Make it so single-member arrays become objects

Among other QoL improvements.

Closes better#25.

jsonschema2db.py

@@ -1,3 +1,4 @@
+from collections.abc import Iterable
 import change_case
 import csv
 import datetime
@@ -37,6 +38,7 @@ def __init__(self, schema, database_flavor, postgres_schema=None, debug=False,
         self._database_flavor = database_flavor
         self._debug = debug
         self._table_definitions = {}
+        self._table_required = {root_table: set()} # keyed by _table_definitions
         self._links = {}
         self._backlinks = {}
         self._postgres_schema = postgres_schema
@@ -86,12 +88,18 @@ def __init__(self, schema, database_flavor, postgres_schema=None, debug=False,
                     warnings.warn('Ignoring_column because it is too long: %s.%s' % (table, column))
             columns = sorted(col for col in column_types.keys() if 0 < len(col) <= max_column_length)
             self._table_columns[table] = columns
+            self._table_required[table] = set()
 
-    def _table_name(self, path):
-        return '__'.join(change_case.ChangeCase.camel_to_snake(self._abbreviations.get(p, p)) for p in path)
+    def _table_name(self, path, fallible=False):
+        try:
+            return '__'.join(change_case.ChangeCase.camel_to_snake(self._abbreviations.get(p, p)) for p in path)
+        except Exception as e:
+            warnings.warn('%s: %s getting table name!' % (path, e))
+            if not fallible:
+                raise e
 
-    def _column_name(self, path):
-        return self._table_name(path)  # same
+    def _column_name(self, path, **kwargs):
+        return self._table_name(path, **kwargs)  # same
 
     def _execute(self, cursor, query, args=None, query_ok_to_print=True):
         if self._debug and query_ok_to_print:
@@ -112,9 +120,14 @@ def _traverse(self, schema, tree, path=tuple(), table='root', parent=None, comme
 
         if table not in self._table_definitions:
             self._table_definitions[table] = {}
+            self._table_required[table] = set()
             if comment:
                 self._table_comments[table] = comment
 
+        if 'required' in tree:
+            for req in tree['required']:
+                self._table_required[table].add(self._column_name(path))
+
         definition = None
         new_json_path = json_path
         while '$ref' in tree:
@@ -144,7 +157,22 @@ def _traverse(self, schema, tree, path=tuple(), table='root', parent=None, comme
         elif 'type' not in tree:
             res = {}
             warnings.warn('%s.%s: Type info missing' % (table, self._column_name(path)))
-        elif tree['type'] == 'object':
+
+        if 'type' in tree:
+            if tree['type'] != 'null' and isinstance(tree['type'], Iterable) and 'null' in tree['type']: # nullable
+                tree['type'] = Nullable(next((t for t in tree['type'] if t != 'null')))
+            elif (tpath := self._column_name(path)) is not None and tpath in self._table_required[table]:
+                tree['type'] = NotNullable(tree['type'])
+
+        if tree['type'] == 'array' and 'items' in tree:
+            if len(tree['items']) == 1:
+                tree['type'] = Nullable('object')
+                tree['patternProperties'] = { '[0-9]+': tree['items'] }
+                del tree['items']
+            else:
+                warnings.warn('%s.%s: Arrays with mismatched item types are not yet gracefully handled.' % (table, self._column_name(path)))
+
+        if tree['type'] == 'object':
             print('object:', tree)
             res = {}
             if 'patternProperties' in tree:
@@ -242,7 +270,7 @@ def create_tables(self, con):
 
         :param con: psycopg2 connection object
         '''
-        postgres_types = {'boolean': 'bool', 'number': 'float', 'string': 'text', 'enum': 'text', 'integer': 'bigint', 'timestamp': 'timestamptz', 'date': 'date', 'link': 'integer'}
+        postgres_types = {'boolean': 'bool', 'number': 'float', 'string': 'text', 'enum': 'text', 'integer': 'bigint', 'timestamp': 'timestamptz', 'date': 'date', 'link': 'bigint'}
         with con.cursor() as cursor:
             if self._postgres_schema is not None:
                 self._execute(cursor, 'drop schema if exists %s cascade' % self._postgres_schema)
@@ -253,7 +281,7 @@ def create_tables(self, con):
 
                 create_q = 'create table %s (id %s, "%s" %s not null, "%s" text not null, %s unique ("%s", "%s"), unique (id))' % \
                            (self._postgres_table_name(table), id_data_type, self._item_col_name, postgres_types[self._item_col_type], self._prefix_col_name,
-                            ''.join('"%s" %s, ' % (c, postgres_types[t]) for c, t in zip(columns, types)),
+                            ''.join('"%s" %s %s, ' % (c, postgres_types[str(t)], 'not null' if isinstance(t, NotNullable) else 'null' if isinstance(t, Nullable) else '') for c, t in zip(columns, types)),
                             self._item_col_name, self._prefix_col_name)
                 self._execute(cursor, create_q)
                 if table in self._table_comments:
@@ -442,8 +470,17 @@ def __init__(self, *args, **kwargs):
         return super(JSONSchemaToPostgres, self).__init__(*args, **kwargs)
 
 
+
 class JSONSchemaToRedshift(JSONSchemaToDatabase):
     '''Shorthand for JSONSchemaToDatabase(..., database_flavor='redshift')'''
     def __init__(self, *args, **kwargs):
         kwargs['database_flavor'] = 'redshift'
         return super(JSONSchemaToRedshift, self).__init__(*args, **kwargs)
+
+
+class Nullable(str):
+    def __eq__(a, b):
+        return str(a).__eq__(str(b))
+
+class NotNullable(str):
+    __eq__ = Nullable.__eq__