use different port for websocket server

README.md

@@ -1,32 +1,35 @@
 # ctts
 
 This project was inspired by wanting **a way of seeing what the
-[sky brightness](https://en.wikipedia.org/wiki/Sky_brightness) is at different points around the globe, and how that is
-changing over time**..
+[sky brightness](https://en.wikipedia.org/wiki/Sky_brightness)
+is over the entire earth surface, and how it changes over time**..
 
-Needless to say, we cannot have sensors for this sort of thing
-everwhere we would want a measurement, therefore we need a way
-of performing this measurement
-in a less direct way..
+Given that it would be infeasible to have [sensors](http://unihedron.com/projects/darksky/TSL237-E32.pdf)
+everywhere that we want a brightness measurement, it would make
+sense to have a way of performing this measurement indirectly.
 
 ---
 
 The approach this project takes is to model the relationship
 between a set of independent variables and the dependent variable
-(sky brightness) available in a [public dataset](http://www.unihedron.com/projects/darksky/database/?csv=true).
+(sky brightness) available in a [public dataset](http://www.unihedron.com/projects/darksky/database/?csv=true) using
+pytorch.
 
 H3 is then used to discretize cells over the earth that form
 the basis of the requests made to the api server serving up
 the sky brightness.
 
-Another component pushes brightness values from the model onto
-rabbitmq while saving them to postgres, to enable historical lookup.
+This component pushes brightness values from the model onto
+rabbitmq while another component consumes them - inserting to
+postgres to enable historical lookup.
 
 ## running locally
 
-this will spin up the process of the prediction producer container repeatedly asking the api server for sky brightness
-measurements across all [resolution 0 h3 cells](https://h3geo.org/docs/core-library/restable/) and publishing to
-rabbitmq, which the consumer container reads from and stores in postgres.
+This will spin up the process of the prediction producer container
+repeatedly asking the api container for sky brightness measurements
+_at the current time_ across all [resolution 0 h3 cells](https://h3geo.org/docs/core-library/restable/);
+publishing them to rabbitmq, which the consumer container reads from
+and stores messages (in postgres container).
 
 ```shell
 # create the volume for weather data
@@ -38,3 +41,26 @@ docker volume create --name open-meteo-data
 # run the containers
 docker-compose up --build
 ```
+
+Rabbitmq will take time to start up, at which time `producer` and
+`consumer` containers will attempt restart to form connection.
+Once rabbitmq does start, there should be output like this:
+
+```shell
+producer-1   | 2024-07-27 23:48:09,301 [INFO] publishing brightness message {'uuid': '86967a4d-d6a5-4421-8db1-73d9d0d45ea5', 'lat': 11.509775527199592, 'lon': -55.499062349013, 'h3_id': '805ffffffffffff', 'utc_iso': '2024-07-27T23:48:09.301613', 'utc_ns': 1722124089301613056, 'mpsas': 8.6264, 'model_version': '0.1.0'}
+producer-1   | 2024-07-27 23:48:09,304 [INFO] 805ffffffffffff has had 2 predictions published
+consumer-1   | 2024-07-27 23:48:09,321 [INFO] inserting brightness message for 805ffffffffffff
+consumer-1   | 2024-07-27 23:48:09,327 [INFO] broadcasting to 0 websocket clients on consumer:8090
+api-1        | 2024-07-27 23:48:09,922 [INFO] 172.19.0.7:43300 - "GET /api/v1/predict?lat=16.702868303031234&lon=-13.374845104752373 HTTP/1.1" 200
+producer-1   | 2024-07-27 23:48:09,926 [INFO] HTTP Request: GET http://api:8000/api/v1/predict?lat=16.702868303031234&lon=-13.374845104752373 "HTTP/1.1 200 OK"
+producer-1   | 2024-07-27 23:48:09,927 [INFO] publishing brightness message {'uuid': '01d5912d-752b-4611-8de9-4f092dc21c5c', 'lat': 16.702868303031234, 'lon': -13.374845104752373, 'h3_id': '8055fffffffffff', 'utc_iso': '2024-07-27T23:48:09.927243', 'utc_ns': 1722124089927243008, 'mpsas': 4.5332, 'model_version': '0.1.0'}
+producer-1   | 2024-07-27 23:48:09,930 [INFO] 8055fffffffffff has had 3 predictions published
+consumer-1   | 2024-07-27 23:48:09,945 [INFO] inserting brightness message for 8055fffffffffff
+consumer-1   | 2024-07-27 23:48:09,950 [INFO] broadcasting to 0 websocket clients on consumer:8090
+api-1        | 2024-07-27 23:48:10,540 [INFO] 172.19.0.7:43300 - "GET /api/v1/predict?lat=-7.460529604384309&lon=84.45314174117765 HTTP/1.1" 200
+producer-1   | 2024-07-27 23:48:10,544 [INFO] HTTP Request: GET http://api:8000/api/v1/predict?lat=-7.460529604384309&lon=84.45314174117765 "HTTP/1.1 200 OK"
+producer-1   | 2024-07-27 23:48:10,547 [INFO] publishing brightness message {'uuid': 'd015ee51-d60e-4722-9fb3-f442fd3c52e3', 'lat': -7.460529604384309, 'lon': 84.45314174117765, 'h3_id': '8087fffffffffff', 'utc_iso': '2024-07-27T23:48:10.547142', 'utc_ns': 1722124090547142144, 'mpsas': 9.0234, 'model_version': '0.1.0'}
+producer-1   | 2024-07-27 23:48:10,549 [INFO] 8087fffffffffff has had 4 predictions published
+consumer-1   | 2024-07-27 23:48:10,565 [INFO] inserting brightness message for 8087fffffffffff
+consumer-1   | 2024-07-27 23:48:10,572 [INFO] broadcasting to 0 websocket clients on consumer:8090
+```

docker-compose.yml

@@ -73,7 +73,7 @@ services:
   consumer:
     build: ./pc
     ports:
-      - "8080:8080"
+      - "8090:8090"
     environment:
       AMQP_HOST: "rabbitmq"
       PG_DATABASE: "postgres"

pc/pc/config.py

@@ -14,4 +14,4 @@
 AMQP_PREDICTION_QUEUE = os.getenv("AMQP_PREDICTION_QUEUE", "prediction")
 
 WS_HOST = os.getenv("WS_HOST", "consumer")
-WS_PORT = int(os.getenv("WS_PORT", 8080))
+WS_PORT = int(os.getenv("WS_PORT", 8090))

pc/pc/main.py

@@ -9,7 +9,7 @@
 from pc.model import BrightnessMessage
 from pc.websockets_handler import WebSocketsHandler
 
-logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
+logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
 log = logging.getLogger(__name__)
 
 websockets_handler = WebSocketsHandler()

pc/pc/websockets_handler.py

@@ -28,6 +28,6 @@ async def register_client(websocket):
 
     async def broadcast(self, message: BrightnessMessage):
         """send the message to all websockets"""
-        log.info(f"broadcasting to {len(self.clients)} clients")
+        log.info(f"broadcasting to {len(self.clients)} websocket clients on {WS_HOST}:{WS_PORT}")
         message_json = json.dumps(asdict(message))
         broadcast(self.clients, message_json)