feat: add date-based OHIP data reporting alongside XML

arr_web/arr_data_runtime.py

@@ -0,0 +1,48 @@
+"""Explicit direct-data launch configuration. Construction makes no hotel requests."""
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Callable
+
+from arr_web.arr_data_executor import DirectARRExecutor
+from arr_web.arr_downloads import PersistentARRDownloads
+from integrations.ohip.arr_data import ARRDataSource
+from integrations.ohip.audit_arr_capture import protected_read
+from integrations.ohip.capture_job import atomic_json, fingerprint, job_lock, private_directory, sync_directory
+from integrations.ohip.collect_arr_source import require, strict_json
+
+
+@dataclass(frozen=True)
+class DirectARRSource:
+    root: Path
+    hotel_id: str
+    credential_file: Path | None = None
+    transport_factory: Callable | None = None
+    page_size: int = 100
+    max_pages: int = 100
+    max_records: int = 10000
+    max_requests: int = 25000
+
+
+def compose_direct_arr_downloads(source: DirectARRSource, processing):
+    check_schema = getattr(processing.repository, "assert_data_source_schema", None)
+    if callable(check_schema):
+        check_schema()
+    root = Path(source.root).absolute()
+    reader = ARRDataSource(root / "source", source.hotel_id, credential_file=source.credential_file,
+        transport_factory=source.transport_factory, page_size=source.page_size, max_pages=source.max_pages,
+        max_records=source.max_records, max_requests=source.max_requests)
+    executor = DirectARRExecutor(root=root / "processing", source=reader, policy=processing.policy,
+        object_store=processing.object_store, repository=processing.repository, ingestion=processing.ingestion,
+        processor=processing.processor)
+    private_directory(root)
+    sync_directory(root.parent)
+    identity = executor.source_identity()
+    with job_lock(root):
+        pin = root / "source.json"
+        if not pin.exists():
+            require(not (root / "queue").exists() and not (root / "processing").exists(),
+                    "unbound_data_runtime_state")
+            atomic_json(pin, identity, replace=False)
+        require(fingerprint(strict_json(protected_read(pin, 65536))) == fingerprint(identity),
+                "data_runtime_source_conflict")
+        return PersistentARRDownloads(root / "queue", executor)