biocypher 0.5.40__tar.gz → 0.5.41__tar.gz

{biocypher-0.5.40 → biocypher-0.5.41}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: biocypher
-Version: 0.5.40
+Version: 0.5.41
 Summary: A unifying framework for biomedical research knowledge graphs
 Home-page: https://github.com/biocypher/biocypher
 License: MIT

{biocypher-0.5.40 → biocypher-0.5.41}/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 @@ import neo4j_utils
 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 @@ class _Neo4jDriver:
 
         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-0.5.40 → biocypher-0.5.41}/biocypher/_metadata.py

@@ -19,7 +19,7 @@ import importlib.metadata
 
 import toml
 
-_VERSION = "0.5.40"
+_VERSION = "0.5.41"
 
 
 def get_metadata():

{biocypher-0.5.40 → biocypher-0.5.41}/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-0.5.40 → biocypher-0.5.41}/biocypher/write/_write.py

@@ -25,7 +25,7 @@ from typing import TYPE_CHECKING
 
 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-0.5.40 → biocypher-0.5.41}/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
@@ -22,6 +24,19 @@ 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.
@@ -29,6 +44,7 @@ class _Neo4jBatchWriter(_BatchWriter):
         Returns:
             str: The default location for the neo4j admin import location
         """
+
         return "bin/"
 
     def _write_array_string(self, string_list):
@@ -263,9 +279,32 @@ class _Neo4jBatchWriter(_BatchWriter):
         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 @@ class _Neo4jBatchWriter(_BatchWriter):
             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 @@ class _Neo4jBatchWriter(_BatchWriter):
         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

{biocypher-0.5.40 → biocypher-0.5.41}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "biocypher"
-version = "0.5.40"
+version = "0.5.41"
 description = "A unifying framework for biomedical research knowledge graphs"
 authors = [
     "Sebastian Lobentanzer <sebastian.lobentanzer@gmail.com>",