Using tencentdb_ai Plug-In to Call a Large Model
Last updated: 2025-04-30 18:12:05
Currently, both basic large models and AI application development are developing rapidly. TencentDB for PostgreSQL provides the tencentdb_ai plug-in to facilitate your easy calling of accessible large model APIs in TencentDB for PostgreSQL instances and complete application development in various scenarios. This plug-in supports instances with a major version of PostgreSQL 17.
Note:
Using tencentdb_ai requires that you have already created an instance with a major version of PostgreSQL 17.Creating the
tencentdb_ai plug-in will associate with the creation of the pgcrypto plug-in. Please be informed.tencentdb_ai Plug-In Function Description
Adding a Model
The function definition of
add_model is as follows:tencentdb_ai.add_model(model_name NAME,version TEXT,region TEXT,json_path JSONPATH);
The parameter interpretations are as follows:
model name: model name, such as
hunyuan-lite, deepseek-r1.version: common parameters of this model. If this model does not require the
version parameter, it can be left blank.region: the common parameter
region of this model. If this model does not require the region parameter, it can be left blank.json_path: Specifies how to extract the
JSON response of this model when calling it. Equivalent to calling the jsonb_path_query_first function on the model's output. If this parameter is not set, the original response of the model will be output as is.View Registered Models
The function definition of
list_models is as follows:tencentdb_ai.list_models(void);
Updating Model
The function definition of
update_model_attr is as follows:tencentdb_ai.update_model_attr(model_name NAME,attr_name text,attr_value text);
The parameter interpretations are as follows:
model name: model name.
attr_name: model attribute that needs to be updated. The updatable model attributes are
version, region, json_path, SecretId, SecretKey.attr_value: value of the attribute that needs to be updated.
Deleting a Model
The function definition of
delete_model is as follows:tencentdb_ai.delete_model(model_name NAME);
The parameter interpretations are as follows:
model name: model name.
Model Invocation
The call_model function is defined as follows:
tencentdb_ai.call_model(model_name NAME,common_params TEXT[],api_params TEXT[]);
The parameter interpretations are as follows:
model name: model name.
common_params: common parameters needed to call the model, such as
ARRAY['Action: ChatCompletions', 'Version: 2023-09-01']api_params: proprietary parameters required to invoke the model, for example
ARRAY['"Stream": false', '"Model": "hunyuan-lite"', '"Messages": [{"Role": "user", "Content": "Hello"}]']Besides using the
call_model function to proceed with model invocation, you can also use the following APIs to perform model invocation in specific scenarios.Fixed Scenario Interface
Dialogue
chat_completions function definition is as follows:tencentdb_ai.chat_completions(model_name NAME,content TEXT,args TEXT[] default NULL);
The parameter interpretations are as follows:
model name: model name.
content: conversation content.
args: other parameters for invoking the model, default value is NULL.
Call example:
SELECT tencentdb_ai.chat_completions('hunyuan-lite', 'Hello');chat_completions----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------{"Response":{"RequestId":"*******-****-****-****-***********","Note":"The above content is AI generation. It does not represent the developer's standpoint. Do not delete or modify this tag.","Choices":[{"Index":0,"Message":{"Role":"assistant","Co"}}]}}Hello! Glad to chat with you. What would you like to share with me today?nTokens":17,"TotalTokens":20}}}(1 row)
Text to Vector
The
get_embedding function is defined as follows:tencentdb_ai.get_embedding(model_name NAME,content TEXT[]);
The parameter interpretations are as follows:
model name: model name.
content: transformed content.
Call example:
damoxing1 => SELECT tencentdb_ai.get_embedding('hunyuan-embedding',ARRAY['Hello', 'PostgreSQL']);
Text Sorting
The
run_rerank function is defined as follows:tencentdb_ai.run_rerank(model_name NAME,query TEXT,documents TEXT[],args TEXT[] DEFAULT NULL);
The parameter interpretations are as follows:
model name: model name.
query: text sorting problem.
documents: content to be textually sorted.
args: specified parameters for a specific model.
damoxing1=> SELECT COUNT(*) FROM tencentdb_ai.run_rerank('lke-reranker-base', 'large model of knowledge engine', ARRAY['HunYuan Large Model', 'Tencent knowledge engine']);count-------2(1 row)
Plugin Usage Example
The following describes the plugin usage process one by one:
1. Create a plugin
damoxing=> CREATE EXTENSION tencentdb_ai CASCADE;NOTICE: installing required extension "pgcrypto"CREATE EXTENSION
Check plug-in creation completed
damoxing=> SELECT * FROM pg_extension;oid | extname | extowner | extnamespace | extrelocatable | extversion | extconfig | extcondition-------+--------------+----------+--------------+----------------+------------+-----------+--------------14275 | plpgsql | 10 | 11 | f | 1.0 | |16440 | pgcrypto | 16437 | 2200 | t | 1.3 | |16477 | tencentdb_ai | 16437 | 2200 | t | 1.0 | {16479} | {""}(3 rows)
2. Add a large model
damoxing=> SELECT tencentdb_ai.add_model('hunyuan-lite','2023-09-01', NULL, NULL);add_model-----------(1 row)
3. Add SecretId and SecretKey for large model API calls
damoxing=> SELECT tencentdb_ai.update_model_attr('hunyuan-lite', 'SecretId', 'AKID***************');update_model_attr-------------------(1 row)damoxing=> SELECT tencentdb_ai.update_model_attr('hunyuan-lite', 'SecretKey', '********************');update_model_attr-------------------(1 row)
4. View the currently added model list
damoxing=> SELECT * FROM tencentdb_ai.list_models();-[ RECORD 1 ]--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------model_name | hunyuan-litejson_path |secretid | ***************************secretkey | ***************************version |region |id_random | 516234453022key_random | 134577899689
5. Call a large model
In a chat scenario, you can use the
call_model API.damoxing=> SELECT tencentdb_ai.call_model('hunyuan-lite',ARRAY['Action: ChatCompletions', 'Version: 2023-09-01'],ARRAY['"Stream": false', '"Model": "hunyuan-lite"', '"Messages": [{"Role": "user", "Content": "Hello"}]']);damoxing(> damoxing(>call_model------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------{"Response":{"RequestId":"*******-****-****-****-***********","Note":"The above content is AI generation. It does not represent the developer's standpoint. Do not delete or modify this tag.","ChVoices":[{"Index":0,"Message":{"Role":"assistant","Content":"Hello! Glad to talk with you. What can I help you with? Whether it's about life, work, or study.For any problems, whether related to Xi or other aspects, I will do my best to offer you help.","Usage":{"PromptTokens":3,"CompletionTokens":33,"TotalTokens":36}}}(1 row)damoxing=>
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback