Merge pull request #68 from AnswerDotAI/fix/array-for-gemini

Add 'items' field to JSON schema for Gemini compatibility

Author: jph00

01_funccall.ipynb

@@ -193,7 +193,7 @@
     "    if t is empty: raise TypeError('Missing type')\n",
     "    tmap = {int:\"integer\", float:\"number\", str:\"string\", bool:\"boolean\", list:\"array\", dict:\"object\"}\n",
     "    tmap.update({k.__name__: v for k, v in tmap.items()})\n",
-    "    if getattr(t, '__origin__', None) in (list,tuple):\n",
+    "    if getattr(t, '__origin__', None) in (list,tuple,set):\n",
     "        args = getattr(t, '__args__', None)\n",
     "        item_type = \"object\" if not args else tmap.get(t.__args__[0].__name__, \"object\")\n",
     "        return \"array\", item_type\n",
@@ -222,7 +222,10 @@
     {
      "data": {
       "text/plain": [
-       "(('array', 'integer'), ('integer', None), ('integer', None))"
+       "(('array', 'integer'),\n",
+       " ('array', 'integer'),\n",
+       " ('integer', None),\n",
+       " ('integer', None))"
       ]
      },
      "execution_count": null,
@@ -231,7 +234,7 @@
     }
    ],
    "source": [
-    "_types(list[int]), _types(int), _types('int')"
+    "_types(list[int]), _types(set[int]), _types(int), _types('int')"
    ]
   },
   {
@@ -435,9 +438,11 @@
     "\n",
     "def _handle_type(t, defs):\n",
     "    \"Handle a single type, creating nested schemas if necessary\"\n",
+    "    ot = ifnone(get_origin(t), t)\n",
     "    if t is NoneType: return {'type': 'null'}\n",
     "    if t in custom_types: return {'type':'string', 'format':t.__name__}\n",
-    "    if t in (dict, list, tuple, set): return {'type': _types(t)[0]}\n",
+    "    if ot is dict: return {'type': _types(t)[0]} \n",
+    "    if ot in (list, tuple, set): return {'type': _types(t)[0], 'items':{}}\n",
     "    if isinstance(t, type) and not issubclass(t, (int, float, str, bool)) or inspect.isfunction(t):\n",
     "        defs[t.__name__] = _get_nested_schema(t)\n",
     "        return {'$ref': f'#/$defs/{t.__name__}'}\n",
@@ -465,6 +470,51 @@
     "_handle_type(int, None), _handle_type(Path, None)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a43e9134",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "({'type': 'array', 'items': {}},\n",
+       " {'type': 'array', 'items': {}},\n",
+       " {'type': 'array', 'items': {}})"
+      ]
+     },
+     "execution_count": null,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# gemini expect `items` to be defined for arrays\n",
+    "_handle_type(list, None), _handle_type(tuple[str], None), _handle_type(set[str], None)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cf6d417e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "({'type': 'object'}, {'type': 'object'})"
+      ]
+     },
+     "execution_count": null,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "_handle_type(dict, None), _handle_type(dict[str,str], None)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -1710,14 +1760,14 @@
      "output_type": "stream",
      "text": [
       "Traceback (most recent call last):\n",
-      "  File \"/var/folders/51/b2_szf2945n072c0vj2cyty40000gn/T/ipykernel_6265/2052945749.py\", line 14, in python\n",
+      "  File \"/tmp/ipykernel_3890/2052945749.py\", line 14, in python\n",
       "    try: return _run(code, glb, loc)\n",
       "                ^^^^^^^^^^^^^^^^^^^^\n",
-      "  File \"/var/folders/51/b2_szf2945n072c0vj2cyty40000gn/T/ipykernel_6265/1858893181.py\", line 18, in _run\n",
+      "  File \"/tmp/ipykernel_3890/1858893181.py\", line 18, in _run\n",
       "    try: exec(compiled_code, glb, loc)\n",
       "         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
       "  File \"<ast>\", line 1, in <module>\n",
-      "  File \"/var/folders/51/b2_szf2945n072c0vj2cyty40000gn/T/ipykernel_6265/2052945749.py\", line 9, in handler\n",
+      "  File \"/tmp/ipykernel_3890/2052945749.py\", line 9, in handler\n",
       "    def handler(*args): raise TimeoutError()\n",
       "                        ^^^^^^^^^^^^^^^^^^^^\n",
       "TimeoutError\n",
@@ -2051,7 +2101,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "7ac04e80-7bb9-4b52-8285-454684605d47",
+   "id": "73bca085",
    "metadata": {},
    "outputs": [],
    "source": [

toolslm/funccall.py

@@ -25,7 +25,7 @@ def _types(t:type)->tuple[str,Optional[str]]:
     if t is empty: raise TypeError('Missing type')
     tmap = {int:"integer", float:"number", str:"string", bool:"boolean", list:"array", dict:"object"}
     tmap.update({k.__name__: v for k, v in tmap.items()})
-    if getattr(t, '__origin__', None) in (list,tuple):
+    if getattr(t, '__origin__', None) in (list,tuple,set):
         args = getattr(t, '__args__', None)
         item_type = "object" if not args else tmap.get(t.__args__[0].__name__, "object")
         return "array", item_type
@@ -58,9 +58,11 @@ def _param(
 
 def _handle_type(t, defs):
     "Handle a single type, creating nested schemas if necessary"
+    ot = ifnone(get_origin(t), t)
     if t is NoneType: return {'type': 'null'}
     if t in custom_types: return {'type':'string', 'format':t.__name__}
-    if t in (dict, list, tuple, set): return {'type': _types(t)[0]}
+    if ot is dict: return {'type': _types(t)[0]} 
+    if ot in (list, tuple, set): return {'type': _types(t)[0], 'items':{}}
     if isinstance(t, type) and not issubclass(t, (int, float, str, bool)) or inspect.isfunction(t):
         defs[t.__name__] = _get_nested_schema(t)
         return {'$ref': f'#/$defs/{t.__name__}'}
@@ -235,7 +237,7 @@ def call_func(fc_name, fc_inputs, ns, raise_on_err=True):
         if raise_on_err: raise e from None
         else: return traceback.format_exc()
 
-# %% ../01_funccall.ipynb #7ac04e80-7bb9-4b52-8285-454684605d47
+# %% ../01_funccall.ipynb #73bca085
 async def call_func_async(fc_name, fc_inputs, ns, raise_on_err=True):
     "Awaits the function `fc_name` with the given `fc_inputs` using namespace `ns`."
     res = call_func(fc_name, fc_inputs, ns, raise_on_err=raise_on_err)