Issue/1638 model metadata

doc/source/examples/metadata.nblink

@@ -0,0 +1,3 @@
+{
+  "path": "../../../examples/models/metadata/metadata.ipynb"
+}

doc/source/python/python_component.md

@@ -103,7 +103,7 @@ To return metrics associated with a call create a method with signature as shown
     def metrics(self) -> List[Dict]:
 ```
 
-This method should return a Dictionary of metrics as described in the [custom metrics](../analytics/custom_metrics.md) docs.
+This method should return a Dictionary of metrics as described in the [custom metrics](../analytics/analytics.md#custom-metrics) docs.
 
 An illustrative example is shown below:
 
@@ -443,9 +443,53 @@ class Model:
 
 ### REST Metadata Endpoint
 The python wrapper will automatically expose a `/metadata` endpoint to return metadata about the loaded model.
-It is up to the developer to implement a `metadata` method in their class to provide an arbitrary `dict` back containing the model metadata.
+It is up to the developer to implement a `metadata` method in their class to provide a `dict` back containing the model metadata.
 
-Note: future work will most likely standardize `/metadata` endpoint and change behaviour of this method. See [this](https://github.com/SeldonIO/seldon-core/issues/1638) GitHub issue for details.
+#### Example format:
+```python
+class Model:
+    ...
+
+    def init_metadata(self):
+
+        meta = {
+            "name": "model-name",
+            "versions": ["model-version"],
+            "platform": "platform-name",
+            "inputs": [{"name": "input", "datatype": "BYTES", "shape": [1]}],
+            "outputs": [{"name": "output", "datatype": "BYTES", "shape": [1]}],
+        }
+
+        return meta
+```
+
+#### Validation
+Output of developer-defined `metadata` method will be validated to follow the [kfserving dataplane proposal](https://github.com/kubeflow/kfserving/blob/master/docs/predict-api/v2/required_api.md#model-metadata) protocol, see [this](https://github.com/SeldonIO/seldon-core/issues/1638) GitHub issue for details:
+```
+$metadata_model_response =
+{
+  "name" : $string,
+  "versions" : [ $string, ... ] #optional,
+  "platform" : $string,
+  "inputs" : [ $metadata_tensor, ... ],
+  "outputs" : [ $metadata_tensor, ... ]
+}
+```
+with
+```
+$metadata_tensor =
+{
+  "name" : $string,
+  "datatype" : $string,
+  "shape" : [ $number, ... ]
+}
+```
+
+If validation fails server will reply with `500` response `MICROSERVICE_BAD_METADATA` when requested for `metadata`.
+
+#### Examples:
+- [Basic Examples for Model with Metadata](../examples/metadata.html)
+- [SKLearn Server example with MinIO](../examples/minio-sklearn.html)
 
 
 

examples/dvc/.dvc/.gitignore

@@ -0,0 +1,9 @@
+/config.local
+/updater
+/lock
+/updater.lock
+/tmp
+/state-journal
+/state-wal
+/state
+/cache

examples/dvc/.gitignore

@@ -0,0 +1,2 @@
+.env
+/model.joblib

examples/dvc/Makefile

@@ -0,0 +1,9 @@
+env:
+	python3 -m venv .env
+	./.env/bin/pip install --upgrade pip setuptools
+	./.env/bin/pip install -r requirements.txt
+
+train:
+	.env/bin/python train_iris.py
+
+model: env train

examples/dvc/deploy.yaml

@@ -0,0 +1,19 @@
+
+apiVersion: machinelearning.seldon.io/v1
+kind: SeldonDeployment
+metadata:
+  name: minio-dvc-sklearn
+spec:
+  annotations:
+    seldon.io/executor: "true"
+  name: iris
+  predictors:
+  - componentSpecs:
+    graph:
+      children: []
+      implementation: SKLEARN_SERVER
+      modelUri: s3://dvc-iris
+      envSecretRefName: seldon-init-container-secret
+      name: classifier
+    name: default
+    replicas: 1

examples/dvc/metadata.yaml

@@ -0,0 +1,12 @@
+
+name: iris
+versions: [iris/dvc:8104914e6936da9864603b9bc4be2114]
+platform: sklearn
+inputs:
+- datatype: BYTES
+  name: input
+  shape: [ 1, 4 ]
+outputs:
+- datatype: BYTES
+  name: output
+  shape: [ 3 ]

examples/dvc/model.dvc

@@ -0,0 +1,15 @@
+md5: 99124ac10a601ab1d9f07f9c392b5d89
+cmd: make model
+deps:
+- md5: 65fd61883993b68d1937bdc36c59b20c
+  path: Makefile
+- md5: 7e8ce9f96492fee21db6a59c2b52f34d
+  path: requirements.txt
+- md5: 49c19c3ea9deb642066c0a457181cfbf
+  path: train_iris.py
+outs:
+- md5: 8104914e6936da9864603b9bc4be2114
+  path: model.joblib
+  cache: true
+  metric: false
+  persist: false

examples/dvc/requirements.txt

@@ -0,0 +1,4 @@
+scikit-learn == 0.20.3
+numpy >= 1.8.2
+joblib >= 0.13.0
+

examples/dvc/secret.yaml

@@ -0,0 +1,11 @@
+
+apiVersion: v1
+kind: Secret
+metadata:
+  name: seldon-init-container-secret
+type: Opaque
+stringData:
+  AWS_ACCESS_KEY_ID: minioadmin
+  AWS_SECRET_ACCESS_KEY: minioadmin
+  AWS_ENDPOINT_URL: http://minio.minio-system.svc.cluster.local:9000
+  USE_SSL: "false"