How to configure the Content Communities industry template-OpenSearch(Open Search)-阿里云帮助中心

Procedure

1. In the OpenSearch Industry Algorithm Edition console, navigate to the Instance Management page and click Create Instance.

2. On the purchase page, set Edition to Industry Algorithm Edition and Industry Type to Content Communities. For Cluster Preferences, select a dedicated cluster type (such as General-purpose, Compute-optimized, or Storage-optimized) based on your business needs. Configure the other parameters as you would for the General-purpose Edition.

3. On the order confirmation page, review your configuration, including the region, application name, application type, edition, industry type, purchase specifications, storage capacity, and computing resources. After verifying that the configuration is correct, select the OpenSearch Pay-As-You-Go Service Agreement checkbox, and click Activate Now.

4. After the instance is created, go to the Instance Management page, find your application in the list, and click Configure.

The configuration wizard guides you through the following steps: Feature Selection > Application Schema > Index Schema > Data Source > Complete.

(1) Feature Selection: For Selected Vertical, select IT. Select the template features as needed (all are selected by default). When finished, click Next.

Template features include Query Analysis (such as Term Weight Analysis Dictionary for IT Content, Synonym Dictionary for IT Content, and Text Vectorization), Sort Policy (such as multimodal search, Text Relevance, and Vector Relevance), and Drop-down Suggestions.

(2) Application Schema: At the top of the page, review the available field tags: Article Title, Article Details, Article Author, Article Summary, Reply Count, Article Tags, and Creation/Modification Time. In the table below, assign the correct field tag and data type to each field. When finished, click Next.

Note: The system uses field tags to automatically generate the industry-enhanced Index Schema, query analysis rules, and sort policies. You must assign a field tag to its corresponding field in the Application Schema.

(3) Index Schema: On the Index Schema configuration page, add an index tag and set the analysis method for each index field. When finished, click Next.

Note: The system uses index tags to automatically generate query analysis rules and sort policies. You must assign an index tag to its corresponding index in the Index Schema. Follow these rules:

1. Fields that you tagged as Article Details or Article Title in the Application Schema must be configured as a composite index. You must also apply the Default Index index tag and set the Analysis Method to Industry - Content IT Analysis.

2. Fields that you tagged as Article Title in the Application Schema must be assigned the Vector Index index tag, and their Analysis Method must be set to Vector - Content IT.

(4) Data Source: OpenSearch supports several cloud data sources, including RDS, MaxCompute, and PolarDB.

Note: You can also upload data to OpenSearch using the API.

(5) Complete:

A message confirms that the Industry Algorithm Edition application is configured. You can now perform the following actions: 1. Configure query analysis using the feature package from your selected industry template. 2. Configure a sort policy using the selected sort policy models. 3. Configure and use the Drop-down Suggestions feature.

Query test

Search test

Once the indexes are built, go to the Feature Extensions > Search Test page to test the index that uses the text analyzer from the Content Communities industry template:

For example, to test the default index, enter the following query clause: default:'company'

Vector search

To test the vector search feature, follow these steps:

1. Go to the Search Algorithm Center > Retrieval Configuration > Query Analysis Configuration page. In the list, find the sys_default query analysis rule, which is generated automatically by the industry template. Click sys_default to open the rule editor. Confirm that Industry Type is set to Content Communities and Selected Vertical is set to IT. In the feature description section, select the Text Vectorization checkbox, select vector for Vector Index, and click OK.

2. Go to the Feature Extensions > Search Test page and add the following required parameters to run a test:

Required parameters:
query=default:'company'
qp=sys_default
raw_query=company

You can view the vector result in the query analysis details. In the search results, click the Query Analysis Process link. A dialog box opens, showing each analysis step, such as spell check, normalization, tokenization, term weight analysis, and text vectorization. The text vectorization step shows the vector index is vector and the vector dimension is 128.

Note

The vector index takes effect only after you enable the Text Vectorization feature in the query analysis configuration.
The index on which you configure query analysis must use the Industry - Content IT Analysis analyzer, and its industry type must be Content Communities. For the Text Vectorization feature, the associated vector index field must use the Vector - Content IT analyzer.

Multimodal retrieval

The Industry Algorithm Edition for Content Communities automatically generates a multimodal search policy named sys_strategy for your instance. To test this feature, go to the Feature Extensions > Search Test page and add the following parameters:

Required parameters:
query=default:'company'
qp=sys_default
raw_query=company
search_strategy=sys_strategy

Usage notes

You can upgrade an application from General-purpose Edition to Industry Algorithm Edition, but you cannot downgrade an application from Industry Algorithm Edition to General-purpose Edition.
The Vector - Content IT analyzer cannot be used independently. It must be used with an index that uses the Industry - Content IT Analysis analyzer and has the Text Vectorization feature enabled in its query analysis configuration.
The Industry Algorithm Edition for Content Communities is available only for applications that use a dedicated cluster.
To have the system automatically generate vectors for a field, set the field data type to TEXT and the analysis method to Vector - Content IT in the index configuration.
In the Application Schema, you must assign the correct field tag to each field. Otherwise, an error occurs.
In the Index Schema, you must assign the correct index tag to each index. Otherwise, you cannot proceed to the next step.