[DOP-22336] Add logic for handling WebDAV transfers (#194)

IlyasDevelopment · web-flow · commit 49273f2e7b96 · 2025-02-03T13:48:45.000+03:00
diff --git a/.env.docker b/.env.docker
@@ -142,6 +142,13 @@ TEST_SAMBA_USER=syncmaster
 TEST_SAMBA_PASSWORD=test_only
 TEST_SAMBA_AUTH_TYPE=NTLMv2
 
+TEST_WEBDAV_HOST_FOR_CONFTEST=webdav
+TEST_WEBDAV_PORT_FOR_CONFTEST=80
+TEST_WEBDAV_HOST_FOR_WORKER=webdav
+TEST_WEBDAV_PORT_FOR_WORKER=80
+TEST_WEBDAV_USER=syncmaster
+TEST_WEBDAV_PASSWORD=test_only
+
 SPARK_CONF_DIR=/app/tests/spark/hive/conf/
 HADOOP_CONF_DIR=/app/tests/spark/hadoop/
 HIVE_CONF_DIR=/app/tests/spark/hive/conf/
diff --git a/.env.local b/.env.local
@@ -129,6 +129,13 @@ export TEST_SAMBA_USER=syncmaster
 export TEST_SAMBA_PASSWORD=test_only
 export TEST_SAMBA_AUTH_TYPE=NTLMv2
 
+export TEST_WEBDAV_HOST_FOR_CONFTEST=localhost
+export TEST_WEBDAV_PORT_FOR_CONFTEST=8010
+export TEST_WEBDAV_HOST_FOR_WORKER=webdav
+export TEST_WEBDAV_PORT_FOR_WORKER=80
+export TEST_WEBDAV_USER=syncmaster
+export TEST_WEBDAV_PASSWORD=test_only
+
 export SPARK_CONF_DIR=./tests/spark/hive/conf/
 export HADOOP_CONF_DIR=./tests/spark/hadoop/
 export HIVE_CONF_DIR=./tests/spark/hive/conf/
diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
@@ -60,6 +60,10 @@ jobs:
     name: Samba tests
     uses: ./.github/workflows/samba-tests.yml
 
+  webdav_tests:
+    name: WebDAV tests
+    uses: ./.github/workflows/webdav-tests.yml
+
   scheduler_tests:
     name: Scheduler tests
     uses: ./.github/workflows/scheduler-tests.yml
@@ -72,7 +76,7 @@ jobs:
     name: Tests done
     runs-on: ubuntu-latest
 
-    needs: [unit_tests, scheduler_tests, oracle_tests, clickhouse_tests, mssql_tests, mysql_tests, hive_tests, hdfs_tests, s3_tests, sftp_tests, ftp_tests, ftps_tests, samba_tests]
+    needs: [unit_tests, scheduler_tests, oracle_tests, clickhouse_tests, mssql_tests, mysql_tests, hive_tests, hdfs_tests, s3_tests, sftp_tests, ftp_tests, ftps_tests, samba_tests, webdav_tests]
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
diff --git a/.github/workflows/webdav-tests.yml b/.github/workflows/webdav-tests.yml
@@ -0,0 +1,79 @@
+name: WebDAV tests
+on:
+  workflow_call:
+
+env:
+  DEFAULT_PYTHON: '3.12'
+
+jobs:
+  test:
+    name: Run WebDAV tests
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Cache jars
+        uses: actions/cache@v4
+        with:
+          path: ./cached_jars
+          key: ${{ runner.os }}-python-${{ env.DEFAULT_PYTHON }}-test-webdav
+          restore-keys: |
+            ${{ runner.os }}-python-${{ env.DEFAULT_PYTHON }}-test-webdav
+            ${{ runner.os }}-python-
+
+      - name: Build Worker Image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          tags: mtsrus/syncmaster-worker:${{ github.sha }}
+          target: test
+          file: docker/Dockerfile.worker
+          load: true
+          cache-from: mtsrus/syncmaster-worker:develop
+
+      - name: Docker compose up
+        run: |
+          docker compose -f docker-compose.test.yml --profile all down -v --remove-orphans
+          docker compose -f docker-compose.test.yml --profile webdav up -d --wait --wait-timeout 200
+        env:
+          WORKER_IMAGE_TAG: ${{ github.sha }}
+
+      - name: Run WebDAV Tests
+        run: |
+          docker compose -f ./docker-compose.test.yml --profile webdav exec -T worker coverage run -m pytest -vvv -s -m "worker and webdav"
+
+      - name: Dump worker logs on failure
+        if: failure()
+        uses: jwalton/gh-docker-logs@v2
+        with:
+          images: mtsrus/syncmaster-worker
+          dest: ./logs
+
+      # This is important, as coverage is exported after receiving SIGTERM
+      - name: Shutdown
+        if: always()
+        run: |
+          docker compose -f docker-compose.test.yml --profile all down -v --remove-orphans
+
+      - name: Upload worker logs
+        uses: actions/upload-artifact@v4
+        if: failure()
+        with:
+          name: worker-logs-webdav
+          path: logs/*
+
+      - name: Upload coverage results
+        uses: actions/upload-artifact@v4
+        with:
+          name: coverage-webdav
+          path: reports/*
+          # https://github.com/actions/upload-artifact/issues/602
+          include-hidden-files: true
diff --git a/Makefile b/Makefile
@@ -125,6 +125,10 @@ test-integration-samba: test-db ##@Test           Run integration tests for Samb
 	docker compose -f docker-compose.test.yml --profile samba up -d --wait $(DOCKER_COMPOSE_ARGS)
 	${POETRY} run pytest ./tests/test_integration -m samba $(PYTEST_ARGS)
 
+test-integration-webdav: test-db ##@Test           Run integration tests for WebDAV
+	docker compose -f docker-compose.test.yml --profile webdav up -d --wait $(DOCKER_COMPOSE_ARGS)
+	${POETRY} run pytest ./tests/test_integration -m webdav $(PYTEST_ARGS)
+
 test-integration: test-db ##@Test           Run all integration tests
 	docker compose -f docker-compose.test.yml --profile all up -d --wait $(DOCKER_COMPOSE_ARGS)
 	${POETRY} run pytest ./tests/test_integration $(PYTEST_ARGS)
diff --git a/README.rst b/README.rst
@@ -44,6 +44,7 @@ List of currently supported connections:
 * FTPS
 * SFTP
 * Samba
+* WebDAV
 
 Current Data.SyncMaster implementation provides following components:
 
diff --git a/docker-compose.test.yml b/docker-compose.test.yml
@@ -125,7 +125,7 @@ services:
         condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
-    profiles: [worker, scheduler, s3, oracle, hdfs, hive, clickhouse, mysql, mssql, sftp, ftp, ftps, samba, all]
+    profiles: [worker, scheduler, s3, oracle, hdfs, hive, clickhouse, mysql, mssql, sftp, ftp, ftps, samba, webdav, all]
 
   test-postgres:
     image: postgres
@@ -139,7 +139,7 @@ services:
       interval: 30s
       timeout: 5s
       retries: 3
-    profiles: [s3, oracle, clickhouse, mysql, mssql, hdfs, hive, sftp, ftp, ftps, samba, all]
+    profiles: [s3, oracle, clickhouse, mysql, mssql, hdfs, hive, sftp, ftp, ftps, samba, webdav, all]
 
   test-s3:
     image: bitnami/minio:latest
@@ -209,7 +209,6 @@ services:
     platform: linux/amd64
     profiles: [mysql, all]
 
-
   metastore-hive:
     image: postgres
     restart: unless-stopped
@@ -225,7 +224,7 @@ services:
       interval: 30s
       timeout: 5s
       retries: 3
-    profiles: [hive, hdfs, s3, sftp, ftp, ftps, samba, all]
+    profiles: [hive, hdfs, s3, sftp, ftp, ftps, samba, webdav, all]
 
   keycloak:
     image: quay.io/keycloak/keycloak:latest
@@ -264,7 +263,7 @@ services:
       HIVE_METASTORE_DB_USER: test_hive
       HIVE_METASTORE_DB_PASSWORD: test_hive
     # writing spark dataframe to s3, sftp, ftp, ftps xml file fails without running hive metastore server
-    profiles: [hive, hdfs, s3, sftp, ftp, ftps, samba, all]
+    profiles: [hive, hdfs, s3, sftp, ftp, ftps, samba, webdav, all]
 
   test-sftp:
     image: ${SFTP_IMAGE:-linuxserver/openssh-server}
@@ -327,6 +326,21 @@ services:
     entrypoint: [/custom_entrypoint.sh]
     profiles: [samba, all]
 
+  webdav:
+    image: ${WEBDAV_IMAGE:-chonjay21/webdav:latest}
+    restart: unless-stopped
+    environment:
+      - APP_USER_NAME=syncmaster
+      - APP_USER_PASSWD=test_only
+      - APP_UID=1000
+      - APP_GID=1000
+    ports:
+      - 8010:80
+    volumes:
+      # Remove after https://github.com/chonjay21/docker-webdav/pull/3
+      - ./docker/webdav/on_post_init.sh:/sources/webdav/eventscripts/on_post_init.sh
+    profiles: [webdav, all]
+
 volumes:
   postgres_test_data:
   rabbitmq_test_data:
diff --git a/docker/webdav/on_post_init.sh b/docker/webdav/on_post_init.sh
@@ -0,0 +1,5 @@
+#!/usr/bin/env bash
+set -e
+
+# allow create files and directories
+chown -R www-data:www-data /var/webdav
diff --git a/docs/changelog/next_release/194.feature.rst b/docs/changelog/next_release/194.feature.rst
@@ -0,0 +1 @@
+Add logic for handling WebDAV transfers
diff --git a/syncmaster/dto/connections.py b/syncmaster/dto/connections.py
@@ -132,3 +132,13 @@ class SambaConnectionDTO(ConnectionDTO):
     domain: str = ""
     port: int | None = None
     type: ClassVar[str] = "samba"
+
+
+@dataclass
+class WebDAVConnectionDTO(ConnectionDTO):
+    host: str
+    port: int
+    user: str
+    password: str
+    protocol: Literal["http", "https"] = "https"
+    type: ClassVar[str] = "webdav"
diff --git a/syncmaster/dto/transfers.py b/syncmaster/dto/transfers.py
@@ -115,3 +115,8 @@ class FTPSTransferDTO(FileTransferDTO):
 @dataclass
 class SambaTransferDTO(FileTransferDTO):
     type: ClassVar[str] = "samba"
+
+
+@dataclass
+class WebDAVTransferDTO(FileTransferDTO):
+    type: ClassVar[str] = "webdav"
diff --git a/syncmaster/worker/controller.py b/syncmaster/worker/controller.py
@@ -18,6 +18,7 @@
     S3ConnectionDTO,
     SambaConnectionDTO,
     SFTPConnectionDTO,
+    WebDAVConnectionDTO,
 )
 from syncmaster.dto.transfers import (
     ClickhouseTransferDTO,
@@ -32,6 +33,7 @@
     S3TransferDTO,
     SambaTransferDTO,
     SFTPTransferDTO,
+    WebDAVTransferDTO,
 )
 from syncmaster.exceptions.connection import ConnectionTypeNotRecognizedError
 from syncmaster.worker.handlers.base import Handler
@@ -47,6 +49,7 @@
 from syncmaster.worker.handlers.file.s3 import S3Handler
 from syncmaster.worker.handlers.file.samba import SambaHandler
 from syncmaster.worker.handlers.file.sftp import SFTPHandler
+from syncmaster.worker.handlers.file.webdav import WebDAVHandler
 from syncmaster.worker.settings import WorkerAppSettings
 
 logger = logging.getLogger(__name__)
@@ -113,6 +116,11 @@
         SambaConnectionDTO,
         SambaTransferDTO,
     ),
+    "webdav": (
+        WebDAVHandler,
+        WebDAVConnectionDTO,
+        WebDAVTransferDTO,
+    ),
 }
 
 
diff --git a/syncmaster/worker/handlers/file/webdav.py b/syncmaster/worker/handlers/file/webdav.py
@@ -0,0 +1,31 @@
+# SPDX-FileCopyrightText: 2023-2024 MTS PJSC
+# SPDX-License-Identifier: Apache-2.0
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from onetl.connection import SparkLocalFS, WebDAV
+
+from syncmaster.dto.connections import WebDAVConnectionDTO
+from syncmaster.worker.handlers.file.protocol import FileProtocolHandler
+
+if TYPE_CHECKING:
+    from pyspark.sql import SparkSession
+
+
+class WebDAVHandler(FileProtocolHandler):
+    connection_dto: WebDAVConnectionDTO
+
+    def connect(self, spark: SparkSession) -> None:
+        self.connection = WebDAV(
+            host=self.connection_dto.host,
+            port=self.connection_dto.port,
+            protocol=self.connection_dto.protocol,
+            user=self.connection_dto.user,
+            password=self.connection_dto.password,
+            ssl_verify=False,
+        ).check()
+        self.local_connection = SparkLocalFS(
+            spark=spark,
+        ).check()
diff --git a/syncmaster/worker/spark.py b/syncmaster/worker/spark.py
@@ -64,7 +64,7 @@ def get_packages(connection_type: str) -> list[str]:
         spark_version = pyspark.__version__
         return SparkS3.get_packages(spark_version=spark_version) + file_formats_spark_packages
 
-    if connection_type in ("hdfs", "sftp", "ftp", "ftps", "samba"):
+    if connection_type in ("hdfs", "sftp", "ftp", "ftps", "samba", "webdav"):
         return file_formats_spark_packages
 
     # If the database type does not require downloading .jar packages
diff --git a/tests/settings.py b/tests/settings.py
@@ -98,6 +98,14 @@ class TestSettings(BaseSettings):
     TEST_SAMBA_PASSWORD: str
     TEST_SAMBA_AUTH_TYPE: str = "NTLMv2"
 
+    TEST_WEBDAV_HOST_FOR_CONFTEST: str
+    TEST_WEBDAV_PORT_FOR_CONFTEST: int
+    TEST_WEBDAV_HOST_FOR_WORKER: str
+    TEST_WEBDAV_PORT_FOR_WORKER: int
+    TEST_WEBDAV_PROTOCOL: str = "http"
+    TEST_WEBDAV_USER: str
+    TEST_WEBDAV_PASSWORD: str
+
     @model_validator(mode="before")
     def check_sid_and_service_name(cls, values):
         sid = values.get("TEST_ORACLE_SID")
diff --git a/tests/test_integration/test_run_transfer/connection_fixtures/__init__.py b/tests/test_integration/test_run_transfer/connection_fixtures/__init__.py
@@ -117,3 +117,13 @@
 from tests.test_integration.test_run_transfer.connection_fixtures.spark_fixtures import (
     spark,
 )
+from tests.test_integration.test_run_transfer.connection_fixtures.webdav_fixtures import (
+    prepare_webdav,
+    webdav_connection,
+    webdav_file_connection,
+    webdav_file_connection_with_path,
+    webdav_file_df_connection,
+    webdav_file_df_connection_with_path,
+    webdav_for_conftest,
+    webdav_for_worker,
+)
diff --git a/tests/test_integration/test_run_transfer/connection_fixtures/spark_fixtures.py b/tests/test_integration/test_run_transfer/connection_fixtures/spark_fixtures.py
@@ -75,7 +75,7 @@ def spark(settings: Settings, request: FixtureRequest) -> SparkSession:
             )
         )
 
-    if set(markers).intersection({"hdfs", "s3", "sftp", "ftp", "ftps", "samba"}):
+    if set(markers).intersection({"hdfs", "s3", "sftp", "ftp", "ftps", "samba", "webdav"}):
         # excel version is hardcoded due to https://github.com/nightscape/spark-excel/issues/902
         file_formats_spark_packages: list[str] = XML.get_packages(
             spark_version=pyspark.__version__,
diff --git a/tests/test_integration/test_run_transfer/connection_fixtures/webdav_fixtures.py b/tests/test_integration/test_run_transfer/connection_fixtures/webdav_fixtures.py
diff --git a/tests/test_integration/test_run_transfer/test_webdav.py b/tests/test_integration/test_run_transfer/test_webdav.py

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+Add logic for handling WebDAV transfers`