test(file_processors): send the test id on raw process_file requests

The four file-processor integration tests that upload a file started
failing under the recording harness with 400 "Test ID is required for
file ID allocation".

These tests call the endpoint with bare requests.post rather than the OGX
client, so nothing injects the provider-data header the recorder reads
the test id from. That did not matter before: a direct upload was handed
straight to the processor and no file record was ever created. Staging
the upload into the Files API means an ID now gets allocated per request,
and in record/replay mode allocation is deterministic and refuses to run
without a test id.

Adds a test_context_headers fixture following the same idiom as
tests/integration/files/test_files.py, and passes it on the four upload
requests. Outside record/replay mode get_test_context() is empty and the
fixture sends nothing, so the docling workflow is unaffected.

Signed-off-by: Charlie Doern <cdoern@redhat.com>

Author: cdoern

tests/integration/file_processors/test_file_processors.py

@@ -10,6 +10,23 @@
 import pytest
 import requests
 
+from ogx.core.testing_context import get_test_context
+
+
+@pytest.fixture
+def test_context_headers() -> dict[str, str]:
+    """Propagate the test id on raw requests.
+
+    These tests bypass the OGX client, so nothing injects the header the
+    recorder relies on. A direct upload is now staged into the Files API before
+    processing, and deterministic file-ID allocation refuses to run without a
+    test id.
+    """
+    test_id = get_test_context()
+    if not test_id:
+        return {}
+    return {"X-OGX-Provider-Data": json.dumps({"__test_id": test_id})}
+
 
 @pytest.fixture(autouse=True)
 def skip_if_no_file_processor_provider(ogx_client, require_server):
@@ -45,11 +62,12 @@ def test_pdf_content(test_pdf_path: Path) -> bytes:
 class TestFileProcessors:
     """Provider-agnostic integration tests for the file-processors/process endpoint."""
 
-    def test_process_file_basic(self, process_url, test_pdf_content):
+    def test_process_file_basic(self, process_url, test_pdf_content, test_context_headers):
         """Test basic file processing without chunking."""
         resp = requests.post(
             process_url,
             files={"file": ("test.pdf", test_pdf_content, "application/pdf")},
+            headers=test_context_headers,
             timeout=120,
         )
         assert resp.status_code == 200, f"Unexpected status: {resp.status_code} {resp.text}"
@@ -69,12 +87,13 @@ def test_process_file_basic(self, process_url, test_pdf_content):
         assert chunk["content"] is not None
         assert len(chunk["content"].strip()) > 0
 
-    def test_process_file_with_auto_chunking(self, process_url, test_pdf_content):
+    def test_process_file_with_auto_chunking(self, process_url, test_pdf_content, test_context_headers):
         """Test file processing with auto chunking strategy."""
         chunking_strategy = json.dumps({"type": "auto"})
         resp = requests.post(
             process_url,
             files={"file": ("test.pdf", test_pdf_content, "application/pdf")},
+            headers=test_context_headers,
             data={"chunking_strategy": chunking_strategy},
             timeout=120,
         )
@@ -96,7 +115,7 @@ def test_process_file_with_auto_chunking(self, process_url, test_pdf_content):
             assert "document_id" in chunk["metadata"]
             assert chunk["metadata"]["filename"] == "test.pdf"
 
-    def test_process_file_with_static_chunking(self, process_url, test_pdf_content):
+    def test_process_file_with_static_chunking(self, process_url, test_pdf_content, test_context_headers):
         """Test file processing with static chunking strategy."""
         chunking_strategy = json.dumps(
             {
@@ -110,6 +129,7 @@ def test_process_file_with_static_chunking(self, process_url, test_pdf_content):
         resp = requests.post(
             process_url,
             files={"file": ("test.pdf", test_pdf_content, "application/pdf")},
+            headers=test_context_headers,
             data={"chunking_strategy": chunking_strategy},
             timeout=120,
         )
@@ -125,12 +145,13 @@ def test_process_file_with_static_chunking(self, process_url, test_pdf_content):
             chunk_ids.add(chunk["chunk_id"])
             assert chunk["chunk_metadata"]["content_token_count"] > 0
 
-    def test_chunk_id_uniqueness(self, process_url, test_pdf_content):
+    def test_chunk_id_uniqueness(self, process_url, test_pdf_content, test_context_headers):
         """Test chunk IDs are unique across chunks."""
         chunking_strategy = json.dumps({"type": "auto"})
         resp = requests.post(
             process_url,
             files={"file": ("test.pdf", test_pdf_content, "application/pdf")},
+            headers=test_context_headers,
             data={"chunking_strategy": chunking_strategy},
             timeout=120,
         )