Per segment chunks

.github/workflows/full.yml

@@ -165,6 +165,8 @@ jobs:
         id: run_tests
         run: |
           pytest tests/python/
+          ONE_RUNNING_JOB_IN_QUEUE_PER_USER="true" pytest tests/python/rest_api/test_queues.py
+          CVAT_ALLOW_STATIC_CACHE="true" pytest -k "TestTaskData" tests/python
 
       - name: Creating a log file from cvat containers
         if: failure() && steps.run_tests.conclusion == 'failure'

.github/workflows/main.yml

@@ -177,8 +177,9 @@ jobs:
           COVERAGE_PROCESS_START: ".coveragerc"
         run: |
           pytest tests/python/ --cov --cov-report=json
-          for COVERAGE_FILE in `find -name "coverage*.json" -type f -printf "%f\n"`; do mv ${COVERAGE_FILE} "${COVERAGE_FILE%%.*}_0.json"; done
           ONE_RUNNING_JOB_IN_QUEUE_PER_USER="true" pytest tests/python/rest_api/test_queues.py --cov --cov-report=json
+          CVAT_ALLOW_STATIC_CACHE="true" pytest -k "TestTaskData" tests/python --cov --cov-report=json
+          for COVERAGE_FILE in `find -name "coverage*.json" -type f -printf "%f\n"`; do mv ${COVERAGE_FILE} "${COVERAGE_FILE%%.*}_0.json"; done
 
       - name: Uploading code coverage results as an artifact
         uses: actions/upload-artifact@v4

.github/workflows/schedule.yml

@@ -170,6 +170,12 @@ jobs:
           pytest tests/python/
           pytest tests/python/ --stop-services
 
+          ONE_RUNNING_JOB_IN_QUEUE_PER_USER="true" pytest tests/python/rest_api/test_queues.py
+          pytest tests/python/ --stop-services
+
+          CVAT_ALLOW_STATIC_CACHE="true" pytest tests/python
+          pytest tests/python/ --stop-services
+
       - name: Unit tests
         env:
           HOST_COVERAGE_DATA_DIR: ${{ github.workspace }}

changelog.d/20240812_161617_mzhiltso_job_chunks.md

@@ -0,0 +1,24 @@
+### Added
+
+- A server setting to enable or disable storage of permanent media chunks on the server filesystem
+  (<https://github.com/cvat-ai/cvat/pull/8272>)
+- \[Server API\] `GET /api/jobs/{id}/data/?type=chunk&index=x` parameter combination.
+  The new `index` parameter allows to retrieve job chunks using 0-based index in each job,
+  instead of the `number` parameter, which used task chunk ids.
+  (<https://github.com/cvat-ai/cvat/pull/8272>)
+
+### Changed
+
+- Job assignees will not receive frames from adjacent jobs in chunks
+  (<https://github.com/cvat-ai/cvat/pull/8272>)
+
+### Deprecated
+
+- \[Server API\] `GET /api/jobs/{id}/data/?type=chunk&number=x` parameter combination
+  (<https://github.com/cvat-ai/cvat/pull/8272>)
+
+
+### Fixed
+
+- Various memory leaks in video reading on the server
+  (<https://github.com/cvat-ai/cvat/pull/8272>)