Add JdbcChatMemory #1528

leijendary · 2024-10-11T21:51:32Z

An option to use a generic JdbcChatMemory with a starter for users that want to use SQL as chat memory.

...ngframework/ai/autoconfigure/chat/memory/pgvector/PgVectorChatMemoryAutoConfigurationIT.java

...ector-store/src/test/java/org/springframework/ai/chat/memory/PgVectorChatMemoryConfigIT.java

...ai-pgvector-store/src/test/java/org/springframework/ai/chat/memory/PgVectorChatMemoryIT.java

ogbozoyan · 2024-11-23T13:05:35Z

Is it will be usable with VectorStoreChayMemoryAdvisor? Or what is the point ?

leijendary · 2024-11-23T15:00:27Z

@ogbozoyan VectorStoreChatMemoryAdvisor is optional if we have this change. PgVectorChatMemory is an implementation of ChatMemory and the postgres version of CassandraChatMemory which is specialized just for the chat memory/history.

ogbozoyan · 2024-11-23T15:11:03Z

@leijendary which advisor you should use with that ChatMemory to process saving ?

leijendary · 2024-11-23T15:15:14Z

@ogbozoyan You can use MessageChatMemoryAdvisor and pass in the instance of the PgVectorChatMemory. Then for RAG, you can use RetrievalAugmentationAdvisor. This is what I did for my chatClient bean:

@Bean
fun chatClient(builder: ChatClient.Builder, vectorStore: VectorStore, chatMemory: ChatMemory): ChatClient {
    val memoryAdvisor = MessageChatMemoryAdvisor(chatMemory)
    val documentRetriever = VectorStoreDocumentRetriever.builder()
        .vectorStore(vectorStore)
        .build()
    val ragAdvisor = RetrievalAugmentationAdvisor.builder()
        .documentRetriever(documentRetriever)
        .build()

    return builder
        .defaultAdvisors(memoryAdvisor, ragAdvisor)
        .build()
}

ogbozoyan · 2024-11-23T15:20:29Z

@ogbozoyan You can use MessageChatMemoryAdvisor and pass in the instance of the PgVectorChatMemory. Then for RAG, you can use RetrievalAugmentationAdvisor. This is what I did for my chatClient bean:

@Bean
fun chatClient(builder: ChatClient.Builder, vectorStore: VectorStore, chatMemory: ChatMemory): ChatClient {
    val memoryAdvisor = MessageChatMemoryAdvisor(chatMemory)
    val documentRetriever = VectorStoreDocumentRetriever.builder()
        .vectorStore(vectorStore)
        .build()
    val ragAdvisor = RetrievalAugmentationAdvisor.builder()
        .documentRetriever(documentRetriever)
        .build()

    return builder
        .defaultAdvisors(memoryAdvisor, ragAdvisor)
        .build()
}

in this example you passing VectorStore, may be you meant ?

fun chatClient(builder: ChatClient.Builder, **pgVectorChatMemory: PgVectorChatMemory**, chatMemory: ChatMemory): ChatClient {
     val memoryAdvisor = MessageChatMemoryAdvisor(**pgVectorChatMemory**)
     val documentRetriever = VectorStoreDocumentRetriever.builder()
         .vectorStore(vectorStore)
         .build()
     val ragAdvisor = RetrievalAugmentationAdvisor.builder()
         .documentRetriever(documentRetriever)
         .build()
 
     return builder
         .defaultAdvisors(memoryAdvisor, ragAdvisor)
         .build()
}

leijendary · 2024-11-23T15:24:33Z

@ogbozoyan

@ogbozoyan You can use MessageChatMemoryAdvisor and pass in the instance of the PgVectorChatMemory. Then for RAG, you can use RetrievalAugmentationAdvisor. This is what I did for my chatClient bean:

@Bean
fun chatClient(builder: ChatClient.Builder, vectorStore: VectorStore, chatMemory: ChatMemory): ChatClient {
    val memoryAdvisor = MessageChatMemoryAdvisor(chatMemory)
    val documentRetriever = VectorStoreDocumentRetriever.builder()
        .vectorStore(vectorStore)
        .build()
    val ragAdvisor = RetrievalAugmentationAdvisor.builder()
        .documentRetriever(documentRetriever)
        .build()

    return builder
        .defaultAdvisors(memoryAdvisor, ragAdvisor)
        .build()
}

in this example you passing VectorStore, may be you meant ?

fun chatClient(builder: ChatClient.Builder, **pgVectorChatMemory: PgVectorChatMemory**, chatMemory: ChatMemory): ChatClient {
     val memoryAdvisor = MessageChatMemoryAdvisor(**pgVectorChatMemory**)
     val documentRetriever = VectorStoreDocumentRetriever.builder()
         .vectorStore(vectorStore)
         .build()
     val ragAdvisor = RetrievalAugmentationAdvisor.builder()
         .documentRetriever(documentRetriever)
         .build()
 
     return builder
         .defaultAdvisors(memoryAdvisor, ragAdvisor)
         .build()
}

The code I provided is correct. I am using it right now. The ChatMemory instance is automatically picked up by spring as PgVectorChatMemory because of PgVectorChatMemoryAutoConfiguration

eddumelendez · 2024-11-29T20:45:40Z

I wonder if this should be more generic and instead providing a JdbcChatMemory instead. I would be also nice to provide the proper scripts to create the tables, similar to what spring security and spring session does.

leijendary · 2024-11-29T21:17:45Z

@eddumelendez I agree. I also originally designed it to be the same as the existing CassandraChatMemory and just used the same columns. I thought of changing the columns to the following:

CREATE TABLE ai_chat_memory (
    session_id character varying(36) NOT NULL,
    content text NOT NULL,
    type character varying(10) NOT NULL, -- USER/ASSISTANT
    "timestamp" timestamp without time zone NOT NULL DEFAULT NOW()
);

Do you also mean that the table should not be automatically created and instead be created manually by the developer?

eddumelendez · 2024-11-29T22:07:53Z

Do you also mean that the table should not be automatically created and instead be created manually by the developer?

it should be created automatically by detecting the database. I meant having something like https://github.com/spring-projects/spring-session/blob/main/spring-session-jdbc/src/main/resources/org/springframework/session/jdbc/schema-postgresql.sql that spring ai can execute, so, if new database should be supported then adding a script should be enough

ogbozoyan · 2024-11-30T09:41:26Z

type character varying(10) NOT NULL, -- USER/ASSISTANT

maybe add check ?

CREATE TABLE ai_chat_memory (
    session_id character varying(36) NOT NULL,
    content text NOT NULL,
    type character varying(10) NOT NULL CHECK (type IN ('USER', 'ASSISTANT')), -- USER/ASSISTANT
    "timestamp" timestamp without time zone NOT NULL DEFAULT NOW()
);

sobychacko · 2025-03-26T22:35:37Z

@leijendary @ogbozoyan We are reviewing this PR currently. Could you rebase the changes against the latest main branch? Please keep in mind that we refactored auto-configuration and starters in a significant manner recently. See this blog for more details: https://spring.io/blog/2025/03/26/spring-ai-update-to-snapshots. The previous monolithic auto-configuration module has been removed in favor of dedicated auto-config modules. Can you adjust your PR for that? Also, I couldn't run your tests in the chat-memory module. I didn't try the autoconfig tests yet. Thanks!

leijendary · 2025-03-28T13:41:28Z

@sobychacko I pushed the changes.

sobychacko · 2025-04-02T22:04:16Z

...at-memory/src/main/resources/org/springframework/ai/chat/memory/jdbc/schema-drop-mariadb.sql

@@ -0,0 +1 @@
+DROP TABLE IF EXISTS ai_chat_memory;


Are these SQL scripts necessary for the JDBC chat memory? I see they are used in the integration test, but it's unclear how a user would use it. The docs below don't mention them either. Could you clarify? At the very least, we need to inform the end user of the existence of these scripts. I see some of these scripts are used from the auto-configuration, but how would an end user that is not using auto configuration makes use of these scripts?

I updated the docs and specified that the chat memory table will be automatically created using autoconfiguration and how to disable it. I also removed the sql files that are not used.

sobychacko · 2025-04-02T22:04:31Z

...bc-chat-memory/src/main/resources/org/springframework/ai/chat/memory/jdbc/schema-mariadb.sql

+);
+
+CREATE INDEX IF NOT EXISTS ai_chat_memory_conversation_id_timestamp_idx
+ON ai_chat_memory(conversation_id, `timestamp` DESC);


Same comment as above.

I updated the docs and specified that the chat memory table will be automatically created using autoconfiguration and how to disable it. I also removed the sql files that are not used.

sobychacko · 2025-04-02T22:04:38Z

...chat-memory/src/main/resources/org/springframework/ai/chat/memory/jdbc/schema-postgresql.sql

+);
+
+CREATE INDEX IF NOT EXISTS ai_chat_memory_conversation_id_timestamp_idx
+ON ai_chat_memory(conversation_id, "timestamp" DESC);


Same comment as above.

I updated the docs and specified that the chat memory table will be automatically created using autoconfiguration and how to disable it. I also removed the sql files that are not used.

sobychacko · 2025-04-02T22:04:49Z

...memory/src/main/resources/org/springframework/ai/chat/memory/jdbc/schema-drop-postgresql.sql

@@ -0,0 +1 @@
+DROP TABLE IF EXISTS ai_chat_memory;


Same comment as above.

This file is now removed. See comment above.

sobychacko · 2025-04-05T17:58:31Z

@leijendary I see that you are making some updates to the PR. I was in the process of making your PR conformant with the new structure we introduced in the latest main. Since you are already making changes, could you restructure your PR to be consistent with the latest main? For e.g. move the jdbc chat memory module under the top-level memory folder and rename the module to reflect the new structure etc. Please ping here if you can do that. Otherwise, we can take care of it on the merge. (If you end up making the changes, I would suggest squashing all your commits into a single one after rebasing to the latest main). Thanks!

Signed-off-by: leijendary <jonathanleijendekker@gmail.com>

leijendary · 2025-04-06T19:33:12Z

@sobychacko I took care of it and migrated the folder structure based on the latest structure in main. Feel free to update it if I missed anything.

sobychacko · 2025-04-07T19:14:09Z

@leijendary Thank you for the PR. It is now merged upstream via 4be1002.

leijendary changed the title ~~Added ChatMemory implementation for PgVector~~ Add ChatMemory implementation for PgVector Oct 11, 2024

leijendary changed the title ~~Add ChatMemory implementation for PgVector~~ Add PgVectorChatMemory Oct 11, 2024

eddumelendez reviewed Oct 12, 2024

View reviewed changes

tzolov added the Chat Memory label Oct 12, 2024

ogbozoyan mentioned this pull request Nov 24, 2024

Rag Advanced example ThomasVitale/llm-apps-java-spring-ai#7

Open

leijendary changed the title ~~Add PgVectorChatMemory~~ Add JdbcChatMemory Dec 13, 2024

leijendary requested a review from eddumelendez December 13, 2024 23:51

leijendary force-pushed the main branch 7 times, most recently from ac555b9 to af61678 Compare January 25, 2025 10:27

markpollack assigned sobychacko Mar 25, 2025

markpollack added this to the 1.0.0-M7 milestone Mar 25, 2025

leijendary closed this Mar 28, 2025

leijendary force-pushed the main branch from 44c0786 to b3cfa2b Compare March 28, 2025 13:33

leijendary reopened this Mar 28, 2025

sobychacko reviewed Apr 2, 2025

View reviewed changes

leijendary force-pushed the main branch from 41ee3c1 to 6bb4617 Compare April 6, 2025 19:24

feat: JDBC implementation of ChatMemory

11b8b8a

Signed-off-by: leijendary <jonathanleijendekker@gmail.com>

leijendary force-pushed the main branch from 6bb4617 to 11b8b8a Compare April 6, 2025 19:27

sobychacko closed this Apr 7, 2025

Add JdbcChatMemory #1528

Add JdbcChatMemory #1528

Uh oh!

Conversation

leijendary commented Oct 11, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ogbozoyan commented Nov 23, 2024

Uh oh!

leijendary commented Nov 23, 2024

Uh oh!

ogbozoyan commented Nov 23, 2024

Uh oh!

leijendary commented Nov 23, 2024

Uh oh!

ogbozoyan commented Nov 23, 2024

Uh oh!

leijendary commented Nov 23, 2024

Uh oh!

eddumelendez commented Nov 29, 2024

Uh oh!

leijendary commented Nov 29, 2024

Uh oh!

eddumelendez commented Nov 29, 2024

Uh oh!

ogbozoyan commented Nov 30, 2024

Uh oh!

sobychacko commented Mar 26, 2025

Uh oh!

leijendary commented Mar 28, 2025

Uh oh!

sobychacko Apr 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

leijendary Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

sobychacko Apr 2, 2025

Choose a reason for hiding this comment

Uh oh!

leijendary Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

sobychacko Apr 2, 2025

Choose a reason for hiding this comment

Uh oh!

leijendary Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

sobychacko Apr 2, 2025

Choose a reason for hiding this comment

Uh oh!

leijendary Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

sobychacko commented Apr 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

leijendary commented Apr 6, 2025

Uh oh!

sobychacko commented Apr 7, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

leijendary commented Oct 11, 2024 •

edited

Loading

sobychacko Apr 2, 2025 •

edited

Loading

sobychacko commented Apr 5, 2025 •

edited

Loading