HDDS-15468. Replace langchain4j-bom, guice-bom due to leaked dependencies

[adoroszlai]

What changes were proposed in this pull request?

langchain4j-bom and guice-bom leak transitive dependencies defined in their parent POMs.

• https://github.com/langchain4j/langchain4j/blob/0.35.0/langchain4j-bom/pom.xml#L8
• https://github.com/langchain4j/langchain4j/blob/0.35.0/langchain4j-parent/pom.xml#L92-L126
• https://github.com/langchain4j/langchain4j/blob/0.35.0/langchain4j-parent/pom.xml#L158-L433
• https://github.com/google/guice/blob/6.0.0/bom/pom.xml#L8
• https://github.com/google/guice/blob/6.0.0/pom.xml#L127-L199

This overrides some of Ozone's own dependencies. Example for production dependency:

$ cd hadoop-ozone/dist/target/ozone-2.2.0-SNAPSHOT/share/ozone/lib
$ ls slf4j*
slf4j-api-2.0.7.jar
slf4j-reload4j-2.0.17.jar

Some test artifacts are also overridden, which was previously worked around by adding these dependencies again, despite having been imported from their own BOMs:

ozone/pom.xml

Line 1613 in 891cf0e

<!-- langchain4j-bom pins JUnit 5.10.0; override to keep Ozone on ${junit5.version}. -->

ozone/pom.xml

Line 1679 in 891cf0e

<!-- langchain4j-bom pins AWS SDK 2.21.44; override to keep Ozone on ${aws-java-sdk2.version}. -->

This change replaces usage of these two BOMs with individual artifacts.

https://issues.apache.org/jira/browse/HDDS-15468

How was this patch tested?

Verified slf4j-api version after build.

$ cd hadoop-ozone/dist/target/ozone-2.2.0-SNAPSHOT/share/ozone/lib
$ ls slf4j*
slf4j-api-2.0.17.jar
slf4j-reload4j-2.0.17.jar

Compared effective POM before/after the change. Effective POM (with BOMs imported, properties substituted etc.) can be generated as:

mvn -N help:effective-pom -Doutput=effective.xml

https://github.com/adoroszlai/ozone/actions/runs/26864255402

[ArafatKhan2198]

@adoroszlai Thanks for the fix! I pulled down your branch and ran a live test against the Recon chatbot to make sure removing the BOM didn't break the LLM integration. I generated some data using Freon and ran a few queries. Everything is resolving correctly and the chatbot works perfectly.

Here are the responses from the test:

Query: What is the cluster state?

The Ozone cluster is currently in a **healthy state** with **one datanode** and **one pipeline** operational.

Here's a summary of the cluster's state:

-   **Datanodes:**
    -   Total Datanodes: **1**
    -   Healthy Datanodes: **1**
-   **Pipelines:** **1**
-   **Containers:**
    -   Total Containers: **1**
    -   Open Containers: **1**
    -   Missing Containers: **0**
    -   Deleted Containers: **0**
-   **Storage Report:**
    -   Capacity: **149.27 GB** (149,266,245,387 bytes)
    -   Used: **4.38 MB** (4,382,720 bytes)
    -   Remaining: **42.25 GB** (42,251,796,480 bytes)
    -   Filesystem Capacity: **149.28 GB** (149,281,173,504 bytes)
    -   Filesystem Used: **107.03 GB** (107,029,377,024 bytes)
    -   Filesystem Available: **42.25 GB** (42,251,796,480 bytes)
    -   Minimum Free Space: **100 MB** (104,857,600 bytes)
-   **Volumes:** **2**
-   **Buckets:** **1**
-   **Keys:** **0**
-   **Keys Pending Deletion:** **0**

Query: How many datanodes are healthy?

Based on the cluster state, there is **1 healthy datanode**.

Query: Show me all pipelines

There is **1 pipeline** currently in your Ozone cluster.

Here are the details for the pipeline:

- **Pipeline ID**: `71eb2a50-3acd-4a47-acfa-1483d5ad289a`
- **Status**: `OPEN`
- **Leader Node**: `ozone-datanode-1.ozone_default`
- **Replication Type**: `RATIS`
- **Replication Factor**: `ONE`
- **Datanodes in Pipeline**:
  - `ozone-datanode-1.ozone_default` (IP: `172.29.0.4`, UUID: `2a15fd8d-8303-4768-a2f8-13332bb1be71`)
- **Containers**: 1

[ArafatKhan2198]

LGTM +1

[adoroszlai]

Thanks @ArafatKhan2198 for the review.