Test for ocrd-network

Dockerfile

@@ -44,9 +44,11 @@ WORKDIR /data
 CMD ["/usr/local/bin/ocrd", "--help"]
 
 FROM ocrd_core_base as ocrd_core_test
+# Optionally skip make assets with this arg
+ARG SKIP_ASSETS
 WORKDIR /build-ocrd
 COPY Makefile .
-RUN make assets
+RUN if test -z "$SKIP_ASSETS" || test $SKIP_ASSETS -eq 0 ; then make assets ; fi
 COPY tests ./tests
 COPY .gitmodules .
 COPY requirements_test.txt .

tests/network/test_ocrd_all_workflow.py

@@ -0,0 +1,22 @@
+from src.ocrd_network.models import StateEnum
+from tests.network.config import test_config
+from tests.network.utils import (
+    poll_job_till_timeout_fail_or_success,
+    post_ps_workflow_request,
+)
+
+PROCESSING_SERVER_URL = test_config.PROCESSING_SERVER_URL
+
+
+def test_ocrd_all_workflow():
+    # This test is supposed to run with ocrd_all not with just core on its own
+    # Note: the used workflow path is volume mapped
+    path_to_wf = "/ocrd-data/assets/ocrd_all-test-workflow.txt"
+    path_to_mets = "/data/mets.xml"
+    wf_job_id = post_ps_workflow_request(PROCESSING_SERVER_URL, path_to_wf, path_to_mets)
+    job_state = poll_job_till_timeout_fail_or_success(
+        test_url=f"{PROCESSING_SERVER_URL}/workflow/job-simple/{wf_job_id}",
+        tries=30,
+        wait=10
+    )
+    assert job_state == StateEnum.success

tests/network/test_processing_server.py

@@ -1,70 +1,50 @@
-from time import sleep
-from requests import get, post
+from requests import get
 from src.ocrd_network import NETWORK_AGENT_WORKER
 from src.ocrd_network.models import StateEnum
 from tests.base import assets
 from tests.network.config import test_config
+from tests.network.utils import (
+    poll_job_till_timeout_fail_or_success,
+    post_ps_processing_request,
+    post_ps_workflow_request,
+)
 
 PROCESSING_SERVER_URL = test_config.PROCESSING_SERVER_URL
 
 
-def poll_till_timeout_fail_or_success(test_url: str, tries: int, wait: int) -> StateEnum:
-    job_state = StateEnum.unset
-    while tries > 0:
-        sleep(wait)
-        response = get(url=test_url)
-        assert response.status_code == 200, f"Processing server: {test_url}, {response.status_code}"
-        job_state = response.json()["state"]
-        if job_state == StateEnum.success or job_state == StateEnum.failed:
-            break
-        tries -= 1
-    return job_state
-
-
 def test_processing_server_connectivity():
-    test_url = f'{PROCESSING_SERVER_URL}/'
+    test_url = f"{PROCESSING_SERVER_URL}/"
     response = get(test_url)
     assert response.status_code == 200, \
-        f'Processing server is not reachable on: {test_url}, {response.status_code}'
-    message = response.json()['message']
-    assert message.startswith('The home page of'), \
-        f'Processing server home page message is corrupted'
+        f"Processing server is not reachable on: {test_url}, {response.status_code}"
+    message = response.json()["message"]
+    assert message.startswith("The home page of"), \
+        f"Processing server home page message is corrupted"
 
 
 # TODO: The processing workers are still not registered when deployed separately.
 #  Fix that by extending the processing server.
 def test_processing_server_deployed_processors():
-    test_url = f'{PROCESSING_SERVER_URL}/processor'
+    test_url = f"{PROCESSING_SERVER_URL}/processor"
     response = get(test_url)
     processors = response.json()
     assert response.status_code == 200, \
-        f'Processing server: {test_url}, {response.status_code}'
-    assert processors == [], f'Mismatch in deployed processors'
+        f"Processing server: {test_url}, {response.status_code}"
+    assert processors == [], f"Mismatch in deployed processors"
 
 
 def test_processing_server_processing_request():
-    path_to_mets = assets.path_to('kant_aufklaerung_1784/data/mets.xml')
+    path_to_mets = assets.path_to("kant_aufklaerung_1784/data/mets.xml")
     test_processing_job_input = {
         "path_to_mets": path_to_mets,
-        "input_file_grps": ['OCR-D-IMG'],
-        "output_file_grps": ['OCR-D-DUMMY'],
+        "input_file_grps": ["OCR-D-IMG"],
+        "output_file_grps": ["OCR-D-DUMMY"],
         "agent_type": NETWORK_AGENT_WORKER,
         "parameters": {}
     }
-    test_processor = 'ocrd-dummy'
-    test_url = f'{PROCESSING_SERVER_URL}/processor/run/{test_processor}'
-    response = post(
-        url=test_url,
-        headers={"accept": "application/json"},
-        json=test_processing_job_input
-    )
-    # print(response.json())
-    assert response.status_code == 200, \
-        f'Processing server: {test_url}, {response.status_code}'
-    processing_job_id = response.json()["job_id"]
-    assert processing_job_id
-
-    job_state = poll_till_timeout_fail_or_success(
+    test_processor = "ocrd-dummy"
+    processing_job_id = post_ps_processing_request(PROCESSING_SERVER_URL, test_processor, test_processing_job_input)
+    job_state = poll_job_till_timeout_fail_or_success(
         test_url=f"{PROCESSING_SERVER_URL}/processor/job/{processing_job_id}",
         tries=10,
         wait=10
@@ -76,20 +56,8 @@ def test_processing_server_workflow_request():
     # Note: the used workflow path is volume mapped
     path_to_dummy_wf = "/ocrd-data/assets/dummy-workflow.txt"
     path_to_mets = assets.path_to('kant_aufklaerung_1784/data/mets.xml')
-
-    # submit the workflow job
-    test_url = f"{PROCESSING_SERVER_URL}/workflow/run?mets_path={path_to_mets}&page_wise=True"
-    response = post(
-        url=test_url,
-        headers={"accept": "application/json"},
-        files={"workflow": open(path_to_dummy_wf, 'rb')}
-    )
-    # print(response.json())
-    assert response.status_code == 200, f"Processing server: {test_url}, {response.status_code}"
-    wf_job_id = response.json()["job_id"]
-    assert wf_job_id
-
-    job_state = poll_till_timeout_fail_or_success(
+    wf_job_id = post_ps_workflow_request(PROCESSING_SERVER_URL, path_to_dummy_wf, path_to_mets)
+    job_state = poll_job_till_timeout_fail_or_success(
         test_url=f"{PROCESSING_SERVER_URL}/workflow/job-simple/{wf_job_id}",
         tries=30,
         wait=10

tests/network/utils.py

@@ -0,0 +1,50 @@
+from requests import get, post
+from time import sleep
+from src.ocrd_network.models import StateEnum
+
+
+def poll_job_till_timeout_fail_or_success(
+    test_url: str,
+    tries: int = 10,
+    wait: int = 10
+) -> StateEnum:
+    job_state = StateEnum.unset
+    while tries > 0:
+        sleep(wait)
+        response = get(url=test_url)
+        assert response.status_code == 200, f"Processing server: {test_url}, {response.status_code}"
+        job_state = response.json()["state"]
+        if job_state == StateEnum.success or job_state == StateEnum.failed:
+            break
+        tries -= 1
+    return job_state
+
+
+def post_ps_processing_request(ps_server_host: str, test_processor: str, test_job_input: dict) -> str:
+    test_url = f"{ps_server_host}/processor/run/{test_processor}"
+    response = post(
+        url=test_url,
+        headers={"accept": "application/json"},
+        json=test_job_input
+    )
+    # print(response.json())
+    assert response.status_code == 200, \
+        f"Processing server: {test_url}, {response.status_code}"
+    processing_job_id = response.json()["job_id"]
+    assert processing_job_id
+    return processing_job_id
+
+
+# TODO: Can be extended to include other parameters such as page_wise
+def post_ps_workflow_request(ps_server_host: str, path_to_test_wf: str, path_to_test_mets: str) -> str:
+    test_url = f"{ps_server_host}/workflow/run?mets_path={path_to_test_mets}&page_wise=True"
+    response = post(
+        url=test_url,
+        headers={"accept": "application/json"},
+        files={"workflow": open(path_to_test_wf, "rb")}
+    )
+    # print(response.json())
+    assert response.status_code == 200, f"Processing server: {test_url}, {response.status_code}"
+    wf_job_id = response.json()["job_id"]
+    assert wf_job_id
+    return wf_job_id