Merge pull request #65 from Materials-Data-Science-and-Informatics/fix/non_ascii

json dumps wont force ascii characters

Author: mustafasoylu

.somesy.toml

@@ -1,6 +1,6 @@
 [project]
 name = "somesy"
-version = "0.2.1"
+version = "0.3.0"
 description = "A CLI tool for synchronizing software project metadata."
 keywords = ["metadata", "FAIR"]
 license = "MIT"

CITATION.cff

@@ -3,7 +3,7 @@ type: software
 message: If you use this software, please cite it using this metadata.
 
 title: somesy
-version: 0.2.1
+version: 0.3.0
 abstract: A CLI tool for synchronizing software project metadata.
 repository-code: https://github.com/Materials-Data-Science-and-Informatics/somesy
 license: MIT

codemeta.json

@@ -25,7 +25,7 @@
   ],
   "name": "somesy",
   "description": "A CLI tool for synchronizing software project metadata.",
-  "version": "0.2.1",
+  "version": "0.3.0",
   "keywords": [
     "metadata",
     "FAIR"
@@ -62,7 +62,7 @@
     {
       "@type": "Person",
       "givenName": "Jens",
-      "familyName": "Br\u00f6der",
+      "familyName": "Bröder",
       "email": "j.broeder@fz-juelich.de",
       "@id": "https://orcid.org/0000-0001-7939-226X"
     },

src/somesy/codemeta/__init__.py

@@ -1,4 +1,4 @@
-"""Integration with codemetapy (to re-generate codemeta as part of somesy sync)."""
+"""Integration with codemeta.json (to re-generate codemeta as part of somesy sync)."""
 from .writer import Codemeta
 
 __all__ = ["Codemeta"]

src/somesy/codemeta/writer.py

@@ -1,5 +1,4 @@
 """codemeta.json creation module."""
-import json
 import logging
 from collections import OrderedDict
 from pathlib import Path
@@ -9,6 +8,7 @@
 
 from somesy.core.models import Person, ProjectMetadata
 from somesy.core.writer import ProjectMetadataWriter
+from somesy.json_wrapper import json
 
 logger = logging.getLogger("somesy")
 
@@ -87,7 +87,7 @@ def _init_new_file(self) -> None:
         }
         # dump to file
         with self.path.open("w+") as f:
-            json.dump(data, f, indent=2)
+            json.dump(data, f)
 
     def save(self, path: Optional[Path] = None) -> None:
         """Save the codemeta.json file."""
@@ -107,15 +107,14 @@ def save(self, path: Optional[Path] = None) -> None:
 
         with path.open("w") as f:
             # codemeta.json indentation is 2 spaces
-            json.dump(data, f, indent=2)
+            json.dump(data, f)
 
     @staticmethod
     def _from_person(person: Person):
         """Convert project metadata person object to codemeta.json dict for person format."""
         person_dict = {
             "@type": "Person",
         }
-        logger.debug(f"Converting person {person} to codemeta.json format.")
         if person.given_names:
             person_dict["givenName"] = person.given_names
         if person.family_names:
@@ -144,7 +143,6 @@ def _to_person(person) -> Person:
             person_obj["orcid"] = person["@id"].strip()
         if "address" in person:
             person_obj["address"] = person["address"].strip()
-        logger.debug(f"Converting person {person_obj} to pydantic person instance.")
 
         return Person(**person_obj)
 

src/somesy/core/models.py

@@ -115,7 +115,7 @@ def model_dump_json(self, *args, **kwargs):
 
         if by_alias:
             ret = {self.model_fields[k].alias or k: v for k, v in ret.items()}
-        return json.dumps(ret)
+        return json.dumps(ret, ensure_ascii=False)
 
 
 _SOMESY_TARGETS = ["cff", "pyproject", "package_json", "codemeta"]

src/somesy/json_wrapper.py

@@ -0,0 +1,18 @@
+"""Utility functions for somesy."""
+import json
+
+import wrapt
+
+
+@wrapt.decorator
+def json_dump_wrapper(wrapped, instance, args, kwargs):
+    """Wrap json.dump to write non-ascii characters with default indentation."""
+    # Ensure ensure_ascii is set to False
+    kwargs["ensure_ascii"] = False
+    # set indent to 2 if not set
+    kwargs["indent"] = kwargs.get("indent", 2)
+    return wrapped(*args, **kwargs)
+
+
+# Apply the wrapper
+json.dump = json_dump_wrapper(json.dump)

src/somesy/package_json/writer.py

@@ -1,5 +1,4 @@
 """package.json parser and saver."""
-import json
 import logging
 from collections import OrderedDict
 from pathlib import Path
@@ -9,6 +8,7 @@
 
 from somesy.core.models import Person, ProjectMetadata
 from somesy.core.writer import ProjectMetadataWriter
+from somesy.json_wrapper import json
 from somesy.package_json.models import PackageJsonConfig
 
 logger = logging.getLogger("somesy")
@@ -72,7 +72,7 @@ def save(self, path: Optional[Path] = None) -> None:
 
         with path.open("w") as f:
             # package.json indentation is 2 spaces
-            json.dump(self._data, f, indent=2)
+            json.dump(self._data, f)
 
     @staticmethod
     def _from_person(person: Person):

tests/output/test_codemeta.py

@@ -1,5 +1,5 @@
 from somesy.codemeta import Codemeta
-import json
+from somesy.json_wrapper import json
 
 
 def test_update_codemeta(somesy_input, tmp_path):