biocypher 0.5.36__py3-none-any.whl → 0.5.37__py3-none-any.whl

biocypher/_metadata.py

@@ -19,7 +19,7 @@ import importlib.metadata
 
 import toml
 
-_VERSION = "0.5.36"
+_VERSION = "0.5.37"
 
 
 def get_metadata():

biocypher/_write.py

@@ -13,6 +13,7 @@ BioCypher 'offline' module. Handles the writing of node and edge representations
 suitable for import into a DBMS.
 """
 
+import re
 import glob
 
 from ._logger import logger
@@ -22,7 +23,6 @@ logger.debug(f"Loading module {__name__}.")
 from abc import ABC, abstractmethod
 from types import GeneratorType
 from typing import TYPE_CHECKING, Union, Optional
-from datetime import datetime
 from collections import OrderedDict, defaultdict
 import os
 
@@ -34,7 +34,6 @@ from ._create import BioCypherEdge, BioCypherNode, BioCypherRelAsNode
 __all__ = ["get_writer"]
 
 if TYPE_CHECKING:
-    from ._ontology import Ontology
     from ._translate import Translator
     from ._deduplicate import Deduplicator
 
@@ -954,7 +953,9 @@ class _BatchWriter(ABC):
             bool: The return value. True for success, False otherwise.
         """
         # translate label to PascalCase
-        label_pascal = self.translator.name_sentence_to_pascal(label)
+        label_pascal = self.translator.name_sentence_to_pascal(
+            parse_label(label)
+        )
 
         # list files in self.outdir
         files = glob.glob(
@@ -1086,7 +1087,9 @@ class _Neo4jBatchWriter(_BatchWriter):
             _id = ":ID"
 
             # translate label to PascalCase
-            pascal_label = self.translator.name_sentence_to_pascal(label)
+            pascal_label = self.translator.name_sentence_to_pascal(
+                parse_label(label)
+            )
 
             header = f"{pascal_label}-header.csv"
             header_path = os.path.join(
@@ -1165,7 +1168,9 @@ class _Neo4jBatchWriter(_BatchWriter):
 
         for label, props in self.edge_property_dict.items():
             # translate label to PascalCase
-            pascal_label = self.translator.name_sentence_to_pascal(label)
+            pascal_label = self.translator.name_sentence_to_pascal(
+                parse_label(label)
+            )
 
             # paths
             header = f"{pascal_label}-header.csv"
@@ -1310,6 +1315,43 @@ class _Neo4jBatchWriter(_BatchWriter):
         return import_call
 
 
+def parse_label(label: str) -> str:
+    """
+
+    Check if the label is compliant with Neo4j naming conventions,
+    https://neo4j.com/docs/cypher-manual/current/syntax/naming/, and if not,
+    remove non-compliant characters.
+
+    Args:
+        label (str): The label to check
+    Returns:
+        str: The compliant label
+    """
+    # Check if the name contains only alphanumeric characters, underscore, or dollar sign
+    # and dot (for class hierarchy of BioCypher)
+    allowed_chars = r"a-zA-Z0-9_$ ."
+    matches = re.findall(f"[{allowed_chars}]", label)
+    non_matches = re.findall(f"[^{allowed_chars}]", label)
+    if non_matches:
+        non_matches = list(set(non_matches))
+        logger.warning(
+            f"Label is not compliant with Neo4j naming rules. Removed non compliant characters: {non_matches}"
+        )
+
+    def first_character_compliant(character: str) -> bool:
+        return character.isalpha() or character == "$"
+
+    if not first_character_compliant(matches[0]):
+        for c in matches:
+            if first_character_compliant(c):
+                matches = matches[matches.index(c) :]
+                break
+        logger.warning(
+            "Label does not start with an alphabetic character or with $. Removed non compliant characters."
+        )
+    return "".join(matches).strip()
+
+
 class _ArangoDBBatchWriter(_Neo4jBatchWriter):
     """
     Class for writing node and edge representations to disk using the format

{biocypher-0.5.36.dist-info → biocypher-0.5.37.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: biocypher
-Version: 0.5.36
+Version: 0.5.37
 Summary: A unifying framework for biomedical research knowledge graphs
 Home-page: https://github.com/biocypher/biocypher
 License: MIT

{biocypher-0.5.36.dist-info → biocypher-0.5.37.dist-info}/RECORD

@@ -12,13 +12,13 @@ biocypher/_deduplicate.py,sha256=BBvfpXzu6L5YDY5FdtXxnf8YlsbJpbCE8RdUoKsm0n0,494
 biocypher/_get.py,sha256=3Kpky3blfNf1JwxKWLsZxTU2aTP_C4sUe8OpiyYj63I,10810
 biocypher/_logger.py,sha256=NGXe3hZA79WSujfOgpcxHBf8N2QAfrmvM1LFDpsGK2U,3185
 biocypher/_mapping.py,sha256=ERSNH2Bg19145KytxbFE4BInPaiP-LWW7osOBot29Eo,9304
-biocypher/_metadata.py,sha256=7k16RMYEquR4rs2kSUGc2uJs0T-HntytZ2cZerQtsFk,1658
+biocypher/_metadata.py,sha256=-yWk0B2JDJXt34dPea_5x4t2YASfP8wc6iRVvRKdB1A,1658
 biocypher/_misc.py,sha256=g5B-PO_XJlYEJC7kEVRdCXeB2NW0ZSVr_5KqTEk2ldk,5877
 biocypher/_ontology.py,sha256=53hHroH4K9MbwueK2pAbdkidMRBVH2adlQ66QbI_BiE,26734
 biocypher/_pandas.py,sha256=GVCFM68J7yBjh40MpkNVgD8qT1RFMrrIjMOtD3iKsf4,3040
 biocypher/_translate.py,sha256=nj4Y60F0U3JBH36N2dh5pFcC8Ot86rskJ2ChJwje9dI,16494
-biocypher/_write.py,sha256=nvb75OwElu8fLUp0FjEBqQ1VNpx6iRrk-t7v_TOlDhg,68165
-biocypher-0.5.36.dist-info/LICENSE,sha256=SjUaQkq671iQUZOxEUpC4jvJxXOlfSiHTTueyz9kXJM,1065
-biocypher-0.5.36.dist-info/METADATA,sha256=0BlPiw51Ym03ecbQ4uBIDCexAU31tbRa1Iqr19GSbiE,10573
-biocypher-0.5.36.dist-info/WHEEL,sha256=Zb28QaM1gQi8f4VCBhsUklF61CTlNYfs9YAZn-TOGFk,88
-biocypher-0.5.36.dist-info/RECORD,,
+biocypher/_write.py,sha256=EjAnNzayVKvBuIVLw3gY8T9fTnfIPaTGODu275IaRJ8,69554
+biocypher-0.5.37.dist-info/LICENSE,sha256=SjUaQkq671iQUZOxEUpC4jvJxXOlfSiHTTueyz9kXJM,1065
+biocypher-0.5.37.dist-info/METADATA,sha256=iD1g7a9FH1JRfhVynse5qctOAsazZxa0_YDxr-OCDkc,10573
+biocypher-0.5.37.dist-info/WHEEL,sha256=Zb28QaM1gQi8f4VCBhsUklF61CTlNYfs9YAZn-TOGFk,88
+biocypher-0.5.37.dist-info/RECORD,,