Replace main branch with rhoai-3.4 in autorag .md files

Author: MichalSteczko

examples/autorag/readme.md

@@ -165,7 +165,7 @@ To run the Documents RAG Optimization Pipeline, you provide:
 
 | Item | Description |
 |------|-------------|
-| **Pipeline parameters** | `embeddings_models` (list of embedding model IDs), `generation_models` (list of foundation model IDs), `optimization_metric` (e.g. `faithfulness`, `answer_correctness`, `context_correctness`), `llama_stack_vector_database_id` (e.g. `ls_milvus`). See the [pipeline README](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) for full input/output descriptions. |
+| **Pipeline parameters** | `embeddings_models` (list of embedding model IDs), `generation_models` (list of foundation model IDs), `optimization_metric` (e.g. `faithfulness`, `answer_correctness`, `context_correctness`), `llama_stack_vector_database_id` (e.g. `ls_milvus`). See the [pipeline README](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) for full input/output descriptions. |
 
 ---
 
@@ -214,7 +214,7 @@ You run AutoRAG by **running the Documents RAG Optimization Pipeline**:
 
 1. Ensure the **Llama-stack RAG stack** is deployed (see [Llama stack setup](https://github.com/red-hat-data-services/red-hat-ai-examples/blob/llama-stack_sample/examples/llama-stack/SETUP.md)) and that you have created a secret (or connection) with `LLAMA_STACK_CLIENT_BASE_URL` and `LLAMA_STACK_CLIENT_API_KEY` for the pipeline to use.
 2. Ensure the **sample documents** from [data/rh_summit_2026/input_data/](data/rh_summit_2026/input_data/) and the **benchmark** file [benchmark_data.json](data/rh_summit_2026/benchmark_data.json) are uploaded to S3 (same bucket, different paths), and that you have an S3 connection for that data plus a Pipeline Server configured with a results connection for artifacts.
-3. Add the **Documents RAG Optimization Pipeline** as a Pipeline Definition (from [pipelines-components](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline), branch `autox`). You can find it's compiled version [here](pipelines/pipeline.yaml).
+3. Add the **Documents RAG Optimization Pipeline** as a Pipeline Definition (from [pipelines-components](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline), branch `autox`). You can find it's compiled version [here](pipelines/pipeline.yaml).
 4. Create a pipeline run and set the required parameters: use the same connection and bucket for test data and input documents (different object keys); Llama-stack secret name; embeddings_models and generation_models lists; optimization_metric.
 5. **View the results** in the run's Artifacts: leaderboard HTML and RAG pattern artifacts (pattern.json, evaluation_results.json, indexing and inference notebooks).
 
@@ -230,7 +230,7 @@ For a step-by-step walkthrough, see the [Tutorial: Ask questions against Red Hat
 
 ## References
 
-- [Documents RAG Optimization Pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) — Pipeline definition, inputs, outputs, and artifact layout (branch `autox`)
+- [Documents RAG Optimization Pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) — Pipeline definition, inputs, outputs, and artifact layout (branch `autox`)
 - [Llama stack setup](https://github.com/red-hat-data-services/red-hat-ai-examples/blob/llama-stack_sample/examples/llama-stack/SETUP.md) — Installation and configuration for the Llama-stack RAG server (prerequisite for AutoRAG)
 - [IBM ai4rag](https://github.com/IBM/ai4rag) — RAG templates and optimization engine used by the pipeline
 - [Deploying a RAG stack in a project (Red Hat OpenShift AI)](https://docs.redhat.com/en/documentation/red_hat_openshift_ai_self-managed/3.0/html/working_with_llama_stack/deploying-a-rag-stack-in-a-project_rag)

examples/autorag/red_hat_summit_tutorial.md

@@ -53,7 +53,7 @@ The pipeline expects a Kubernetes secret containing the Llama-stack client URL a
 
 **② Create new connection of type `LLAMA_STACK_CONNECTION` inside your project**
 ![Llama Stack connection creation](images/llama_stack_connection_creation.png)
-**Note:** Use the connection’s **resource name** (or secret name) as `llama_stack_secret_name` when creating the pipeline run. For exact key names and options, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) README.
+**Note:** Use the connection’s **resource name** (or secret name) as `llama_stack_secret_name` when creating the pipeline run. For exact key names and options, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) README.
 
 <a id="create-s3-connections"></a>
 
@@ -128,7 +128,7 @@ Sample data is provided in this repository under **`data/rh_summit_2026/`**: inp
 | Step | Action |
 |------|--------|
 | **①** | From **Pipelines**, create a new pipeline run using **Pipeline definitions → ⋮ → Create run** for the Documents RAG Optimization Pipeline you added. |
-| **②** | Set the **Name** of the run and the following run parameters (see the [pipeline README](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) for full descriptions): **test_data_secret_name** and **input_data_secret_name** (same connection name from the single S3 data connection), **test_data_bucket_name** and **input_data_bucket_name** (same bucket), **test_data_key** (path to the benchmark file, e.g. `benchmark_data.json`), **input_data_key** (path to documents folder or prefix, e.g. `input_data/`), **llama_stack_secret_name** (secret with `LLAMA_STACK_CLIENT_BASE_URL` and `LLAMA_STACK_CLIENT_API_KEY`), **embeddings_models** (list of embedding model identifiers to try, e.g. `["ibm/slate-125m-english-rtrvr-v2", "intfloat/multilingual-e5-large"]`), **generation_models** (list of foundation/generation model identifiers, e.g. `["mistralai/mixtral-8x7b-instruct-v01", "ibm/granite-13b-instruct-v2"]`), **optimization_metric** (e.g. `faithfulness`, `answer_correctness`, or `context_correctness`; default `faithfulness`). </br> <span style="color:red">**Set `llama_stack_vector_database_id` (e.g. `ls_milvus`; default is `ls_milvus`) and please note that only `inline::milvus` provider type is currently supported**.</span> |
+| **②** | Set the **Name** of the run and the following run parameters (see the [pipeline README](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) for full descriptions): **test_data_secret_name** and **input_data_secret_name** (same connection name from the single S3 data connection), **test_data_bucket_name** and **input_data_bucket_name** (same bucket), **test_data_key** (path to the benchmark file, e.g. `benchmark_data.json`), **input_data_key** (path to documents folder or prefix, e.g. `input_data/`), **llama_stack_secret_name** (secret with `LLAMA_STACK_CLIENT_BASE_URL` and `LLAMA_STACK_CLIENT_API_KEY`), **embeddings_models** (list of embedding model identifiers to try, e.g. `["ibm/slate-125m-english-rtrvr-v2", "intfloat/multilingual-e5-large"]`), **generation_models** (list of foundation/generation model identifiers, e.g. `["mistralai/mixtral-8x7b-instruct-v01", "ibm/granite-13b-instruct-v2"]`), **optimization_metric** (e.g. `faithfulness`, `answer_correctness`, or `context_correctness`; default `faithfulness`). </br> <span style="color:red">**Set `llama_stack_vector_database_id` (e.g. `ls_milvus`; default is `ls_milvus`) and please note that only `inline::milvus` provider type is currently supported**.</span> |
 | **③** | Ensure the **Pipeline Server** is configured with the results S3 connection from [Create S3 connections](#create-s3-connections), so artifacts are stored in the expected bucket. |
 | **④** | Start the run via **Create run** and wait for it to complete. |
 
@@ -158,7 +158,7 @@ Sample data is provided in this repository under **`data/rh_summit_2026/`**: inp
 **Step ② Review the leaderboard results:**
 ![Pipeline Leaderboard Step](images/leaderboard.png)
 
-For exact artifact paths and layout, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) README (Outputs and "Files stored in user storage" sections).
+For exact artifact paths and layout, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) README (Outputs and "Files stored in user storage" sections).
 
 **Step 3 Locate best pattern directory inside your S3 storage (pattern 8 in this case) and review the generated files:**
 <a id="generated-files"></a>
@@ -198,4 +198,4 @@ Each RAG pattern in the **rag_patterns_artifact** includes two generated noteboo
 Test your own questions against the generated RAG pattern at the end of the notebook.
 
 ![RAG Query result](images/rag_query_response.png)
-The **inference.ipynb** is the main interface for querying your RAG system; the **indexing.ipynb** is needed when you first deploy a pattern or when you add or update documents in the index. For more detail on the contents of each notebook and the pattern configuration, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/main/pipelines/training/autorag/documents_rag_optimization_pipeline) README.
+The **inference.ipynb** is the main interface for querying your RAG system; the **indexing.ipynb** is needed when you first deploy a pattern or when you add or update documents in the index. For more detail on the contents of each notebook and the pattern configuration, see the [documents_rag_optimization_pipeline](https://github.com/red-hat-data-services/pipelines-components/tree/rhoai-3.4/pipelines/training/autorag/documents_rag_optimization_pipeline) README.