Samples and README.md

README.md

@@ -1 +1,69 @@
-# AWS Greengrass ML Inference Playground
+# AWS Greengrass ML Inference Playground
+
+This codifies [AWS ML inference tutorial](https://docs.aws.amazon.com/greengrass/latest/developerguide/ml-console.html). The intend is a consisten repeatable development environment for experimenting and enjoying Greengrass ML Inference.
+
+* Instead of using AWS Console, Python/BotoCLI is used via [Greengo](http://greengo.io)
+* Instead of Raspberry PI, a Ubuntu VM.
+* Instead of a camera, images are dropped to an `/images` folder in a VM. The ML inference Lambda function
+  uses access to local volume to read the images.
+
+
+The instructions below is a rough draft that gives direction and can work with the right amount of extra hacking.
+
+> Read to the end before attempting to run.
+> If it fails, don't get upset: this is work in progress and a trail, not a highway.
+> Roll your sleeves, hack alone, PRs welcome.
+
+
+1. Install [`greengo`](http://greengo.io). Use the head of `master` - I am moving this and `greengo` together.
+
+    ```
+    pip install git+git://github.com/dzimine/greengo.git#egg=greengo
+    ```
+
+2. Create a Greengrass Core definition in AWS. See [./greengo.yaml](./greengo.yaml) for what will be created.
+
+    ```
+    greengo create
+    ```
+
+3. Provision Greengrass Core VM. Vagrant will do the magic of installing Greengrass, placing certificates and configuration,
+    and adding the necessary ML stuff on the GG Core VM.
+
+    ```
+    vagrant up
+    ```
+    Ah, and before you run it, you should place `greengrass-ubuntu-x86-64-VERSION.tar.gz` into `./downloads`.
+    What? You already run it? You should have read the instruction to the end!
+
+4. Deploy Greengrass Core
+
+    ```
+    greengo deploy
+    ```
+
+5. ***Profit!***
+   
+   Login to the Vagrant VM. Drop image files to `/images`. There are few already in `./images`, and a script you can run in VM to randomly drop the files:
+   
+   ```
+   vagrant ssh
+   sudo su
+   cd /vagrant/scripts
+   ./feed.sh
+   ```
+   Watch the GGC log files to see the results:
+   
+   ```
+   tail -f /greengrass/ggc/var/log/user/$AWS_REGION/$AWS_ACCOUNT/GreengrassImageClassification.log
+   ```
+
+   I'll eventually get to posting them to MQTT - or you can take it as an exercise.
+   
+   If you don't like the prediction, it has nothing to do with Greengrass - blame the pretrained [squeezenet1_1 model](http://data.dmlc.ml/mxnet/models/imagenet/squeezenet/) offered in the AWS Tutorial, and bring a better one.
+
+6. The best part: once you've done playing, clean up your AWS resources:
+
+    ```
+    greengo remove
+    ```

scripts/feed.sh

@@ -0,0 +1,12 @@
+files=(../images/*.jpg)
+
+RANDOM=$$$(date +%s)
+
+while [[ 1 ]]; do
+    f=${files[$RANDOM % ${#files[@]}]}
+    echo "Placing $(basename $f)..."
+    d=/images/$(basename $f)
+    cp $f $d
+    mv "$d" "${d%.jpg}.jpeg"
+    sleep 5
+done