chore: READMEs and other adjustments

Ch09/Makefile

@@ -51,5 +51,3 @@ deploy:
 		--region=asia-northeast1 \
 		--display-name=flights \
 		--model=$(MODEL_ID) \
-
-

Ch09/README.md

@@ -25,6 +25,42 @@ This chapter uses Vertex AI to run custom training jobs.
 
 ## Tutorial
 
+## Deploy endpoint
+
+Register a model
+```shell
+make import-model GS_ARTIFACT_URI=MY_GS_URI
+```
+
+Create endpoint
+
+```shell
+gcloud ai endpoints create \
+  --display-name flights \
+  --region asia-northeast1
+```
+
+Deploy the model
+
+```shell
+make deploy ENDPOINT=MY_ENDPOINT MODEL_ID=MY_MODEL_ID
+```
+Undeploy the model
+```shell
+gcloud ai endpoints describe ${ENDPOINT_ID} \
+  --region asia-northeast1
+gcloud ai endpoints undeploy-model ${ENDPOINT_ID} \
+  --region asia-northeast1 \
+  --deployed-model-id ${DEPLOYED_MODEL_ID}
+```
+
+Terminate the endpoint when you are done
+
+```shell
+gcloud ai endpoints delete ${ENDPOINT_ID} \
+    	--region=asia-northeast1
+```
+
 ## Run a training job
 
 ```shell

Ch10/README.md

@@ -7,13 +7,50 @@ As a high level view, below functions will be implemented.
 * Ingest events from Google PubSub
 * Call oneline prediction API built in Chapter 09
 
+## Instructions
+
+### Setting up Bigtable
+
+Create a instance
+
+```shell
+gcloud bigtable instances create flights \
+  --display-name="flights" \
+  --cluster-config=id=datascienceongcp,nodes=1,zone=asia-northeast1-a
+```
+
+Create a table
+
+```shell
+gcloud bigtable instances tables create predictions --instance=flights --column-families=FL
+```
+
+Reading data
+
+See full example of cbt
+https://cloud.google.com/bigtable/docs/create-instance-write-data-cbt-cli#local-shell
+
+```shell
+gcloud components install cbt
+cbt -project elite-caster-125113 -instance flights read predictions
+```
+
+Clean up
+```shell
+gcloud bigtable instances delete flights
+```
+
+### Running apache beam
+
+#### Using direct runner
+
 ```shell
 PROJECT=$(gcloud config get core/project)
 OUTPUT_LOCATION=gs://${PROJECT}/flights/chapter10/output/
 GOOGLE_PROJECT_ID=$PROJECT
 DEPARTURE_DELAY_CSV_PATH=gs://${PROJECT}/flights/chapter8/output/delays.csv
 AI_PLATFORM_LOCATION=$(gcloud config get compute/region)
-AI_PLATFORM_ENDPOINT_ID="4765428530316050432"
+AI_PLATFORM_ENDPOINT_ID=MY_ENDOINT_ID
 TEMP_LOCATION=gs://${PROJECT}/flights/staging
 BIGTABLE_INSTANCE_ID=flights
 BIGTABLE_TABLE_ID=predictions
@@ -32,14 +69,16 @@ mvn compile exec:java \
     --bigtableTableId=$BIGTABLE_TABLE_ID"
 ```
 
+#### Using dataflow , tbc
+
 ```shell
 PROJECT=$(gcloud config get core/project)
 OUTPUT_LOCATION=gs://${PROJECT}/flights/chapter10/output/
 STAGING_LOCATION=gs://${PROJECT}/staging
 GOOGLE_PROJECT_ID=$PROJECT
 DEPARTURE_DELAY_CSV_PATH=gs://${PROJECT}/flights/chapter8/output/delays.csv
 AI_PLATFORM_LOCATION=$(gcloud config get compute/region)
-AI_PLATFORM_ENDPOINT_ID="3976559723712348160"
+AI_PLATFORM_ENDPOINT_ID=MY_ENDOINT_ID
 TEMP_LOCATION=gs://${PROJECT}/flights/staging
 RUNNER=DataflowRunner
 mvn compile exec:java \
@@ -54,17 +93,3 @@ mvn compile exec:java \
     --aiPlatformLocation=$AI_PLATFORM_LOCATION \
     --aiPlatformEndpointId=$AI_PLATFORM_ENDPOINT_ID"
 ```
-
-```shell
-gcloud bigtable instances create flights \
-  --display-name="flights" \
-  --cluster-config=id=datascienceongcp,nodes=1,zone=asia-northeast1-a
-```
-
-```shell
-gcloud bigtable instances tables create predictions --instance=flights --column-families=FL
-```
-
-```shell
-gcloud bigtable instances delete flights
-```