prediction-market-agent-tooling 0.56.0.dev1858__py3-none-any.whl → 0.56.0.dev1860__py3-none-any.whl

prediction_market_agent_tooling/tools/caches/db_cache.py

@@ -114,6 +114,7 @@ def db_cache(
 
     @wraps(func)
     def wrapper(*args: Any, **kwargs: Any) -> Any:
+        # If caching is disabled, just call the function and return it
         if not api_keys.ENABLE_CACHE:
             return func(*args, **kwargs)
 
@@ -183,11 +184,15 @@ def db_cache(
 
         if cached_result:
             logger.info(
-                f"Cache hit for {full_function_name} with args {args_dict} and output {cached_result.result}"
+                # Keep the special [case-hit] identifier so we can easily track it in GCP.
+                f"[cache-hit] Cache hit for {full_function_name} with args {args_dict} and output {cached_result.result}"
             )
             if is_pydantic_model:
+                # If the output contains any Pydantic models, we need to initialise them.
                 try:
-                    return convert_to_pydantic(return_type, cached_result.result)
+                    return convert_return_type_to_pydantic(
+                        return_type, cached_result.result
+                    )
                 except ValueError as e:
                     # In case of backward-incompatible pydantic model, just treat it as cache miss, to not error out.
                     logger.warning(
@@ -199,13 +204,14 @@ def db_cache(
 
         # On cache miss, compute the result
         computed_result = func(*args, **kwargs)
+        # Keep the special [case-miss] identifier so we can easily track it in GCP.
         logger.info(
-            f"Cache miss for {full_function_name} with args {args_dict}, computed the output {computed_result}"
+            f"[cache-miss] Cache miss for {full_function_name} with args {args_dict}, computed the output {computed_result}"
         )
 
-        # If postgres access was specified, save it to dB.
+        # If postgres access was specified, save it.
         if engine is not None and (cache_none or computed_result is not None):
-            # Call the original function
+            # In case of Pydantic outputs, we need to dictionarize it for it to be serializable.
             result_data = (
                 convert_pydantic_to_dict(computed_result)
                 if is_pydantic_model
@@ -229,18 +235,25 @@ def db_cache(
     return cast(FunctionT, wrapper)
 
 
-def contains_pydantic_model(tp: Any) -> bool:
-    if tp is None:
+def contains_pydantic_model(return_type: Any) -> bool:
+    """
+    Check if the return type contains anything that's a Pydantic model (including nested structures, like `list[BaseModel]`, `dict[str, list[BaseModel]]`, etc.)
+    """
+    if return_type is None:
         return False
-    origin = get_origin(tp)
+    origin = get_origin(return_type)
     if origin is not None:
-        return any(contains_pydantic_model(arg) for arg in get_args(tp))
-    if inspect.isclass(tp):
-        return issubclass(tp, BaseModel)
+        return any(contains_pydantic_model(arg) for arg in get_args(return_type))
+    if inspect.isclass(return_type):
+        return issubclass(return_type, BaseModel)
     return False
 
 
-def json_serializer_default_fn(y: Any) -> Any:
+def json_serializer_default_fn(y: DatetimeUTC | timedelta | date) -> str:
+    """
+    Used to serialize objects that don't support it by default into a specific string that can be deserialized out later.
+    If you add something here, also add it to `replace_custom_stringified_objects` below.
+    """
     if isinstance(y, DatetimeUTC):
         return f"DatetimeUTC::{y.isoformat()}"
     elif isinstance(y, timedelta):
@@ -257,6 +270,9 @@ def json_serializer(x: Any) -> str:
 
 
 def replace_custom_stringified_objects(obj: Any) -> Any:
+    """
+    Used to deserialize objects from `json_serializer_default_fn` into their proper form.
+    """
     if isinstance(obj, str):
         if obj.startswith("DatetimeUTC::"):
             iso_str = obj[len("DatetimeUTC::") :]
@@ -283,6 +299,9 @@ def json_deserializer(s: str) -> Any:
 
 
 def convert_pydantic_to_dict(value: Any) -> Any:
+    """
+    Convert Pydantic models to dictionaries, including if they are in nested structures.
+    """
     if isinstance(value, BaseModel):
         return value.model_dump()
     elif isinstance(value, dict):
@@ -295,20 +314,25 @@ def convert_pydantic_to_dict(value: Any) -> Any:
         return value
 
 
-def convert_to_pydantic(model: Any, data: Any) -> Any:
+def convert_return_type_to_pydantic(return_type: Any, data: Any) -> Any:
+    """
+    Used to initialize Pydantic models from anything cached that was originally a Pydantic model in the output. Including models in nested structures.
+    """
     # Get the origin and arguments of the model type
-    origin = get_origin(model)
-    args = get_args(model)
+    origin = get_origin(return_type)
+    args = get_args(return_type)
 
     # Check if the data is a dictionary
     if isinstance(data, dict):
         # If the model has no origin, check if it is a subclass of BaseModel
         if origin is None:
-            if inspect.isclass(model) and issubclass(model, BaseModel):
+            if inspect.isclass(return_type) and issubclass(return_type, BaseModel):
                 # Convert the dictionary to a Pydantic model
-                return model(
+                return return_type(
                     **{
-                        k: convert_to_pydantic(getattr(model, k, None), v)
+                        k: convert_return_type_to_pydantic(
+                            getattr(return_type, k, None), v
+                        )
                         for k, v in data.items()
                     }
                 )
@@ -319,20 +343,25 @@ def convert_to_pydantic(model: Any, data: Any) -> Any:
         elif origin is dict:
             key_type, value_type = args
             return {
-                convert_to_pydantic(key_type, k): convert_to_pydantic(value_type, v)
+                convert_return_type_to_pydantic(
+                    key_type, k
+                ): convert_return_type_to_pydantic(value_type, v)
                 for k, v in data.items()
             }
         else:
             # If the origin is not a dictionary, return the data as is
             return data
     # Check if the data is a list
-    elif isinstance(data, list):
-        # If the origin is a list, convert each item
-        if origin is list:
+    elif isinstance(data, (list, tuple)):
+        # If the origin is a list or tuple, convert each item
+        if origin in {list, tuple}:
             item_type = args[0]
-            return [convert_to_pydantic(item_type, item) for item in data]
+            converted_items = [
+                convert_return_type_to_pydantic(item_type, item) for item in data
+            ]
+            return type(data)(converted_items)
         else:
-            # If the origin is not a list, return the data as is
+            # If the origin is not a list or tuple, return the data as is
             return data
     else:
         # If the data is neither a dictionary nor a list, return it as is

prediction_market_agent_tooling/tools/is_invalid.py

@@ -34,9 +34,10 @@ QUESTION_IS_INVALID_PROMPT = """Main signs about an invalid question (sometimes
   - Which could give an incentive only to specific participants to commit an immoral violent action, but are in practice unlikely.
   - Valid: Will the US be engaged in a military conflict with a UN member state in 2021? (It’s unlikely for the US to declare war in order to win a bet on this market).
   - Valid: Will Derek Chauvin go to jail for the murder of George Flyod? (It’s unlikely that the jurors would collude to make a wrong verdict in order to win this market).
-- Questions with relative dates will resolve as invalid. Dates must be stated in absolute terms, not relative depending on the current time.
+- Questions with relative dates will resolve as invalid. Dates must be stated in absolute terms, not relative depending on the current time. But they can be relative to the even specified in the question itself.
 - Invalid: Who will be the president of the United States in 6 months? ("in 6 months depends on the current time").
 - Invalid: In the next 14 days, will Gnosis Chain gain another 1M users? ("in the next 14 days depends on the current time").
+- Valid: Will GNO price go up 10 days after Gnosis Pay cashback program is annouced? ("10 days after" is relative to the event in the question, so we can determine absolute value).
 - Questions about moral values and not facts will be resolved as invalid.
 - Invalid: "Is it ethical to eat meat?".
 

{prediction_market_agent_tooling-0.56.0.dev1858.dist-info → prediction_market_agent_tooling-0.56.0.dev1860.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: prediction-market-agent-tooling
-Version: 0.56.0.dev1858
+Version: 0.56.0.dev1860
 Summary: Tools to benchmark, deploy and monitor prediction market agents.
 Author: Gnosis
 Requires-Python: >=3.10,<3.12

{prediction_market_agent_tooling-0.56.0.dev1858.dist-info → prediction_market_agent_tooling-0.56.0.dev1860.dist-info}/RECORD

@@ -69,7 +69,7 @@ prediction_market_agent_tooling/tools/betting_strategies/market_moving.py,sha256
 prediction_market_agent_tooling/tools/betting_strategies/minimum_bet_to_win.py,sha256=-FUSuQQgjcWSSnoFxnlAyTeilY6raJABJVM2QKkFqAY,438
 prediction_market_agent_tooling/tools/betting_strategies/stretch_bet_between.py,sha256=THMXwFlskvzbjnX_OiYtDSzI8XVFyULWfP2525_9UGc,429
 prediction_market_agent_tooling/tools/betting_strategies/utils.py,sha256=kpIb-ci67Vc1Yqqaa-_S4OUkbhWSIYog4_Iwp69HU_k,97
-prediction_market_agent_tooling/tools/caches/db_cache.py,sha256=2a1bIDQk1KmVcp6cR8nSDKyT7rWnBQ7oqmqLjw49sps,12196
+prediction_market_agent_tooling/tools/caches/db_cache.py,sha256=QkhtNTs7sjwBhBJpkF_dWO5UZBE8Z1FT03WXskpakQg,13847
 prediction_market_agent_tooling/tools/caches/inmemory_cache.py,sha256=tGHHd9HCiE_hCCtPtloHZQdDfBuiow9YsqJNYi2Tx_0,499
 prediction_market_agent_tooling/tools/contract.py,sha256=s3yo8IbXTcvAJcPfLM0_NbgaEsWwLsPmyVnOgyjq_xI,20919
 prediction_market_agent_tooling/tools/costs.py,sha256=EaAJ7v9laD4VEV3d8B44M4u3_oEO_H16jRVCdoZ93Uw,954
@@ -81,7 +81,7 @@ prediction_market_agent_tooling/tools/httpx_cached_client.py,sha256=0-N1r0zcGKlY
 prediction_market_agent_tooling/tools/image_gen/image_gen.py,sha256=HzRwBx62hOXBOmrtpkXaP9Qq1Ku03uUGdREocyjLQ_k,1266
 prediction_market_agent_tooling/tools/image_gen/market_thumbnail_gen.py,sha256=8A3U2uxsCsOfLjru-6R_PPIAuiKY4qFkWp_GSBPV6-s,1280
 prediction_market_agent_tooling/tools/ipfs/ipfs_handler.py,sha256=CTTMfTvs_8PH4kAtlQby2aeEKwgpmxtuGbd4oYIdJ2A,1201
-prediction_market_agent_tooling/tools/is_invalid.py,sha256=EPZWH642LZZqx40_8ejJgE7Qe4nXLTXQIh4lBf5PDQw,5070
+prediction_market_agent_tooling/tools/is_invalid.py,sha256=Gp8ASGmqXZGdKKQ4U2ZfQj6bx2Wngb4IExIfV54E4j4,5322
 prediction_market_agent_tooling/tools/is_predictable.py,sha256=VGkxSoJ8CSLknloOLzm5J4-us7XImYxVzvpsAzxbpCc,6730
 prediction_market_agent_tooling/tools/langfuse_.py,sha256=jI_4ROxqo41CCnWGS1vN_AeDVhRzLMaQLxH3kxDu3L8,1153
 prediction_market_agent_tooling/tools/langfuse_client_utils.py,sha256=B0PhAQyviFnVbtOCYMxYmcCn66cu9nbqAOIAZcdgiRI,5771
@@ -97,8 +97,8 @@ prediction_market_agent_tooling/tools/tavily/tavily_models.py,sha256=5ldQs1pZe6u
 prediction_market_agent_tooling/tools/tavily/tavily_search.py,sha256=Kw2mXNkMTYTEe1MBSTqhQmLoeXtgb6CkmHlcAJvhtqE,3809
 prediction_market_agent_tooling/tools/utils.py,sha256=W-9SqeCKd51BYMRhDjYPQ7lfNO_zE9EvYpmu2r5WXGA,7163
 prediction_market_agent_tooling/tools/web3_utils.py,sha256=44W8siSLNQxeib98bbwAe7V5C609NHNlUuxwuWIRDiY,11838
-prediction_market_agent_tooling-0.56.0.dev1858.dist-info/LICENSE,sha256=6or154nLLU6bELzjh0mCreFjt0m2v72zLi3yHE0QbeE,7650
-prediction_market_agent_tooling-0.56.0.dev1858.dist-info/METADATA,sha256=K1Igqgk7hNSEVoHZ40ZW3x7s9_rjtsgDIGJ7HrH7nVk,8114
-prediction_market_agent_tooling-0.56.0.dev1858.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-prediction_market_agent_tooling-0.56.0.dev1858.dist-info/entry_points.txt,sha256=m8PukHbeH5g0IAAmOf_1Ahm-sGAMdhSSRQmwtpmi2s8,81
-prediction_market_agent_tooling-0.56.0.dev1858.dist-info/RECORD,,
+prediction_market_agent_tooling-0.56.0.dev1860.dist-info/LICENSE,sha256=6or154nLLU6bELzjh0mCreFjt0m2v72zLi3yHE0QbeE,7650
+prediction_market_agent_tooling-0.56.0.dev1860.dist-info/METADATA,sha256=D5iUZsQ2RbKbhyVvKEtc9A40BfZTJaJWoEqjLreNF_Q,8114
+prediction_market_agent_tooling-0.56.0.dev1860.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+prediction_market_agent_tooling-0.56.0.dev1860.dist-info/entry_points.txt,sha256=m8PukHbeH5g0IAAmOf_1Ahm-sGAMdhSSRQmwtpmi2s8,81
+prediction_market_agent_tooling-0.56.0.dev1860.dist-info/RECORD,,