README
¶
Transformers
We have sample transformers echo, md5 and tar2tf.
| Transformer | Language | Description |
|---|---|---|
echo |
python:3.8.5 |
This returns the original data, with an md5 sum in the response headers. |
go-echo |
golang:1.15 |
This returns the original data, with an md5 sum in the response headers. |
hello_world |
python:3.8.5 |
This always returns Hello World! string on any request. |
md5 |
python:3.8.5 |
This returns the md5 sum of the original data as the response. |
tar2tf |
golang:1.13 |
This returns the transformed TensorFlow compatible data for the input tar files. |
Each of sample transformers contains three main files:
Dockerfile- contains description of Docker image.pod.yaml- contains the spec for the K8s Pod running this transformer.Makefile- used for building and pushing Docker image.
Using sample transformers
Environment Variables
| Env Variable | Default Value | Description |
|---|---|---|
${COMMUNICATION_TYPE} |
"hpush://" |
This is the mode of communication to be used. For more info read this. |
Terms
HostMachine - The place from where the deployment is performed.
Prerequisites
kubectlmust be installed onHostMachineand connected and correctly configured foraistorekubernetes network.dockermust be installed on theHostMachine.
Steps:
- We must set the
${COMMUNICATION_TYPE}environment variable first with appropriate value. eval "echo \"$(cat pod.yaml)\"" > pod_config.yaml- substitutes the environment variablesmake build- builds the Docker imagedocker push <IMAGE_URL>- <IMAGE_URL> can be obtained post the build stepais transform init pod_config.yaml
Directories
Click to show internal directories.
Click to hide internal directories.