Modify step image README.md (#2489)

Author: lhw362950217

docker/step/README.md

@@ -2,14 +2,27 @@
 
 This image is used when submitting Argo workflows to run Tensorflow/PAI Tensorflow jobs. To build this image, you should follow the below steps:
 
-1. Run `docker run --rm -it -v $PWD:/opt/output sqlflow/sqlflow cp -r /usr/local/bin/step /opt/sqlflow/parser/parser-1.0-SNAPSHOT-jar-with-dependencies.jar /opt/output` to copy latest `step` and parser jar file.
-2. Copy `python/sqlflow_submitter` directory to this directory.
+1. Go to SQLFlow root directory
+    ```bash
+    cd $(git rev-parse --show-toplevel)
+    ```
+1. Build the `sqlflow:dev` docker image as described [here](../dev/README.md), only needed when you haven't done it.  For short, you can just execute below command:
+    ```bash
+    docker build -t sqlflow:dev -f docker/dev/Dockerfile .
+    ```
+1. Run `sqlflow:dev` image to build SQLFlow project.  This process will generate a directory called `build` under current directory which contains all the binaries we need.
+    ```bash
+    docker run --rm -it \
+    -v $GOPATH:/root/go \
+    -v $HOME/.m2:/root/.m2 \
+    -v $HOME/.cache:/root/.cache \
+    -v $PWD:/work -w /work \
+    sqlflow:dev
+    ```
+1. Run below command to build the `sqlflow:step` docker image.  This command will package perviously built binaries into the result image.
+    ```bash
+    docker build -t sqlflow:step -f docker/step/Dockerfile .
+    ```
 
-After above steps, we are expecting to have below files under the current directory before we can run `docker build`:
-
-- step
-- parser-1.0-SNAPSHOT-jar-with-dependencies.jar
-- sqlflow_submitter/
-
-Then run `docker build -t sqlflow_tensorflow_step:slim .`, and push this image to some registry that your Kubernetes cluster can reach. To use it in SQLFlow, specify the environment variable `export SQLFLOW_WORKFLOW_STEP_IMAGE=sqlflow_tensorflow_step:slim` to use this image as step image.
+To use it in SQLFlow, specify the environment variable `export SQLFLOW_WORKFLOW_STEP_IMAGE=sqlflow:step` when you start `sqlflow server`, or set the image in a Kubernetes config file, you can see an example [here](https://github.com/sql-machine-learning/sqlflow/blob/f5dc0209fe1bd71c443e82ebb4f6981b06e33542/doc/run/k8s/install-sqlflow.yaml#L16).