structurize 3.6.0__tar.gz → 3.6.1__tar.gz

{structurize-3.6.0/structurize.egg-info → structurize-3.6.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: structurize
-Version: 3.6.0
+Version: 3.6.1
 Summary: Tools to convert from and to JSON Structure from various other schema languages.
 Author-email: Clemens Vasters <clemensv@microsoft.com>
 Classifier: Programming Language :: Python :: 3

{structurize-3.6.0 → structurize-3.6.1}/avrotize/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '3.6.0'
-__version_tuple__ = version_tuple = (3, 6, 0)
+__version__ = version = '3.6.1'
+__version_tuple__ = version_tuple = (3, 6, 1)
 
-__commit_id__ = commit_id = 'g1329b8c45'
+__commit_id__ = commit_id = 'ga0404973f'

{structurize-3.6.0 → structurize-3.6.1}/avrotize/jstructtoavro.py

@@ -8,7 +8,7 @@ This is the reverse operation of avrotojstruct.py.
 import json
 from typing import Any, Dict, List, Union, Optional
 
-from avrotize.common import ANY_VALUE_RECORD, any_value_name, deduplicate_any_value_record, generic_type
+from avrotize.common import ANY_VALUE_RECORD, any_value_name, deduplicate_any_value_record, generic_type, avro_name, avro_name_with_altname
 
 
 class JsonStructureToAvro:
@@ -388,10 +388,16 @@ class JsonStructureToAvro:
         
         fields = []
         for prop_name, prop_schema in properties.items():
+            # Property keys may be non-identifiers (e.g. "dr-type", "1"). Avro field
+            # names must match [A-Za-z_][A-Za-z0-9_]*, so sanitize and preserve the
+            # original wire key as an Avro alias for round-trip fidelity (issue #382).
+            avro_field_name, original_name = avro_name_with_altname(prop_name)
             field = {
-                'name': prop_name,
-                'type': self._convert_type_reference(prop_schema, name_hint=prop_name)
+                'name': avro_field_name,
+                'type': self._convert_type_reference(prop_schema, name_hint=avro_field_name)
             }
+            if original_name is not None:
+                field['aliases'] = [original_name]
             
             # Build documentation with annotations
             doc = self._build_doc_with_annotations(
@@ -418,22 +424,61 @@ class JsonStructureToAvro:
         avro_record['fields'] = fields
         return avro_record
     
+    def _sanitize_enum_symbols(self, symbols: List[Any]) -> tuple:
+        """
+        Sanitize JSON Structure enum values into valid Avro enum symbols.
+
+        Avro enum symbols must match [A-Za-z_][A-Za-z0-9_]* and be unique. JSON
+        Structure enum values may be arbitrary (e.g. "N/A", "KAR-MQTT", 1). This
+        sanitizes each value, resolves collisions with a numeric suffix, and returns
+        both the symbol list and a mapping of sanitized-symbol -> original-value for
+        any symbol that was changed (issue #383).
+
+        Returns:
+            tuple: (avro_symbols, symbol_mapping)
+        """
+        avro_symbols: List[str] = []
+        symbol_mapping: Dict[str, Any] = {}
+        seen = set()
+        for value in symbols:
+            base = avro_name(str(value))
+            candidate = base
+            suffix = 1
+            while candidate in seen:
+                candidate = f"{base}_{suffix}"
+                suffix += 1
+            seen.add(candidate)
+            avro_symbols.append(candidate)
+            if candidate != str(value):
+                symbol_mapping[candidate] = value
+        return avro_symbols, symbol_mapping
+
     def _convert_enum(self, schema: Dict[str, Any], namespace: Optional[str], name: str) -> Dict[str, Any]:
         """Convert JSON Structure enum to Avro enum."""
+        avro_symbols, symbol_mapping = self._sanitize_enum_symbols(schema['enum'])
         avro_enum: Dict[str, Any] = {
             'type': 'enum',
             'name': name,
-            'symbols': schema['enum']
+            'symbols': avro_symbols
         }
         
         if namespace:
             avro_enum['namespace'] = namespace
         
+        doc_parts = []
         if 'description' in schema:
-            avro_enum['doc'] = schema['description']
+            doc_parts.append(schema['description'])
+        if symbol_mapping:
+            doc_parts.append(
+                'Original enum values (Avro symbol -> wire value): '
+                + json.dumps(symbol_mapping))
+        if doc_parts:
+            avro_enum['doc'] = ' '.join(doc_parts)
         
         if 'default' in schema:
-            avro_enum['default'] = schema['default']
+            original_values = list(schema['enum'])
+            if schema['default'] in original_values:
+                avro_enum['default'] = avro_symbols[original_values.index(schema['default'])]
         
         return avro_enum
     
@@ -531,11 +576,12 @@ class JsonStructureToAvro:
             # Build enum type for discriminator
             enum_name = f'{name}Type'
             choice_names = list(choices.keys())
+            discriminator_symbols, _ = self._sanitize_enum_symbols(choice_names)
             
             discriminator_enum: Dict[str, Any] = {
                 'type': 'enum',
                 'name': enum_name,
-                'symbols': choice_names
+                'symbols': discriminator_symbols
             }
             
             if namespace:

{structurize-3.6.0 → structurize-3.6.1}/avrotize/structuretokusto.py

@@ -1,12 +1,41 @@
 """Converts a JSON Structure schema to a Kusto table schema."""
 
 import json
+import re
 import sys
 from typing import Any, List, Optional, Dict, Union
 from avrotize.common import build_flat_type_dict, inline_avro_references, strip_first_doc
 from azure.kusto.data import KustoClient, KustoConnectionStringBuilder, ClientRequestProperties
 
 
+_KUSTO_BARE_IDENTIFIER = re.compile(r'^[A-Za-z_][A-Za-z0-9_]*$')
+
+
+def kusto_column_name(name: str) -> str:
+    """Return a Kusto-quoted column identifier.
+
+    Bare identifiers ([A-Za-z_][A-Za-z0-9_]*) are wrapped as [name]; any other
+    name (e.g. "dr-type") is quoted as ['name'] so hyphens and other characters
+    are valid (issue #382).
+    """
+    if _KUSTO_BARE_IDENTIFIER.match(name):
+        return f"[{name}]"
+    escaped = name.replace('\\', '\\\\').replace("'", "\\'")
+    return f"['{escaped}']"
+
+
+def kusto_json_path(name: str, prefix: str = "$") -> str:
+    """Return a JSONPath expression for an ingestion-mapping property name.
+
+    Bare identifiers use dotted notation (``$.name``); any other name uses
+    bracket notation (``$['dr-type']``) so non-identifier keys are valid.
+    """
+    if _KUSTO_BARE_IDENTIFIER.match(name):
+        return f"{prefix}.{name}"
+    escaped = name.replace('\\', '\\\\').replace("'", "\\'")
+    return f"{prefix}['{escaped}']"
+
+
 def kusto_string_literal(value: str) -> str:
     """Return a KQL verbatim string literal for a raw Python string."""
     return '@"' + value.replace('"', '""') + '"'
@@ -268,7 +297,7 @@ class StructureToKusto:
             if isinstance(prop_schema, dict) and 'const' in prop_schema:
                 continue
             column_type = self.convert_structure_type_to_kusto_type(prop_schema, schema_doc)
-            columns.append(f"   [{column_name}]: {column_type}")
+            columns.append(f"   {kusto_column_name(column_name)}: {column_type}")
         if emit_cloudevents_columns:
             columns.append("   [___type]: string")
             columns.append("   [___source]: string")
@@ -365,7 +394,7 @@ class StructureToKusto:
                 continue
             column_name = prop_name
             mapping_entries.append(
-                f"  {{\"column\": \"{column_name}\", \"path\": \"$.{prop_name}\"}}")
+                f"  {{\"column\": \"{column_name}\", \"path\": \"{kusto_json_path(prop_name)}\"}}")
         kusto.append("```\n[\n" + ",\n".join(mapping_entries) + "\n]\n```\n\n")
 
         if emit_cloudevents_columns:
@@ -383,7 +412,7 @@ class StructureToKusto:
                     continue
                 column_name = prop_name
                 ce_entries.append(
-                    f"  {{\"column\": \"{column_name}\", \"path\": \"$.data.{prop_name}\"}}")
+                    f"  {{\"column\": \"{column_name}\", \"path\": \"{kusto_json_path(prop_name, '$.data')}\"}}")
             kusto.append("```\n[\n" + ",\n".join(ce_entries) + "\n]\n```\n\n")
 
         if emit_cloudevents_columns:

{structurize-3.6.0 → structurize-3.6.1/structurize.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: structurize
-Version: 3.6.0
+Version: 3.6.1
 Summary: Tools to convert from and to JSON Structure from various other schema languages.
 Author-email: Clemens Vasters <clemensv@microsoft.com>
 Classifier: Programming Language :: Python :: 3