aws-sdk-comprehend 1.50.0 → 1.51.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e6b8c990eb8f22c9347bb33832a67cbb7decb7c3e2d3e26f5c55645c6f66112e
-  data.tar.gz: 3480e954c55a8a1fa34a2fc713e7ef70fef60b6b788e4962037b186bad2941d5
+  metadata.gz: 6a9828628812834dcbad9932c47945231e7ba7f05b3948b59a349d0c7058f09f
+  data.tar.gz: 5141ec47d75e05488aa0fd4304c573d2e903b4072d7215f45aa30b1a42c33f80
 SHA512:
-  metadata.gz: 77a1b79f708c26bd95b417c884e8c0fc4c3245e20b4eca2ee4f6cd7a65c9c712617ad6a5123a080ce72bab677cbc41b880103a360253ee84c35eb43e327db97b
-  data.tar.gz: ae140ad7047606ac4e7db331f20aaf81825caa98909f8e39e235ea3d741405e854d38f3d39a04cfb478ec23a32e80e829e4c5df688685698b160ba2e2a380e17
+  metadata.gz: 1f350f13013eb12272326dc8aea2df7dbf74d5e2ee2810febfcfd2badd7ae7169f1a68cff10b8dde8a8bdce477064169bf61a26f3bbf4e5b57077cf0ddd2fc1b
+  data.tar.gz: 4d40609712cf892f49e69f14b20e7a4f4c781b4803336e0bc2803eb4839d5ff072e8143c3d31b74ae10a02b6055349a22374ee72b9123b25c2af5cc5ebcf472d

data/CHANGELOG.md

@@ -1,6 +1,11 @@
 Unreleased Changes
 ------------------
 
+1.51.0 (2021-09-14)
+------------------
+
+* Feature - Amazon Comprehend now allows you to train and run PDF and Word documents for custom entity recognition. With PDF and Word formats, you can extract information from documents containing headers, lists and tables.
+
 1.50.0 (2021-09-01)
 ------------------
 

data/VERSION

@@ -1 +1 @@
-1.50.0
+1.51.0

data/lib/aws-sdk-comprehend/client.rb

@@ -761,6 +761,9 @@ module Aws::Comprehend
     #         {
     #           s3_uri: "S3Uri", # required
     #           attribute_names: ["AttributeNamesListItem"], # required
+    #           annotation_data_s3_uri: "S3Uri",
+    #           source_documents_s3_uri: "S3Uri",
+    #           document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #         },
     #       ],
     #     },
@@ -965,6 +968,9 @@ module Aws::Comprehend
     #         {
     #           s3_uri: "S3Uri", # required
     #           attribute_names: ["AttributeNamesListItem"], # required
+    #           annotation_data_s3_uri: "S3Uri",
+    #           source_documents_s3_uri: "S3Uri",
+    #           document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #         },
     #       ],
     #     },
@@ -1107,6 +1113,10 @@ module Aws::Comprehend
     #   resp.document_classification_job_properties.document_classifier_arn #=> String
     #   resp.document_classification_job_properties.input_data_config.s3_uri #=> String
     #   resp.document_classification_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.document_classification_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.document_classification_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.document_classification_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.document_classification_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.document_classification_job_properties.output_data_config.s3_uri #=> String
     #   resp.document_classification_job_properties.output_data_config.kms_key_id #=> String
     #   resp.document_classification_job_properties.data_access_role_arn #=> String
@@ -1158,6 +1168,9 @@ module Aws::Comprehend
     #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].s3_uri #=> String
     #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].attribute_names #=> Array
     #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].attribute_names[0] #=> String
+    #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].annotation_data_s3_uri #=> String
+    #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].source_documents_s3_uri #=> String
+    #   resp.document_classifier_properties.input_data_config.augmented_manifests[0].document_type #=> String, one of "PLAIN_TEXT_DOCUMENT", "SEMI_STRUCTURED_DOCUMENT"
     #   resp.document_classifier_properties.output_data_config.s3_uri #=> String
     #   resp.document_classifier_properties.output_data_config.kms_key_id #=> String
     #   resp.document_classifier_properties.classifier_metadata.number_of_labels #=> Integer
@@ -1217,6 +1230,10 @@ module Aws::Comprehend
     #   resp.dominant_language_detection_job_properties.end_time #=> Time
     #   resp.dominant_language_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.dominant_language_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.dominant_language_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.dominant_language_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.dominant_language_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.dominant_language_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.dominant_language_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.dominant_language_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.dominant_language_detection_job_properties.data_access_role_arn #=> String
@@ -1301,6 +1318,10 @@ module Aws::Comprehend
     #   resp.entities_detection_job_properties.entity_recognizer_arn #=> String
     #   resp.entities_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.entities_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.entities_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.entities_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.entities_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.entities_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.entities_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.entities_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.entities_detection_job_properties.language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -1357,6 +1378,9 @@ module Aws::Comprehend
     #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].s3_uri #=> String
     #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].attribute_names #=> Array
     #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].attribute_names[0] #=> String
+    #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].annotation_data_s3_uri #=> String
+    #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].source_documents_s3_uri #=> String
+    #   resp.entity_recognizer_properties.input_data_config.augmented_manifests[0].document_type #=> String, one of "PLAIN_TEXT_DOCUMENT", "SEMI_STRUCTURED_DOCUMENT"
     #   resp.entity_recognizer_properties.recognizer_metadata.number_of_trained_documents #=> Integer
     #   resp.entity_recognizer_properties.recognizer_metadata.number_of_test_documents #=> Integer
     #   resp.entity_recognizer_properties.recognizer_metadata.evaluation_metrics.precision #=> Float
@@ -1411,6 +1435,10 @@ module Aws::Comprehend
     #   resp.events_detection_job_properties.end_time #=> Time
     #   resp.events_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.events_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.events_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.events_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.events_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.events_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.events_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.events_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.events_detection_job_properties.language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -1455,6 +1483,10 @@ module Aws::Comprehend
     #   resp.key_phrases_detection_job_properties.end_time #=> Time
     #   resp.key_phrases_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.key_phrases_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.key_phrases_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.key_phrases_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.key_phrases_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.key_phrases_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.key_phrases_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.key_phrases_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.key_phrases_detection_job_properties.language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -1502,6 +1534,10 @@ module Aws::Comprehend
     #   resp.pii_entities_detection_job_properties.end_time #=> Time
     #   resp.pii_entities_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.pii_entities_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.pii_entities_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.pii_entities_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.pii_entities_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.pii_entities_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.pii_entities_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.pii_entities_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.pii_entities_detection_job_properties.redaction_config.pii_entity_types #=> Array
@@ -1549,6 +1585,10 @@ module Aws::Comprehend
     #   resp.sentiment_detection_job_properties.end_time #=> Time
     #   resp.sentiment_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.sentiment_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.sentiment_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.sentiment_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.sentiment_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.sentiment_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.sentiment_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.sentiment_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.sentiment_detection_job_properties.language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -1595,6 +1635,10 @@ module Aws::Comprehend
     #   resp.topics_detection_job_properties.end_time #=> Time
     #   resp.topics_detection_job_properties.input_data_config.s3_uri #=> String
     #   resp.topics_detection_job_properties.input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.topics_detection_job_properties.input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.topics_detection_job_properties.input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.topics_detection_job_properties.input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.topics_detection_job_properties.input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.topics_detection_job_properties.output_data_config.s3_uri #=> String
     #   resp.topics_detection_job_properties.output_data_config.kms_key_id #=> String
     #   resp.topics_detection_job_properties.number_of_topics #=> Integer
@@ -1916,6 +1960,10 @@ module Aws::Comprehend
     #   resp.document_classification_job_properties_list[0].document_classifier_arn #=> String
     #   resp.document_classification_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.document_classification_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.document_classification_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.document_classification_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.document_classification_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.document_classification_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.document_classification_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.document_classification_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.document_classification_job_properties_list[0].data_access_role_arn #=> String
@@ -1986,6 +2034,9 @@ module Aws::Comprehend
     #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].s3_uri #=> String
     #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].attribute_names #=> Array
     #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].attribute_names[0] #=> String
+    #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].annotation_data_s3_uri #=> String
+    #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].source_documents_s3_uri #=> String
+    #   resp.document_classifier_properties_list[0].input_data_config.augmented_manifests[0].document_type #=> String, one of "PLAIN_TEXT_DOCUMENT", "SEMI_STRUCTURED_DOCUMENT"
     #   resp.document_classifier_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.document_classifier_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.document_classifier_properties_list[0].classifier_metadata.number_of_labels #=> Integer
@@ -2065,6 +2116,10 @@ module Aws::Comprehend
     #   resp.dominant_language_detection_job_properties_list[0].end_time #=> Time
     #   resp.dominant_language_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.dominant_language_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.dominant_language_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.dominant_language_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.dominant_language_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.dominant_language_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.dominant_language_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.dominant_language_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.dominant_language_detection_job_properties_list[0].data_access_role_arn #=> String
@@ -2186,6 +2241,10 @@ module Aws::Comprehend
     #   resp.entities_detection_job_properties_list[0].entity_recognizer_arn #=> String
     #   resp.entities_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.entities_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.entities_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.entities_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.entities_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.entities_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.entities_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.entities_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.entities_detection_job_properties_list[0].language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -2267,6 +2326,9 @@ module Aws::Comprehend
     #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].s3_uri #=> String
     #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].attribute_names #=> Array
     #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].attribute_names[0] #=> String
+    #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].annotation_data_s3_uri #=> String
+    #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].source_documents_s3_uri #=> String
+    #   resp.entity_recognizer_properties_list[0].input_data_config.augmented_manifests[0].document_type #=> String, one of "PLAIN_TEXT_DOCUMENT", "SEMI_STRUCTURED_DOCUMENT"
     #   resp.entity_recognizer_properties_list[0].recognizer_metadata.number_of_trained_documents #=> Integer
     #   resp.entity_recognizer_properties_list[0].recognizer_metadata.number_of_test_documents #=> Integer
     #   resp.entity_recognizer_properties_list[0].recognizer_metadata.evaluation_metrics.precision #=> Float
@@ -2341,6 +2403,10 @@ module Aws::Comprehend
     #   resp.events_detection_job_properties_list[0].end_time #=> Time
     #   resp.events_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.events_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.events_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.events_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.events_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.events_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.events_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.events_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.events_detection_job_properties_list[0].language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -2404,6 +2470,10 @@ module Aws::Comprehend
     #   resp.key_phrases_detection_job_properties_list[0].end_time #=> Time
     #   resp.key_phrases_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.key_phrases_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.key_phrases_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.key_phrases_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.key_phrases_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.key_phrases_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.key_phrases_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.key_phrases_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.key_phrases_detection_job_properties_list[0].language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -2467,6 +2537,10 @@ module Aws::Comprehend
     #   resp.pii_entities_detection_job_properties_list[0].end_time #=> Time
     #   resp.pii_entities_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.pii_entities_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.pii_entities_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.pii_entities_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.pii_entities_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.pii_entities_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.pii_entities_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.pii_entities_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.pii_entities_detection_job_properties_list[0].redaction_config.pii_entity_types #=> Array
@@ -2533,6 +2607,10 @@ module Aws::Comprehend
     #   resp.sentiment_detection_job_properties_list[0].end_time #=> Time
     #   resp.sentiment_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.sentiment_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.sentiment_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.sentiment_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.sentiment_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.sentiment_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.sentiment_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.sentiment_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.sentiment_detection_job_properties_list[0].language_code #=> String, one of "en", "es", "fr", "de", "it", "pt", "ar", "hi", "ja", "ko", "zh", "zh-TW"
@@ -2632,6 +2710,10 @@ module Aws::Comprehend
     #   resp.topics_detection_job_properties_list[0].end_time #=> Time
     #   resp.topics_detection_job_properties_list[0].input_data_config.s3_uri #=> String
     #   resp.topics_detection_job_properties_list[0].input_data_config.input_format #=> String, one of "ONE_DOC_PER_FILE", "ONE_DOC_PER_LINE"
+    #   resp.topics_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_action #=> String, one of "TEXTRACT_DETECT_DOCUMENT_TEXT", "TEXTRACT_ANALYZE_DOCUMENT"
+    #   resp.topics_detection_job_properties_list[0].input_data_config.document_reader_config.document_read_mode #=> String, one of "SERVICE_DEFAULT", "FORCE_DOCUMENT_READ_ACTION"
+    #   resp.topics_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types #=> Array
+    #   resp.topics_detection_job_properties_list[0].input_data_config.document_reader_config.feature_types[0] #=> String, one of "TABLES", "FORMS"
     #   resp.topics_detection_job_properties_list[0].output_data_config.s3_uri #=> String
     #   resp.topics_detection_job_properties_list[0].output_data_config.kms_key_id #=> String
     #   resp.topics_detection_job_properties_list[0].number_of_topics #=> Integer
@@ -2720,6 +2802,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -2823,6 +2910,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -2942,6 +3034,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3028,6 +3125,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3133,6 +3235,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3227,6 +3334,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3337,6 +3449,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3447,6 +3564,11 @@ module Aws::Comprehend
     #     input_data_config: { # required
     #       s3_uri: "S3Uri", # required
     #       input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #       document_reader_config: {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       },
     #     },
     #     output_data_config: { # required
     #       s3_uri: "S3Uri", # required
@@ -3881,7 +4003,7 @@ module Aws::Comprehend
         params: params,
         config: config)
       context[:gem_name] = 'aws-sdk-comprehend'
-      context[:gem_version] = '1.50.0'
+      context[:gem_version] = '1.51.0'
       Seahorse::Client::Request.new(handlers, context)
     end
 

data/lib/aws-sdk-comprehend/client_api.rb

@@ -16,6 +16,7 @@ module Aws::Comprehend
     AnyLengthString = Shapes::StringShape.new(name: 'AnyLengthString')
     AttributeNamesList = Shapes::ListShape.new(name: 'AttributeNamesList')
     AttributeNamesListItem = Shapes::StringShape.new(name: 'AttributeNamesListItem')
+    AugmentedManifestsDocumentTypeFormat = Shapes::StringShape.new(name: 'AugmentedManifestsDocumentTypeFormat')
     AugmentedManifestsListItem = Shapes::StructureShape.new(name: 'AugmentedManifestsListItem')
     BatchDetectDominantLanguageItemResult = Shapes::StructureShape.new(name: 'BatchDetectDominantLanguageItemResult')
     BatchDetectDominantLanguageRequest = Shapes::StructureShape.new(name: 'BatchDetectDominantLanguageRequest')
@@ -111,6 +112,10 @@ module Aws::Comprehend
     DocumentClassifierProperties = Shapes::StructureShape.new(name: 'DocumentClassifierProperties')
     DocumentClassifierPropertiesList = Shapes::ListShape.new(name: 'DocumentClassifierPropertiesList')
     DocumentLabel = Shapes::StructureShape.new(name: 'DocumentLabel')
+    DocumentReadAction = Shapes::StringShape.new(name: 'DocumentReadAction')
+    DocumentReadFeatureTypes = Shapes::StringShape.new(name: 'DocumentReadFeatureTypes')
+    DocumentReadMode = Shapes::StringShape.new(name: 'DocumentReadMode')
+    DocumentReaderConfig = Shapes::StructureShape.new(name: 'DocumentReaderConfig')
     DominantLanguage = Shapes::StructureShape.new(name: 'DominantLanguage')
     DominantLanguageDetectionJobFilter = Shapes::StructureShape.new(name: 'DominantLanguageDetectionJobFilter')
     DominantLanguageDetectionJobProperties = Shapes::StructureShape.new(name: 'DominantLanguageDetectionJobProperties')
@@ -192,6 +197,7 @@ module Aws::Comprehend
     ListOfDetectKeyPhrasesResult = Shapes::ListShape.new(name: 'ListOfDetectKeyPhrasesResult')
     ListOfDetectSentimentResult = Shapes::ListShape.new(name: 'ListOfDetectSentimentResult')
     ListOfDetectSyntaxResult = Shapes::ListShape.new(name: 'ListOfDetectSyntaxResult')
+    ListOfDocumentReadFeatureTypes = Shapes::ListShape.new(name: 'ListOfDocumentReadFeatureTypes')
     ListOfDominantLanguages = Shapes::ListShape.new(name: 'ListOfDominantLanguages')
     ListOfEntities = Shapes::ListShape.new(name: 'ListOfEntities')
     ListOfEntityLabels = Shapes::ListShape.new(name: 'ListOfEntityLabels')
@@ -300,6 +306,9 @@ module Aws::Comprehend
 
     AugmentedManifestsListItem.add_member(:s3_uri, Shapes::ShapeRef.new(shape: S3Uri, required: true, location_name: "S3Uri"))
     AugmentedManifestsListItem.add_member(:attribute_names, Shapes::ShapeRef.new(shape: AttributeNamesList, required: true, location_name: "AttributeNames"))
+    AugmentedManifestsListItem.add_member(:annotation_data_s3_uri, Shapes::ShapeRef.new(shape: S3Uri, location_name: "AnnotationDataS3Uri"))
+    AugmentedManifestsListItem.add_member(:source_documents_s3_uri, Shapes::ShapeRef.new(shape: S3Uri, location_name: "SourceDocumentsS3Uri"))
+    AugmentedManifestsListItem.add_member(:document_type, Shapes::ShapeRef.new(shape: AugmentedManifestsDocumentTypeFormat, location_name: "DocumentType"))
     AugmentedManifestsListItem.struct_class = Types::AugmentedManifestsListItem
 
     BatchDetectDominantLanguageItemResult.add_member(:index, Shapes::ShapeRef.new(shape: Integer, location_name: "Index"))
@@ -641,6 +650,11 @@ module Aws::Comprehend
     DocumentLabel.add_member(:score, Shapes::ShapeRef.new(shape: Float, location_name: "Score"))
     DocumentLabel.struct_class = Types::DocumentLabel
 
+    DocumentReaderConfig.add_member(:document_read_action, Shapes::ShapeRef.new(shape: DocumentReadAction, required: true, location_name: "DocumentReadAction"))
+    DocumentReaderConfig.add_member(:document_read_mode, Shapes::ShapeRef.new(shape: DocumentReadMode, location_name: "DocumentReadMode"))
+    DocumentReaderConfig.add_member(:feature_types, Shapes::ShapeRef.new(shape: ListOfDocumentReadFeatureTypes, location_name: "FeatureTypes"))
+    DocumentReaderConfig.struct_class = Types::DocumentReaderConfig
+
     DominantLanguage.add_member(:language_code, Shapes::ShapeRef.new(shape: String, location_name: "LanguageCode"))
     DominantLanguage.add_member(:score, Shapes::ShapeRef.new(shape: Float, location_name: "Score"))
     DominantLanguage.struct_class = Types::DominantLanguage
@@ -815,6 +829,7 @@ module Aws::Comprehend
 
     InputDataConfig.add_member(:s3_uri, Shapes::ShapeRef.new(shape: S3Uri, required: true, location_name: "S3Uri"))
     InputDataConfig.add_member(:input_format, Shapes::ShapeRef.new(shape: InputFormat, location_name: "InputFormat"))
+    InputDataConfig.add_member(:document_reader_config, Shapes::ShapeRef.new(shape: DocumentReaderConfig, location_name: "DocumentReaderConfig"))
     InputDataConfig.struct_class = Types::InputDataConfig
 
     InternalServerException.add_member(:message, Shapes::ShapeRef.new(shape: String, location_name: "Message"))
@@ -945,6 +960,8 @@ module Aws::Comprehend
 
     ListOfDetectSyntaxResult.member = Shapes::ShapeRef.new(shape: BatchDetectSyntaxItemResult)
 
+    ListOfDocumentReadFeatureTypes.member = Shapes::ShapeRef.new(shape: DocumentReadFeatureTypes)
+
     ListOfDominantLanguages.member = Shapes::ShapeRef.new(shape: DominantLanguage)
 
     ListOfEntities.member = Shapes::ShapeRef.new(shape: Entity)

data/lib/aws-sdk-comprehend/types.rb

@@ -20,6 +20,9 @@ module Aws::Comprehend
     #       {
     #         s3_uri: "S3Uri", # required
     #         attribute_names: ["AttributeNamesListItem"], # required
+    #         annotation_data_s3_uri: "S3Uri",
+    #         source_documents_s3_uri: "S3Uri",
+    #         document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #       }
     #
     # @!attribute [rw] s3_uri
@@ -42,11 +45,38 @@ module Aws::Comprehend
     #   job.
     #   @return [Array<String>]
     #
+    # @!attribute [rw] annotation_data_s3_uri
+    #   The S3 prefix to the annotation files that are referred in the
+    #   augmented manifest file.
+    #   @return [String]
+    #
+    # @!attribute [rw] source_documents_s3_uri
+    #   The S3 prefix to the source files (PDFs) that are referred to in the
+    #   augmented manifest file.
+    #   @return [String]
+    #
+    # @!attribute [rw] document_type
+    #   The type of augmented manifest. PlainTextDocument or
+    #   SemiStructuredDocument. If you don't specify, the default is
+    #   PlainTextDocument.
+    #
+    #   * `PLAIN_TEXT_DOCUMENT` A document type that represents any unicode
+    #     text that is encoded in UTF-8.
+    #
+    #   * `SEMI_STRUCTURED_DOCUMENT` A document type with positional and
+    #     structural context, like a PDF. For training with Amazon
+    #     Comprehend, only PDFs are supported. For inference, Amazon
+    #     Comprehend support PDFs, DOCX and TXT.
+    #   @return [String]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/comprehend-2017-11-27/AugmentedManifestsListItem AWS API Documentation
     #
     class AugmentedManifestsListItem < Struct.new(
       :s3_uri,
-      :attribute_names)
+      :attribute_names,
+      :annotation_data_s3_uri,
+      :source_documents_s3_uri,
+      :document_type)
       SENSITIVE = []
       include Aws::Structure
     end
@@ -692,6 +722,9 @@ module Aws::Comprehend
     #             {
     #               s3_uri: "S3Uri", # required
     #               attribute_names: ["AttributeNamesListItem"], # required
+    #               annotation_data_s3_uri: "S3Uri",
+    #               source_documents_s3_uri: "S3Uri",
+    #               document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #             },
     #           ],
     #         },
@@ -939,6 +972,9 @@ module Aws::Comprehend
     #             {
     #               s3_uri: "S3Uri", # required
     #               attribute_names: ["AttributeNamesListItem"], # required
+    #               annotation_data_s3_uri: "S3Uri",
+    #               source_documents_s3_uri: "S3Uri",
+    #               document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #             },
     #           ],
     #         },
@@ -1996,6 +2032,9 @@ module Aws::Comprehend
     #           {
     #             s3_uri: "S3Uri", # required
     #             attribute_names: ["AttributeNamesListItem"], # required
+    #             annotation_data_s3_uri: "S3Uri",
+    #             source_documents_s3_uri: "S3Uri",
+    #             document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #           },
     #         ],
     #       }
@@ -2264,6 +2303,52 @@ module Aws::Comprehend
       include Aws::Structure
     end
 
+    # The input properties for a topic detection job.
+    #
+    # @note When making an API call, you may pass DocumentReaderConfig
+    #   data as a hash:
+    #
+    #       {
+    #         document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #         document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #         feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #       }
+    #
+    # @!attribute [rw] document_read_action
+    #   This enum field will start with two values which will apply to PDFs:
+    #
+    #   * `TEXTRACT_DETECT_DOCUMENT_TEXT` - The service calls
+    #     DetectDocumentText for PDF documents per page.
+    #
+    #   * `TEXTRACT_ANALYZE_DOCUMENT` - The service calls AnalyzeDocument
+    #     for PDF documents per page.
+    #   @return [String]
+    #
+    # @!attribute [rw] document_read_mode
+    #   This enum field provides two values:
+    #
+    #   * `SERVICE_DEFAULT` - use service defaults for Document reading. For
+    #     Digital PDF it would mean using an internal parser instead of
+    #     Textract APIs
+    #
+    #   * `FORCE_DOCUMENT_READ_ACTION` - Always use specified action for
+    #     DocumentReadAction, including Digital PDF.
+    #   @return [String]
+    #
+    # @!attribute [rw] feature_types
+    #   Specifies how the text in an input file should be processed:
+    #   @return [Array<String>]
+    #
+    # @see http://docs.aws.amazon.com/goto/WebAPI/comprehend-2017-11-27/DocumentReaderConfig AWS API Documentation
+    #
+    class DocumentReaderConfig < Struct.new(
+      :document_read_action,
+      :document_read_mode,
+      :feature_types)
+      SENSITIVE = []
+      include Aws::Structure
+    end
+
     # Returns the code for the dominant language in the input text and the
     # level of confidence that Amazon Comprehend has in the accuracy of the
     # detection.
@@ -2931,6 +3016,9 @@ module Aws::Comprehend
     #           {
     #             s3_uri: "S3Uri", # required
     #             attribute_names: ["AttributeNamesListItem"], # required
+    #             annotation_data_s3_uri: "S3Uri",
+    #             source_documents_s3_uri: "S3Uri",
+    #             document_type: "PLAIN_TEXT_DOCUMENT", # accepts PLAIN_TEXT_DOCUMENT, SEMI_STRUCTURED_DOCUMENT
     #           },
     #         ],
     #       }
@@ -3368,7 +3456,7 @@ module Aws::Comprehend
       include Aws::Structure
     end
 
-    # The input properties for a topic detection job.
+    # The input properties for an inference job.
     #
     # @note When making an API call, you may pass InputDataConfig
     #   data as a hash:
@@ -3376,6 +3464,11 @@ module Aws::Comprehend
     #       {
     #         s3_uri: "S3Uri", # required
     #         input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #         document_reader_config: {
+    #           document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #           document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #           feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #         },
     #       }
     #
     # @!attribute [rw] s3_uri
@@ -3402,11 +3495,21 @@ module Aws::Comprehend
     #     documents, such as text messages.
     #   @return [String]
     #
+    # @!attribute [rw] document_reader_config
+    #   The document reader config field applies only for InputDataConfig of
+    #   StartEntitiesDetectionJob.
+    #
+    #   Use DocumentReaderConfig to provide specifications about how you
+    #   want your inference documents read. Currently it applies for PDF
+    #   documents in StartEntitiesDetectionJob custom inference.
+    #   @return [Types::DocumentReaderConfig]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/comprehend-2017-11-27/InputDataConfig AWS API Documentation
     #
     class InputDataConfig < Struct.new(
       :s3_uri,
-      :input_format)
+      :input_format,
+      :document_reader_config)
       SENSITIVE = []
       include Aws::Structure
     end
@@ -4877,6 +4980,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5026,6 +5134,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5173,6 +5286,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5342,6 +5460,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5455,6 +5578,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5609,6 +5737,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5737,6 +5870,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required
@@ -5891,6 +6029,11 @@ module Aws::Comprehend
     #         input_data_config: { # required
     #           s3_uri: "S3Uri", # required
     #           input_format: "ONE_DOC_PER_FILE", # accepts ONE_DOC_PER_FILE, ONE_DOC_PER_LINE
+    #           document_reader_config: {
+    #             document_read_action: "TEXTRACT_DETECT_DOCUMENT_TEXT", # required, accepts TEXTRACT_DETECT_DOCUMENT_TEXT, TEXTRACT_ANALYZE_DOCUMENT
+    #             document_read_mode: "SERVICE_DEFAULT", # accepts SERVICE_DEFAULT, FORCE_DOCUMENT_READ_ACTION
+    #             feature_types: ["TABLES"], # accepts TABLES, FORMS
+    #           },
     #         },
     #         output_data_config: { # required
     #           s3_uri: "S3Uri", # required

data/lib/aws-sdk-comprehend.rb

@@ -48,6 +48,6 @@ require_relative 'aws-sdk-comprehend/customizations'
 # @!group service
 module Aws::Comprehend
 
-  GEM_VERSION = '1.50.0'
+  GEM_VERSION = '1.51.0'
 
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: aws-sdk-comprehend
 version: !ruby/object:Gem::Version
-  version: 1.50.0
+  version: 1.51.0
 platform: ruby
 authors:
 - Amazon Web Services
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-09-01 00:00:00.000000000 Z
+date: 2021-09-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aws-sdk-core