doc: Add CI and CD operation doc for composer language models

docs/archived docs/how-to-process-language-models-with-CI-CD.md

@@ -0,0 +1,88 @@
+# Using botframework-cli tool to process language models for Composer bots
+
+Composer allows users to enable kinds of languages models such as [LUIS](https://www.luis.ai/home) and [QnAMaker](https://www.qnamaker.ai/). Composer can help to preprocess and publish these models to azure services if users provide the resource key and other essential elements. Fortunately, outside Composer, users can also use botframework-cli tool to achieve the same goal with below steps.
+* [`cross-train`](#Cross-train-luis-and-qna-files)
+* [`down-sampling`](#Down-sampling-on-luis-models)
+* [`luis build`](#Luis-models-build)
+* [`qnamaker build`](#Qnamaker-models-build)
+
+## Prerequisites
+[Botframework-cli](https://github.com/microsoft/botframework-cli/blob/main/README.md) tool is based on the Node.js platform and the [OClif](https://github.com/oclif/oclif) framework where it inherits its command line parsing style, and plugin architecture platform.
+
+Install the tool using the following command:
+
+~~~
+$ npm i -g @microsoft/botframework-cli@next
+$ bf plugins:install @microsoft/bf-sampler-cli@beta
+~~~
+
+Type `bf` to make sure the command is installed successfully.
+
+You can find more details about [sampler](https://github.com/microsoft/botframework-cli/blob/beta/packages/sampler/README.md) cli plugin.
+
+You also need to create [luis resources](https://docs.microsoft.com/en-us/azure/cognitive-services/luis/luis-how-to-azure-subscription#create-luis-resources-in-azure-portal) to get the luis authoringKey and [qnamaker resources](https://docs.microsoft.com/en-us/azure/cognitive-services/qnamaker/how-to/set-up-qnamaker-service-azure) to get the qnamaker subscriptionKey if you don't have.
+
+## Process language models with cli tool
+You can use the cli tool to process language models . A sample bot will be used to demonstrate how cli tool works with below steps.
+
+### Create bot with luis and qna language models
+Open Composer and choose to create a bot from template. Select the `Todo with LUIS` example bot. Go to bot root folder.
+
+~~~
+$ cd ToDoBotWithLuisSample-0
+~~~
+
+### Cross train luis and qna files
+First create folder that holds cross-trained language model files and then use bf cli tool to cross train language models(LUIS and QnAMaker) files.
+
+~~~
+$ mkdir generated\interruption
+$ bf luis:cross-train --in . --out generated\interruption --config settings\cross-train.config.json --force
+~~~
+
+You can find more details about [cross-train](https://github.com/microsoft/botframework-cli/tree/main/packages/luis#bf-luiscross-train) cli tool.
+
+### Down sampling on luis models
+Use bf cli tool to do down sampling on luis language model files and overwrite original ones. This step can be skipped if luis model files do not exist.
+
+~~~
+$ bf sampler:sampling --in generated\interruption --out generated\interruption --force
+~~~
+
+You can find more details about [sampling](https://github.com/microsoft/botframework-cli/blob/beta/packages/sampler/README.md#bf-samplersampling) cli tool.
+
+### Luis models build
+Use bf cli tool to build luis models and generate model setting if luis model files exist. Here YOUR_LUIS_AUTHORING_KEY can be got from your luis resources.
+
+~~~
+$ bf luis:build --in generated\interruption --authoringKey {YOUR_LUIS_AUTHORING_KEY} --botName {YOUR_BOT_NAME} --out generated --suffix composer --force --log
+~~~
+
+You can find more details about [luis:build](https://github.com/microsoft/botframework-cli/tree/main/packages/luis#bf-luisbuild) cli tool.
+
+### Qnamaker models build
+Use bf cli tool to build qnamaker models and generate model setting if qnamaker model files exist. Here YOUR_QNAMaker_SUBSCRIPTION_KEY can be got from your qnamaker resources.
+
+~~~
+$ bf qnamaker:build --in generated\interruption --subscriptionKey {YOUR_QNAMaker_SUBSCRIPTION_KEY} --botName {YOUR_BOT_NAME} --out generated --suffix composer --force --log
+~~~
+
+You can find more details about [qnamaker:build](https://github.com/microsoft/botframework-cli/tree/main/packages/qnamaker#bf-qnamakerbuild) cli tool.
+
+### Test bot in emulator
+After language models published and the setting files generated, you can use Emulator to test the bot by following below steps:
+1. Configure appsettings.json
+Go to settings folder and edit the appsettings.json.
+- Fill the luis endpoint and endpointKey of prediction resource. You can find the endpoint and endpointKey after assigning the prediction resource to your models in [LUIS](https://www.luis.ai/home) portal. If there are no available prediction resources to assign, please create one in you luis resources. For more details about prediction resource, please find them [here](https://docs.microsoft.com/en-us/azure/cognitive-services/luis/luis-how-to-azure-subscription#assign-a-resource-to-an-app).
+- Fill the qna endpointKey. You can find it on the console output of qnamaker build cmd or in the service settings page of [QnAMaker](https://www.qnamaker.ai/UserSettings) portal.
+
+2. Copy and launch runtime
+~~~
+$ mkdir runtime
+$ xcopy /S "{YOUR_DOTNET_RUNTIME_FOLDER}" "runtime" // you can get the runtime by ejecting it in composer or find it in your composer installation folder
+$ cd runtime
+$ dotnet run --project azurewebapp // launch runtime
+~~~
+
+3. Open Emulator and connect the bot with bot url set to http://localhost:3979/api/messages.
+4. Communicate with the bot.