Float aware mini

.gitignore

@@ -14,3 +14,6 @@ _build/
 
 # coverage.py default output file
 .coverage
+
+# IDE cruft
+.idea/

docs/performance.rst

@@ -36,7 +36,7 @@ Both of these approaches will be much faster than using `load/s()`, since
 they avoid loading the parts of the document we didn't care about.
 
 Both `Object` and `Array` have a `mini` property that returns their entire
-content as a minified Python `str`. A message router for example would only
+content as a minified Python `bytes`. A message router for example would only
 parse the document and retrieve a single property, the destination, and forward
 the payload without ever turning it into a Python object. Here's a (bad)
 example:

simdjson/__init__.pyi

@@ -52,7 +52,7 @@ class Object(Mapping[str, SimValue]):
         ...
 
     @property
-    def mini(self) -> str:
+    def mini(self) -> bytes:
         ...
 
 
@@ -73,7 +73,7 @@ class Array(Sequence[SimValue]):
         ...
 
     @property
-    def mini(self) -> str:
+    def mini(self) -> bytes:
         ...