpty = true))
         String documentS3Bucket,
     @TemplateProperty(
             group = "input",
             label = "Document path",
-            description = "S3 document path to be processed")
-        @NotBlank
+            description = "S3 document path to be processed",
+            condition =
+                @TemplateProperty.PropertyCondition(
+                    property = "input.documentLocationType",
+                    equals = "S3"),
+            constraints = @TemplateProperty.PropertyConstraints(notEmpty = true))
         String documentName,
     @TemplateProperty(
             group = "input",
             label = "Document version",
             description = "S3 document version to be processed",
-            optional = true)
+            optional = true,
+            condition =
+                @TemplateProperty.PropertyCondition(
+                    property = "input.documentLocationType",
+                    equals = "S3"))
         String documentVersion,
     @TemplateProperty(
             label = "Analyze tables",
@@ -150,14 +180,25 @@ public record TextractRequestData(
                 @TemplateProperty.PropertyCondition(
                     property = "input.executionType",
                     equals = "ASYNC"))
-        String outputConfigS3Prefix) {
+        String outputConfigS3Prefix,
+    @TemplateProperty(
+            group = "input",
+            label = "Document",
+            feel = Property.FeelMode.required,
+            type = TemplateProperty.PropertyType.String,
+            condition =
+                @TemplateProperty.PropertyCondition(
+                    property = "input.documentLocationType",
+                    equals = "UPLOADED"),
+            constraints = @TemplateProperty.PropertyConstraints(notEmpty = true))
+        Document document) {
   @TemplateProperty(ignore = true)
   public static final String WRONG_OUTPUT_VALUES_MSG =
       "Output S3 bucket must be filled in if output S3 prefix is filled in";
 
   @TemplateProperty(ignore = true)
   public static final String WRONG_NOTIFICATION_VALUES_MSG =
-      "either both notification values role ARN and topic ARN must be filled in or none of them";
+      "Either both notification values role ARN and topic ARN must be filled in or none of them";
 
   @AssertTrue(message = WRONG_NOTIFICATION_VALUES_MSG)
   public boolean isValidNotificationProperties() {

diff --git a/...-textract/src/test/java/io/camunda/connector/textract/caller/AsyncTextractCallerTest.java b/...-textract/src/test/java/io/camunda/connector/textract/caller/AsyncTextractCallerTest.java
@@ -14,6 +14,7 @@
 import com.amazonaws.services.textract.AmazonTextractAsyncClient;
 import com.amazonaws.services.textract.model.StartDocumentAnalysisRequest;
 import com.amazonaws.services.textract.model.StartDocumentAnalysisResult;
+import io.camunda.connector.textract.model.DocumentLocationType;
 import io.camunda.connector.textract.model.TextractExecutionType;
 import io.camunda.connector.textract.model.TextractRequestData;
 import org.junit.jupiter.api.Test;
@@ -100,6 +101,7 @@ void callWithoutOutputS3BucketShouldNotCreateOutputObj() {
   private TextractRequestData prepareReqData(String roleArn, String topicArn) {
     return new TextractRequestData(
         TextractExecutionType.ASYNC,
+        DocumentLocationType.S3,
         "test-bucket",
         "test-object",
         "1",
@@ -113,12 +115,14 @@ private TextractRequestData prepareReqData(String roleArn, String topicArn) {
         roleArn,
         topicArn,
         "outputBucket",
-        "prefix");
+        "prefix",
+        null);
   }
 
   private TextractRequestData prepareReqDataWithoutOutputS3Bucket() {
     return new TextractRequestData(
         TextractExecutionType.ASYNC,
+        DocumentLocationType.S3,
         "test-bucket",
         "test-object",
         "1",
@@ -132,6 +136,7 @@ private TextractRequestData prepareReqDataWithoutOutputS3Bucket() {
         "roleArn",
         "topicArn",
         "",
-        "prefix");
+        "prefix",
+        null);
   }
 }