More explicit error message when corpus_id is not set in environment

To avoid the ugly 500 errors that env get when they try to use API endpoints with corpus_id=None.

There already was a warning but this one is more explicit.

arkindex_worker/worker/base.py

@@ -140,7 +140,7 @@ class BaseWorker(object):
         self.process_information = None
         # corpus_id will be updated in configure() using the worker_run's corpus
         # or in configure_for_developers() from the environment
-        self.corpus_id = None
+        self._corpus_id = None
         self.user_configuration = {}
         self.model_configuration = {}
         self.support_cache = support_cache
@@ -155,6 +155,17 @@ class BaseWorker(object):
         # Define API Client
         self.setup_api_client()
 
+    @property
+    def corpus_id(self) -> str:
+        """
+        ID of the corpus on which the worker is executed.
+        Has to be set through the `ARKINDEX_CORPUS_ID` variable in **read-only** mode.
+        Raises an Exception when trying to access when unset.
+        """
+        if not self._corpus_id:
+            raise Exception("Missing ARKINDEX_CORPUS_ID environment variable")
+        return self._corpus_id
+
     @property
     def is_read_only(self) -> bool:
         """
@@ -199,11 +210,7 @@ class BaseWorker(object):
             logger.warning("Running without any extra configuration")
 
         # Define corpus_id from environment
-        self.corpus_id = os.environ.get("ARKINDEX_CORPUS_ID")
-        if not self.corpus_id:
-            logger.warning(
-                "'ARKINDEX_CORPUS_ID' was not set in the environment. Any API request involving a `corpus_id` will fail."
-            )
+        self._corpus_id = os.environ.get("ARKINDEX_CORPUS_ID")
 
         # Define model_version_id from environment
         self.model_version_id = os.environ.get("ARKINDEX_MODEL_VERSION_ID")
@@ -229,7 +236,7 @@ class BaseWorker(object):
         self.process_information = worker_run["process"]
 
         # Load corpus id
-        self.corpus_id = worker_run["process"]["corpus"]
+        self._corpus_id = worker_run["process"]["corpus"]
 
         # Load worker version information
         worker_version = worker_run["worker_version"]