Merge pull request Azure-Samples#140 from Azure-Samples/pmishra/rag_doc_updates

updated docs for the advanced RAG solution

Author: vivekchettiar

Solution_Accelerators/Advanced_RAG/FAQ.md

@@ -96,7 +96,7 @@ This error occurs when your user or service principal does not have the necessar
 This error occurs because your account or service principal does not have the "Search Index Data Reader" role assigned for the selected index. Ask your service administrator to grant you this role in the Azure Portal → Access Control (IAM) section of your Azure AI Search resource.
 
 ### How to keep my data updated?
-It is best that you have a schedule for re-running ingestion with every new set of data to make sure your index is updated. To do so follow instructions [here.](./src/skills/ingestion/README_RETAIL.md/#keep-your-index-updated-by-re-running-ingestion)
+It is best that you have a schedule for re-running ingestion with every new set of data to make sure your index is updated. To do so follow instructions [here.](./src/skills/ingestion/README_FINANCIAL.md/#running-ingestion-service-locally)
 
 ### Why is my frontend deployment script taking long?
 Frontend Deployment needs to compress Node build files, this can take a while using using built in Windows Zip. To speed up the process install 7Zip (https://www.7-zip.org/).

Solution_Accelerators/Advanced_RAG/README.md

@@ -12,9 +12,10 @@ To read more about the underlying design principles, architecture, and solution
 * For more information on best practices on evaluation, architecture or validation of the solution design and outputs, please see the 'Additional Resources' section in solution documentation: [Advanced RAG Solution Accelerator Documentation](./docs/Advanced%20RAG%20Solution%20Accelerator%20Documentation.pdf).
 
 ## Table of Contents
+
 - [Use Case: Copilot for Financial Reports](#use-case-copilot-for-financial-reports)
-- [Features](#Features)
-- [Architecture](#Architecture)
+- [Features](#features)
+- [Architecture](#architecture)
 - [Getting Started](#getting-started)
 - [Code of Conduct](#code-of-conduct)
 - [Responsible AI Guidelines](#responsible-ai-guidelines)
@@ -90,4 +91,4 @@ This dataset is released under the Community Data License Agreement – Permissi
 
 ## License
 
-This project is licensed under the MIT License. See the [LICENSE](./LICENSE) file for details.
+This project is licensed under the MIT License. See the [LICENSE](./LICENSE) file for details.

Solution_Accelerators/Advanced_RAG/SETUP.md

@@ -225,7 +225,7 @@ When you begin to setup the AI Assistant deployment, you must must have `Microso
 
 ![ConstosoSub](./docs/media/contososubscription.png)
 
-If you have successfully completed the [Azure Resources setup](#azure-resources-setup), the required resources for end-to-end local setup should already be provisioned. You can proceed by verifying the existence of the following resources in your Azure Resource Group:
+If you have successfully completed the [Azure Resources setup](#setting-up-azure-resources), the required resources for end-to-end local setup should already be provisioned. You can proceed by verifying the existence of the following resources in your Azure Resource Group:
 
 - Azure Key Vault
 - Azure OpenAI
@@ -421,7 +421,7 @@ More details on how to use the tool can be found in the [evaluation tools readme
 
 ### Azure Kubernetes Setup
 **Note:**
-Setting up the Azure Kubernetes Service (AKS) and application gateway is only needed if you want the services to be deployed to cloud. Users can proceed with setting up the keyvault and then follow the instructions to [run the solution locally in VSCode](#local-setup). Once that succeeds, they can then come and setup the AKS.
+Setting up the Azure Kubernetes Service (AKS) and application gateway is only needed if you want the services to be deployed to cloud. Users can proceed with setting up the keyvault and then follow the instructions to [run the solution locally in VSCode](#running-services-locally). Once that succeeds, they can then come and setup the AKS.
 
 1. To work with the AKS cluster, developers will need `Azure Kubernetes Service RBAC Admin` role assigned.
 
@@ -520,7 +520,7 @@ Additionally, developers would need `Cosmos DB Built-in Data Contributor` role w
     ```
 
 ### Keyvault
-1. To work with keyvault, developers will need GET, LIST and SET permissions to keyvault secrets. Refer [AKS](#aks) to grant these permissions via access policies tab.
+1. To work with keyvault, developers will need GET, LIST and SET permissions to keyvault secrets. Refer [AKS](#azure-kubernetes-setup) to grant these permissions via access policies tab.
 2. Navigate through `config.py` files across micro-services in the `<repo root>\Advanced_RAG\src` folder: config_hub, skills\search, orchestrator_rag, data, session_manager and add any un-populated secrets with appropriate values in the keyvault. Examples:
     ```
     az keyvault secret set --vault-name <key vault name> --name "KEYVAULT-URI" --value "https://<keyvault name>.vault.azure.net/"

Solution_Accelerators/Advanced_RAG/docs/demo_script/Analyzing Microsoft Financial Performance Demo Script.pdf

@@ -58,7 +58,7 @@ The tool defaults to calling RAG Bot's Dev endpoint. You can point `rag_eval` to
 ]
 ```
 
-RAG Bot is non-deterministic. You can repeat the same query multiple times to get a more accurate evaluation. 
+RAG Bot is non-deterministic. You can repeat the same query multiple times to get a more accurate evaluation.
 ```json
 "args": [
     "--eval_flow=evaluate_end_to_end",
@@ -91,7 +91,7 @@ The Configuration Service is a separate application for uploading and managing R
 
 The `--orchestrator_config` can be set to a specific version ID for an existing configuration in the Configuration Service. The tool will then use this configuration for the evaluation. You can also provide an absolute path to a local config json file which will be automatically uploaded to the Configuration Service. The version ID of the uploaded configuration will be the same experiment ID as the evaluation.
 
-**Leverage an existing configuration:** 
+**Leverage an existing configuration:**
 ```json
 "args": [
     "--eval_flow=evaluate_end_to_end",
@@ -110,11 +110,11 @@ The `--orchestrator_config` can be set to a specific version ID for an existing
 ]
 ```
 
-Similarly, configuring the search index can be done with the `--search_config` argument. 
+Similarly, configuring the search index can be done with the `--search_config` argument.
 
 
 
-## [Important] Upload Evaluation Flow to Prompt Flow for `evaluate_end_to_end` 
+## [Important] Upload Evaluation Flow to Prompt Flow for `evaluate_end_to_end`
 1) Go to prompt_flow_resources/RAG_E2E_Eval_Flow/flow.dag.yaml and replace \<AZURE-OPENAI-CONNECTION> with your Azure OpenAI connection name
 2) Compress all the files under prompt_flow_resources/RAG_E2E_Eval_Flow into a ZIP File (Note: flow.dag.yaml must be in the root of your ZIP File)
 3) Navigate to ml.azure.com and go to your workspace
@@ -176,7 +176,7 @@ Example with formatted Markdown table:
 | Why did the investment in associates and joint ventures change significantly?  | Investment in associates and joint ventures increased due to the acquisition of a new subsidiary. | 0                  | 2.0                    |
 | What is the largest net currency exposure for the Group?                        | The Group has a net exposure of AED 7.24b to Indian Rupee.                                        | 1                  | 7.0                    |
 | Why did share of profit from investment in associates change?                   | The company recognized Gain on acquisition of additional stake in XYZ Ltd.                         | 2                  | 2.0                    |
-| What was the Net Profit and Net Profit growth for the year?                     | Net Profit, before non-controlling interests, increased by 15% to AED 5.6b.                        | 3                  | 1.0                    |ntrolling interests, ...  ...                  3                     1.0  
+| What was the Net Profit and Net Profit growth for the year?                     | Net Profit, before non-controlling interests, increased by 15% to AED 5.6b.                        | 3                  | 1.0                    |ntrolling interests, ...  ...                  3                     1.0
 
 Runs can also be viewed in the Azure Machine Learning workspace under the `RAG_E2E_Eval_Flow` flow.
 
@@ -196,25 +196,4 @@ RAG Bot's answers are saved to new version of the `RAG-Eval-Results` AML data as
 
 By default, the tool uses the `RAG_E2E_Eval_Flow`. This flow can be extended to support additional evaluation scenarios, such as RAG evaluations (groundedness, relevance, etc.). A new run shhoul be created for each successful evaluation.
 
-![alt text](documentation/flow_ui.png)
-
-## Roadmap
-
-### V1 Feature TODO
-- [X] Configuration service support.
-- [X] Make the evaluation dataset version configurable. 
-- [X] Dataset configuration: Duplicates per sample.
-- [X] Unify evaluation flows.
-
-### V2 Feature TODO
-- [ ] Migrate E2E Evaluation to Prompty.
-- [ ] Support context evaluations such as relevance and groundedness. 
-- [ ] Support running from VS Code.
-- [ ] Support running from Azure DevOps Build Pipeline
-- [ ] Expose tool at pip package with changelog. Must include a script to run, ideally as an ADO pipeline.
-- [ ] Python API which returns experiment results in-memory.
-- [ ] Dataset configuration: Paraphrases per sample in E2E Dataset.
-
-### Testing TODO
-- [X] Write README that others can follow without guidance.
-- [ ] Confirm that others can run the tool.
+![alt text](documentation/flow_ui.png)