Merge branch 'release/1.3' of github.com:NVIDIA/NeMo-Agent-Toolkit into david-fe-simple-calc-custom-e2e

Signed-off-by: David Gardner <[email protected]>

Author: dagardner-nv

.gitlab-ci.yml

@@ -42,6 +42,7 @@ variables:
   NAT_CI_ETCD_HOST: "etcd"
   NAT_CI_MILVUS_HOST: "milvus"
   NAT_CI_MYSQL_HOST: "mysql"
+  NAT_CI_OPENSEARCH_URL: "http://opensearch:9200"
   NAT_CI_REDIS_HOST: "redis"
   NAT_CI_S3_HOST: "minio"
   UV_CACHE_DIR: .uv-cache
@@ -94,6 +95,11 @@ test:python_tests:
         ETCD_ENDPOINTS: etcd:2379
         MINIO_ADDRESS: minio:9000
       command: ["milvus", "run", "standalone"]
+    - name: opensearchproject/opensearch:2.11.1
+      alias: opensearch
+      variables:
+        discovery.type: "single-node"
+        plugins.security.disabled: "true"
 
   script:
     - echo "Running tests"

docs/source/workflows/llms/using-local-llms.md

@@ -32,10 +32,11 @@ Regardless of the model you choose, the process is the same for downloading the
 
 ### Install the Simple Web Query Example
 
-First, ensure the current working directory is the root of the NeMo Agent toolkit repository. Then, install the simple web query example so we have the `webpage_query` tool available.
+First, ensure the current working directory is the root of the NeMo Agent toolkit repository. Then, install NAT and the simple web query example.
 
 ```bash
-pip install -e examples/getting_started/simple_web_query
+uv pip install -e .
+uv pip install -e examples/getting_started/simple_web_query
 ```
 
 ### Downloading the NIM Containers
@@ -112,7 +113,7 @@ llms:
   nim_llm:
     _type: nim
     base_url: "http://localhost:8000/v1"
-    model_name: nvidia/llama3.1-nemotron-nano-4b-v1.1
+    model_name: nvidia/Llama-3.1-Nemotron-Nano-4B-v1.1
 
 embedders:
   nv-embedqa-e5-v5:
@@ -142,10 +143,11 @@ vLLM provides an [OpenAI-Compatible Server](https://docs.vllm.ai/en/latest/getti
 
 ### Install the Simple Web Query Example
 
-First, ensure the current working directory is the root of the NeMo Agent toolkit repository. Then, install the simple web query example so we have the `webpage_query` tool available.
+First, ensure the current working directory is the root of the NeMo Agent toolkit repository. Then, install NAT and the simple web query example.
 
 ```bash
-pip install -e examples/getting_started/simple_web_query
+uv pip install -e .
+uv pip install -e examples/getting_started/simple_web_query
 ```
 
 ### Serving the Models

docs/source/workflows/sizing-calc.md

@@ -21,6 +21,16 @@ The NVIDIA NeMo Agent toolkit provides a sizing calculator to estimate the GPU c
 
 The sizing calculator uses the [evaluation](evaluate.md) and [profiling](./profiler.md) systems in the NeMo Agent toolkit.
 
+## Prerequisites
+Sizing calculator uses the profiler subsystem. Ensure that it is installed by running the following command:
+```bash
+uv pip install -e ".[profiling]"
+```
+If you are installing from a package, you need to install the `nvidia-nat[profiling]` package by running the following command:
+```bash
+uv pip install "nvidia-nat[profiling]"
+```
+
 ## Overview
 
 This guide assumes that you have an LLM hosted by an isolated GPU cluster, for which you want to perform the sizing calculations for.
@@ -40,6 +50,10 @@ mkdir -p ${CALC_OUTPUT_DIR}
 
 cp examples/evaluation_and_profiling/simple_calculator_eval/configs/config-sizing-calc.yml $CONFIG_FILE
 ```
+Install the simple calculator example:
+```bash
+uv pip install -e examples/evaluation_and_profiling/simple_calculator_eval
+```
 
 Edit `.tmp/sizing_calc/config-sizing-calc.yml` file by adding a `base_url` parameter for the `llms.nim_llm` section for your cluster. Then, if needed, change the `llms.nim_llm.model_name`.