Dynamically deploy models to match request targets

.gitignore

@@ -85,7 +85,7 @@ ipython_config.py
 # pyenv
 #   For a library or package, you might want to ignore these files since the code is
 #   intended to run in multiple environments; otherwise, check them in:
-# .python-version
+.python-version
 
 # pipenv
 #   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.

README.md

@@ -41,10 +41,6 @@ through environment variables. Before deploying the Gateway, make sure to set th
 either by exporting them in the shell or by creating a `.env` file in the root directory of the
 project. The following variables are required:
 
-* `MLFLOW_TRACKING_URI`: The URI for the MLflow tracking server.
-* `CMS_PROJECT_NAME`: The name of the Docker project where the CogStack ModelServe stack is running.
-* `CMS_HOST_URL` (optional): Useful when running CogStack ModelServe instances behind a proxy. If
-  omitted, the Gateway will attempt to reach the services directly over the internal Docker network.
 * `CMG_SCHEDULER_MAX_CONCURRENT_TASKS`: The max number of concurrent tasks the scheduler can handle.
 * `CMG_DB_USER`: The username for the PostgreSQL database.
 * `CMG_DB_PASSWORD`: The password for the PostgreSQL database.
@@ -65,37 +61,29 @@ not allowed in MinIO bucket names). The configuration should be saved in a `.env
 directory of the project before running Docker Compose (or sourced directly in the shell):
 
 ```shell
-CMS_PROJECT_NAME=<cms-docker-compose-project-name>  # e.g. cms
-
-# (optional) Useful when running CMS behind a proxy
-CMS_HOST_URL=https://<proxy-docker-service-name>/cms  # e.g. https://proxy/cms
-
 CMG_SCHEDULER_MAX_CONCURRENT_TASKS=1
 
 # Postgres
 CMG_DB_USER=admin
 CMG_DB_PASSWORD=admin
-CMG_DB_HOST=postgres
+CMG_DB_HOST=db
 CMG_DB_PORT=5432
 CMG_DB_NAME=cmg_tasks
 
 # RabbitMQ
 CMG_QUEUE_USER=admin
 CMG_QUEUE_PASSWORD=admin
-CMG_QUEUE_HOST=rabbitmq
+CMG_QUEUE_HOST=queue
 CMG_QUEUE_PORT=5672
 CMG_QUEUE_NAME=cmg_tasks
 
 # MinIO
 CMG_OBJECT_STORE_ACCESS_KEY=admin
 CMG_OBJECT_STORE_SECRET_KEY=admin123
-CMG_OBJECT_STORE_HOST=minio
+CMG_OBJECT_STORE_HOST=object-store
 CMG_OBJECT_STORE_PORT=9000
 CMG_OBJECT_STORE_BUCKET_TASKS=cmg-tasks
 CMG_OBJECT_STORE_BUCKET_RESULTS=cmg-results
-
-# MLflow (use container IP when running locally)
-MLFLOW_TRACKING_URI=http://<mlflow-docker-service-name>:<mlflow-port>  # e.g. http://mlflow-ui:5000
 ```
 
 To install the CogStack Model Gateway, clone the repository and run `docker compose` inside the root

cogstack_model_gateway/common/config/__init__.py

@@ -0,0 +1,142 @@
+import json
+import logging
+import os
+
+from dotenv import load_dotenv
+from pydantic import ValidationError
+
+from cogstack_model_gateway.common.config.models import Config
+
+log = logging.getLogger("cmg.common")
+
+CONFIG_FILE = os.getenv("CONFIG_FILE", "config.json")
+
+_config_instance: Config | None = None
+
+
+def _load_json_config() -> dict:
+    """Load configuration from JSON file."""
+    try:
+        with open(CONFIG_FILE) as f:
+            config = json.load(f)
+            log.info(f"Loaded configuration from {CONFIG_FILE}")
+            return config
+    except FileNotFoundError:
+        log.warning(f"Config file {CONFIG_FILE} not found, using defaults.")
+        return {}
+    except json.JSONDecodeError:
+        log.error(f"Config file {CONFIG_FILE} is not a valid JSON file.")
+        raise
+
+
+def _add_from_env_vars(target: dict, mapping: dict) -> None:
+    """Recursively construct configuration dict from environment variables.
+
+    Recursively populate `target` from `mapping` where mapping leaf values are environment variable
+    names. If the env var is present, the corresponding key is set in `target`.
+
+    Example mapping:
+    {
+        "db": {"user": "CMG_DB_USER", "host": "CMG_DB_HOST"}
+    }
+    """
+    for key, val in mapping.items():
+        if isinstance(val, dict):
+            sub = target.setdefault(key, {})
+            _add_from_env_vars(sub, val)
+        else:
+            if (env_val := os.getenv(val)) is not None:
+                target[key] = env_val
+
+
+def _create_config_from_env_vars(env_map: dict) -> dict:
+    """Create a configuration dictionary from environment variables."""
+    config: dict = {}
+    _add_from_env_vars(config, env_map)
+    return config
+
+
+def _load_env_vars() -> dict:
+    """Load configuration dict from environment variables, skipping None values."""
+    load_dotenv()
+
+    env_map = {
+        "db": {
+            "user": "CMG_DB_USER",
+            "password": "CMG_DB_PASSWORD",
+            "name": "CMG_DB_NAME",
+            "host": "CMG_DB_HOST",
+            "port": "CMG_DB_PORT",
+        },
+        "object_store": {
+            "host": "CMG_OBJECT_STORE_HOST",
+            "port": "CMG_OBJECT_STORE_PORT",
+            "access_key": "CMG_OBJECT_STORE_ACCESS_KEY",
+            "secret_key": "CMG_OBJECT_STORE_SECRET_KEY",
+            "bucket_tasks": "CMG_OBJECT_STORE_BUCKET_TASKS",
+            "bucket_results": "CMG_OBJECT_STORE_BUCKET_RESULTS",
+        },
+        "queue": {
+            "user": "CMG_QUEUE_USER",
+            "password": "CMG_QUEUE_PASSWORD",
+            "name": "CMG_QUEUE_NAME",
+            "host": "CMG_QUEUE_HOST",
+            "port": "CMG_QUEUE_PORT",
+        },
+        "scheduler": {
+            "max_concurrent_tasks": "CMG_SCHEDULER_MAX_CONCURRENT_TASKS",
+            "metrics_port": "CMG_SCHEDULER_METRICS_PORT",
+        },
+        "ripper": {"interval": "CMG_RIPPER_INTERVAL", "metrics_port": "CMG_RIPPER_METRICS_PORT"},
+    }
+
+    return _create_config_from_env_vars(env_map)
+
+
+def load_config() -> Config:
+    """Load and validate configuration from JSON file and environment variables.
+
+    This function:
+    1. Loads JSON configuration for tasks and models
+    2. Loads environment variables for runtime services
+    3. Merges them into a unified structure
+    4. Validates everything with Pydantic schema
+    5. Caches the result for subsequent calls
+
+    Returns:
+        Config: Fully validated configuration object
+
+    Raises:
+        ValidationError: If configuration doesn't match schema
+        JSONDecodeError: If config file is not valid JSON
+    """
+    global _config_instance
+
+    if _config_instance is None:
+        json_config, env_config = _load_json_config(), _load_env_vars()
+
+        merged_config = {
+            **env_config,
+            "cms": json_config.get("cms", {}),
+            "models": json_config.get("models", {}),
+            "labels": json_config.get("labels", {}),
+            # Only include tracking if explicitly provided in JSON
+            # Otherwise, the validator will copy from cms.tracking
+            **({"tracking": json_config["tracking"]} if "tracking" in json_config else {}),
+        }
+
+        try:
+            _config_instance = Config.model_validate(merged_config)
+            log.debug(f"Loaded config: {_config_instance.model_dump_json()}")
+        except ValidationError as e:
+            log.error(f"Configuration validation failed: {e}")
+            raise
+
+    return _config_instance
+
+
+def get_config() -> Config:
+    """Get the current configuration instance."""
+    if _config_instance is None:
+        raise RuntimeError("Config not initialized. Call load_config() first.")
+    return _config_instance