From 5f76e7e8fcb68631e141c6cc1a221f2201c9328d Mon Sep 17 00:00:00 2001 From: Jonathan van Westendorp Date: Thu, 9 Jul 2026 14:21:43 +0200 Subject: [PATCH 1/3] discoveryengine: add enable_llm_layout_parsing and enable_get_processed_document to data store --- mmv1/products/discoveryengine/DataStore.yaml | 20 +++++++++++++++++++ ...ment_processing_config_layout_full.tf.tmpl | 14 +++++++------ 2 files changed, 28 insertions(+), 6 deletions(-) diff --git a/mmv1/products/discoveryengine/DataStore.yaml b/mmv1/products/discoveryengine/DataStore.yaml index 61b868d539ab..1f8124281d8e 100644 --- a/mmv1/products/discoveryengine/DataStore.yaml +++ b/mmv1/products/discoveryengine/DataStore.yaml @@ -313,6 +313,16 @@ properties: description: | If true, the LLM based annotation is added to the image during parsing. required: false + - name: 'enableLlmLayoutParsing' + type: Boolean + description: | + If true, the pdf layout will be refined using an LLM. + required: false + - name: 'enableGetProcessedDocument' + type: Boolean + description: | + If true, the processed document will be made available for the GetProcessedDocument API. + required: false - name: 'structuredContentTypes' type: Array description: | @@ -401,6 +411,16 @@ properties: description: | If true, the LLM based annotation is added to the image during parsing. required: false + - name: 'enableLlmLayoutParsing' + type: Boolean + description: | + If true, the pdf layout will be refined using an LLM. + required: false + - name: 'enableGetProcessedDocument' + type: Boolean + description: | + If true, the processed document will be made available for the GetProcessedDocument API. + required: false - name: 'structuredContentTypes' type: Array description: | diff --git a/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl b/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl index b114fcb9544b..4480b17dbdc9 100644 --- a/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl +++ b/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl @@ -9,12 +9,14 @@ resource "google_discovery_engine_data_store" "document_processing_config_layout document_processing_config { default_parsing_config { layout_parsing_config { - enable_table_annotation = true - enable_image_annotation = true - structured_content_types = ["shareholder-structure"] - exclude_html_elements = ["nav", "footer"] - exclude_html_classes = ["overlay", "screenreader"] - exclude_html_ids = ["cookie-banner"] + enable_table_annotation = true + enable_image_annotation = true + enable_llm_layout_parsing = true + enable_get_processed_document = true + structured_content_types = ["shareholder-structure"] + exclude_html_elements = ["nav", "footer"] + exclude_html_classes = ["overlay", "screenreader"] + exclude_html_ids = ["cookie-banner"] } } chunking_config { From d4c7c6586c3d821ec3fa43e86311b4c74296a709 Mon Sep 17 00:00:00 2001 From: Jonathan van Westendorp Date: Fri, 10 Jul 2026 09:08:54 +0200 Subject: [PATCH 2/3] discoveryengine: cover layout_parsing_config override fields in example --- ...astore_document_processing_config_layout_full.tf.tmpl | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl b/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl index 4480b17dbdc9..e56b5446e25f 100644 --- a/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl +++ b/mmv1/templates/terraform/samples/services/discoveryengine/discoveryengine_datastore_document_processing_config_layout_full.tf.tmpl @@ -25,5 +25,14 @@ resource "google_discovery_engine_data_store" "document_processing_config_layout include_ancestor_headings = true } } + parsing_config_overrides { + file_type = "pdf" + layout_parsing_config { + enable_table_annotation = true + enable_image_annotation = true + enable_llm_layout_parsing = true + enable_get_processed_document = true + } + } } } From bd63e46e43e30b524ce5368b776a8677db741f48 Mon Sep 17 00:00:00 2001 From: Jonathan van Westendorp Date: Fri, 10 Jul 2026 20:57:17 +0200 Subject: [PATCH 3/3] drop redundant required:false on new layout parsing fields --- mmv1/products/discoveryengine/DataStore.yaml | 4 ---- 1 file changed, 4 deletions(-) diff --git a/mmv1/products/discoveryengine/DataStore.yaml b/mmv1/products/discoveryengine/DataStore.yaml index 1f8124281d8e..700f768d8328 100644 --- a/mmv1/products/discoveryengine/DataStore.yaml +++ b/mmv1/products/discoveryengine/DataStore.yaml @@ -317,12 +317,10 @@ properties: type: Boolean description: | If true, the pdf layout will be refined using an LLM. - required: false - name: 'enableGetProcessedDocument' type: Boolean description: | If true, the processed document will be made available for the GetProcessedDocument API. - required: false - name: 'structuredContentTypes' type: Array description: | @@ -415,12 +413,10 @@ properties: type: Boolean description: | If true, the pdf layout will be refined using an LLM. - required: false - name: 'enableGetProcessedDocument' type: Boolean description: | If true, the processed document will be made available for the GetProcessedDocument API. - required: false - name: 'structuredContentTypes' type: Array description: |