Merge pull request biocypher#321 from mbaric758/neo4j_v5_patch

Neo4j version 5 patch

.github/actions/setup/action.yaml

@@ -28,7 +28,7 @@ runs:
       with:
         version: 1.5.1
         virtualenvs-create: true
-        virtualenvs-in-project: true
+        virtualenvs-in-project: false
     - name: Check Poetry installation
       run: poetry --version
       shell: bash

.github/actions/test/action.yaml

@@ -1,5 +1,8 @@
 name: 'Test and code quality'
 description: 'Run tests and code quality checks'
+inputs:
+  NEO4J_VERSION:
+    description: 'Neo4j version'
 
 runs:
   using: "composite"
@@ -12,7 +15,7 @@ runs:
       uses: douglascamata/setup-docker-macos-action@v1-alpha
       if: ${{ runner.os == 'macOS' }}
     - name: Start Neo4j Docker
-      run: docker run --restart always --publish=7474:7474 --publish=7687:7687 --env NEO4J_AUTH=neo4j/your_password_here --env NEO4J_PLUGINS='["apoc"]' --env=NEO4J_ACCEPT_LICENSE_AGREEMENT=yes -d neo4j:4.4-enterprise
+      run: docker run --restart always --publish=7474:7474 --publish=7687:7687 --env NEO4J_AUTH=neo4j/your_password_here --env NEO4J_PLUGINS='["apoc"]' --env=NEO4J_ACCEPT_LICENSE_AGREEMENT=yes -d neo4j:${{ inputs.NEO4J_VERSION }}
       shell: bash
       if: ${{ runner.os != 'Windows' }}
     - name: Start Postgres Docker
@@ -24,16 +27,14 @@ runs:
     #----------------------------------------------
     - name: Run Tests (Windows)
       run: |
-        source .venv/scripts/activate
-        pytest --version
-        pytest --password=your_password_here
+        poetry run pytest --version
+        poetry run pytest --password=your_password_here
       shell: bash
       if: runner.os == 'Windows'
     - name: Run tests (Linux and MacOS)
       run: |
-        source .venv/bin/activate
-        pytest --version
-        pytest --password=your_password_here
+        poetry run pytest --version
+        poetry run pytest --password=your_password_here
       shell: bash
       if: runner.os != 'Windows'
     - name: Check code quality

.github/workflows/pr.yaml

@@ -9,6 +9,11 @@ jobs:
       matrix:
         os: ["ubuntu-latest", "macos-latest", "windows-latest"]
         python-version: ["3.9", "3.10", "3.11"]
+        neo4j-version: ["4.4-enterprise", "5.17.0-enterprise"]
+        exclude:
+          # Windows tests don't use the PostgreSQL/Neo4j Docker containers
+          - os: "windows-latest"
+            neo4j-version: "5.17.0-enterprise"
     runs-on: ${{ matrix.os }}
     env:
       POETRY_VERSION: 1.5.1
@@ -34,8 +39,11 @@ jobs:
       uses: ./.github/actions/install
       with:
         PYTHON_VERSION: ${{ matrix.python-version }}
+
     #----------------------------------------------
     #      run tests and code quality checks
     #----------------------------------------------
     - name: Run tests
       uses: ./.github/actions/test
+      with:
+        NEO4J_VERSION: ${{ matrix.neo4j-version }}

biocypher/_connect.py

@@ -11,11 +11,12 @@
 """
 BioCypher 'online' mode. Handles connection and manipulation of a running DBMS.
 """
+import subprocess
+
 from ._logger import logger
 
 logger.debug(f"Loading module {__name__}.")
 
-from typing import Optional
 from collections.abc import Iterable
 import itertools
 
@@ -24,7 +25,6 @@
 from . import _misc
 from ._config import config as _config
 from ._create import BioCypherEdge, BioCypherNode
-from ._ontology import Ontology
 from ._translate import Translator
 
 __all__ = ["_Neo4jDriver"]
@@ -137,16 +137,43 @@ def _create_constraints(self):
 
         logger.info("Creating constraints for node types in config.")
 
+        major_neo4j_version = int(self._get_neo4j_version().split(".")[0])
         # get structure
         for leaf in self.translator.ontology.mapping.extended_schema.items():
-            label = _misc.sentencecase_to_pascalcase(leaf[0])
+            label = _misc.sentencecase_to_pascalcase(leaf[0], sep=r"\s\.")
             if leaf[1]["represented_as"] == "node":
-                s = (
-                    f"CREATE CONSTRAINT `{label}_id` "
-                    f"IF NOT EXISTS ON (n:`{label}`) "
-                    "ASSERT n.id IS UNIQUE"
-                )
-                self._driver.query(s)
+                if major_neo4j_version >= 5:
+                    s = (
+                        f"CREATE CONSTRAINT `{label}_id` "
+                        f"IF NOT EXISTS FOR (n:`{label}`) "
+                        "REQUIRE n.id IS UNIQUE"
+                    )
+                    self._driver.query(s)
+                else:
+                    s = (
+                        f"CREATE CONSTRAINT `{label}_id` "
+                        f"IF NOT EXISTS ON (n:`{label}`) "
+                        "ASSERT n.id IS UNIQUE"
+                    )
+                    self._driver.query(s)
+
+    def _get_neo4j_version(self):
+        """Get neo4j version."""
+        try:
+            neo4j_version = self._driver.query(
+                """
+                    CALL dbms.components()
+                    YIELD name, versions, edition
+                    UNWIND versions AS version
+                    RETURN version AS version
+                """,
+            )[0][0]["version"]
+            return neo4j_version
+        except Exception as e:
+            logger.warning(
+                f"Error detecting Neo4j version: {e} use default version 4.0.0."
+            )
+            return "4.0.0"
 
     def add_nodes(self, id_type_tuples: Iterable[tuple]) -> tuple:
         """

biocypher/_misc.py

@@ -205,7 +205,7 @@ def sentencecase_to_snakecase(s: str) -> str:
     return stringcase.snakecase(s).lower()
 
 
-def sentencecase_to_pascalcase(s: str) -> str:
+def sentencecase_to_pascalcase(s: str, sep: str = r"\s") -> str:
     """
     Convert sentence case to PascalCase.
 
@@ -215,7 +215,11 @@ def sentencecase_to_pascalcase(s: str) -> str:
     Returns:
         string in PascalCase form
     """
-    return re.sub(r"(?:^| )([a-zA-Z])", lambda match: match.group(1).upper(), s)
+    return re.sub(
+        r"(?:^|[" + sep + "])([a-zA-Z])",
+        lambda match: match.group(1).upper(),
+        s,
+    )
 
 
 def to_lower_sentence_case(s: str) -> str:

biocypher/write/_write.py

@@ -25,7 +25,7 @@
 
 from biocypher._config import config as _config
 
-__all__ = ["get_writer"]
+__all__ = ["get_writer", "DBMS_TO_CLASS"]
 
 if TYPE_CHECKING:
     from biocypher._translate import Translator

biocypher/write/graph/_neo4j.py

@@ -1,4 +1,6 @@
 import os
+import re
+import subprocess
 
 from biocypher._logger import logger
 from biocypher.write._batch_writer import parse_label, _BatchWriter
@@ -22,13 +24,27 @@ class _Neo4jBatchWriter(_BatchWriter):
         - _write_array_string
     """
 
+    def __init__(self, *args, **kwargs):
+        """
+        Constructor.
+
+        Check the version of Neo4j and adds a command scope if version >= 5.
+
+        Returns:
+            _Neo4jBatchWriter: An instance of the writer.
+        """
+
+        # Should read the configuration and setup import_call_bin_prefix.
+        super().__init__(*args, **kwargs)
+
     def _get_default_import_call_bin_prefix(self):
         """
         Method to provide the default string for the import call bin prefix.
 
         Returns:
             str: The default location for the neo4j admin import location
         """
+
         return "bin/"
 
     def _write_array_string(self, string_list):
@@ -263,9 +279,32 @@ def _construct_import_call(self) -> str:
         Returns:
             str: a bash command for neo4j-admin import
         """
+        import_call_neo4j_v4 = self._get_import_call(
+            "import", "--database=", "--force="
+        )
+        import_call_neo4j_v5 = self._get_import_call(
+            "database import full", "", "--overwrite-destination="
+        )
+        neo4j_version_check = f"version=$({self._get_default_import_call_bin_prefix()}neo4j-admin --version | cut -d '.' -f 1)"
+
+        import_script = f"#!/bin/bash\n{neo4j_version_check}\nif [[ $version -ge 5 ]]; then\n\t{import_call_neo4j_v5}\nelse\n\t{import_call_neo4j_v4}\nfi"
+        return import_script
+
+    def _get_import_call(
+        self, import_cmd: str, database_cmd: str, wipe_cmd: str
+    ) -> str:
+        """Get parametrized import call for Neo4j 4 or 5+.
+
+        Args:
+            import_cmd (str): The import command to use.
+            database_cmd (str): The database command to use.
+            wipe_cmd (str): The wipe command to use.
+
+        Returns:
+            str: The import call.
+        """
         import_call = (
-            f"{self.import_call_bin_prefix}neo4j-admin import "
-            f"--database={self.db_name} "
+            f"{self.import_call_bin_prefix}neo4j-admin {import_cmd} "
             f'--delimiter="{self.escaped_delim}" '
             f'--array-delimiter="{self.escaped_adelim}" '
         )
@@ -276,7 +315,7 @@ def _construct_import_call(self) -> str:
             import_call += f"--quote='{self.quote}' "
 
         if self.wipe:
-            import_call += f"--force=true "
+            import_call += f"{wipe_cmd}true "
         if self.skip_bad_relationships:
             import_call += "--skip-bad-relationships=true "
         if self.skip_duplicate_nodes:
@@ -290,4 +329,6 @@ def _construct_import_call(self) -> str:
         for header_path, parts_path in self.import_call_edges:
             import_call += f'--relationships="{header_path},{parts_path}" '
 
+        # Database needs to be at the end starting with Neo4j 5.0+.
+        import_call += f"{database_cmd}{self.db_name} "
         return import_call

docs/output/neo4j.md

@@ -7,13 +7,10 @@ refer the reader to the Neo4j documentation for more details.
 
 ## Install Neo4j
 
-```{note}
 Neo4j provide a [Neo4j Desktop
 application](https://neo4j.com/download-center/#desktop) that can be used to
 create a local instance of Neo4j. The desktop application provides information
-about the DBMS folder and can open a terminal at the DBMS root location. The
-import call generated by BioCypher can then be copied into the terminal of the
-Neo4j Desktop application for each corresponding DBMS.
+about the DBMS folder and can open a terminal at the DBMS root location.
 
 Neo4j is also available as a command line interface (CLI) tool. To use the CLI
 with the BioCypher admin import call, directory structures and permissions need
@@ -24,7 +21,6 @@ on your system.
 
 Be mindful that different versions of Neo4j may differ in features and thus are
 also documented differently.
-```
 
 ```{note}
 We use the APOC library for Neo4j, which is not included automatically, but
@@ -141,9 +137,14 @@ header and data files for each entity type, the import call conveniently
 aggregates this information into one command, detailing the location of all
 files on disk, so no data need to be copied around.
 
-```{caution}
-The generated import call is currently only compatible with Neo4j version 4.
-Version 5 support coming soon!
+```{note}
+The generated import call differs between Neo4j version 4 and 5.
+Starting from major version 5, Neo4j ``import`` command needs the
+``database`` scope. BioCypher takes care of this.
+The generated import script ``neo4j-admin-import-call.sh``
+first checks the Neo4j database version and uses the correct
+import statement for the detected version. Therefore make sure to run
+the script from the targeted DBMS root location.
 ```
 
 Neo4j can manage multiple projects, each with multiple DBMS (database management