Merge pull request #281 from SylphAI-Inc/model_provider

add model providers display and code

README.md

@@ -142,6 +142,14 @@ You have full control over the prompt template, the model you use, and the outpu
   <img src="https://raw.githubusercontent.com/SylphAI-Inc/LightRAG/main/docs/source/_static/images/AdalFlow_task_pipeline.png" alt="AdalFlow Task Pipeline">
 </p>
 
+Many providers and models accessible via the same interface:
+
+<p align="center">
+  <img src="https://raw.githubusercontent.com/SylphAI-Inc/LightRAG/main/docs/source/_static/images/multi-providers.png" alt="AdalFlow Model Providers">
+</p>
+
+[All available model providers](https://adalflow.sylph.ai/apis/components/components.model_client.html)
+
 
 <!-- LLMs are like water; they can be shaped into anything, from GenAI applications such as chatbots, translation, summarization, code generation, and autonomous agents to classical NLP tasks like text classification and named entity recognition. They interact with the world beyond the model’s internal knowledge via retrievers, memory, and tools (function calls). Each use case is unique in its data, business logic, and user experience.
 

adalflow/CHANGELOG.md

@@ -1,3 +1,8 @@
+
+## [0.2.6] - 2024-11-25
+### Improved
+- Add default `max_tokens=512` to the `AnthropicAPIClient` to avoid the error when the user does not provide the `max_tokens` in the prompt.
+
 ## [0.2.5] - 2024-10-28
 
 ### Fixed

adalflow/adalflow/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.2.5"
+__version__ = "0.2.6"
 
 from adalflow.core.component import Component, fun_to_component
 from adalflow.core.container import Sequential

adalflow/adalflow/components/model_client/anthropic_client.py

@@ -44,7 +44,10 @@ class AnthropicAPIClient(ModelClient):
 
     Visit https://docs.anthropic.com/en/docs/intro-to-claude for more api details.
 
-    Ensure "max_tokens" are set.
+    Note:
+
+    As antropic API needs users to set max_tokens, we set up a default value of 512 for the max_tokens.
+    You can override this value by passing the max_tokens in the model_kwargs.
 
     Reference: 8/1/2024
     - https://docs.anthropic.com/en/docs/about-claude/models
@@ -64,6 +67,7 @@ def __init__(
         self.chat_completion_parser = (
             chat_completion_parser or get_first_message_content
         )
+        self.default_max_tokens = 512
 
     def init_sync_client(self):
         api_key = self._api_key or os.getenv("ANTHROPIC_API_KEY")
@@ -116,6 +120,8 @@ def convert_inputs_to_api_kwargs(
             api_kwargs["messages"] = [
                 {"role": "user", "content": input},
             ]
+            if "max_tokens" not in api_kwargs:
+                api_kwargs["max_tokens"] = self.default_max_tokens
             # if input and input != "":
             #     api_kwargs["system"] = input
         else:

adalflow/adalflow/components/model_client/bedrock_client.py

@@ -34,13 +34,27 @@ def get_first_message_content(completion: Dict) -> str:
 
 class BedrockAPIClient(ModelClient):
     __doc__ = r"""A component wrapper for the Bedrock API client.
+
+    Note:
+
+    This client needs a lot more work to be fully functional.
+    (1) Setup the AWS credentials.
+    (2) Access to the modelId.
+    (3) Convert the modelId to standard model.
+
+    To setup the AWS credentials, follow the instructions here:
+    https://docs.aws.amazon.com/bedrock/latest/userguide/getting-started.html
+
+    Additionally, this medium article is a good reference:
+    https://medium.com/@harangpeter/setting-up-aws-bedrock-for-api-based-text-inference-dc25ab2b216b
+
     Visit https://docs.aws.amazon.com/bedrock/latest/APIReference/welcome.html for more api details.
     """
 
     def __init__(
         self,
-        aws_profile_name=None,
-        aws_region_name=None,
+        aws_profile_name="default",
+        aws_region_name="us-west-2",  # Use a supported default region
         aws_access_key_id=None,
         aws_secret_access_key=None,
         aws_session_token=None,
@@ -91,6 +105,8 @@ def init_sync_client(self):
             aws_session_token=aws_session_token,
         )
         bedrock_runtime = session.client(service_name="bedrock-runtime", config=config)
+
+        self._client = session.client(service_name="bedrock")
         return bedrock_runtime
 
     def init_async_client(self):
@@ -117,6 +133,21 @@ def track_completion_usage(self, completion: Dict) -> CompletionUsage:
             total_tokens=usage["totalTokens"],
         )
 
+    def list_models(self):
+        # Initialize Bedrock client (not runtime)
+
+        try:
+            response = self._client.list_foundation_models()
+            models = response.get("models", [])
+            for model in models:
+                print(f"Model ID: {model['modelId']}")
+                print(f"  Name: {model['name']}")
+                print(f"  Description: {model['description']}")
+                print(f"  Provider: {model['provider']}")
+                print("")
+        except Exception as e:
+            print(f"Error listing models: {e}")
+
     def convert_inputs_to_api_kwargs(
         self,
         input: Optional[Any] = None,

adalflow/adalflow/components/model_client/google_client.py

@@ -34,24 +34,27 @@ class GoogleGenAIClient(ModelClient):
 
     Info: 8/1/2024
     Tested: gemini-1.0-pro, gemini-1.5-pro-latest
-    class UsageMetadata(proto.Message):
 
-        prompt_token_count: int = proto.Field(
-            proto.INT32,
-            number=1,
-        )
-        cached_content_token_count: int = proto.Field(
-            proto.INT32,
-            number=4,
-        )
-        candidates_token_count: int = proto.Field(
-            proto.INT32,
-            number=2,
-        )
-        total_token_count: int = proto.Field(
-            proto.INT32,
-            number=3,
-        )
+    .. code-block:: python
+
+        class UsageMetadata(proto.Message):
+
+            prompt_token_count: int = proto.Field(
+                proto.INT32,
+                number=1,
+            )
+            cached_content_token_count: int = proto.Field(
+                proto.INT32,
+                number=4,
+            )
+            candidates_token_count: int = proto.Field(
+                proto.INT32,
+                number=2,
+            )
+            total_token_count: int = proto.Field(
+                proto.INT32,
+                number=3,
+            )
     """
 
     def __init__(self, api_key: Optional[str] = None):

adalflow/adalflow/core/types.py

@@ -39,7 +39,7 @@
     GroqAPIClient,
     OpenAIClient,
     GoogleGenAIClient,
-    OllamaClient
+    OllamaClient,
 )
 
 

adalflow/pyproject.toml

@@ -1,7 +1,7 @@
 [tool.poetry]
 name = "adalflow"
 
-version = "0.2.5"
+version = "0.2.6"
 description = "The Library to Build and Auto-optimize LLM Applications"
 authors = ["Li Yin <[email protected]>"]
 readme = "README.md"
@@ -118,10 +118,20 @@ url = "https://pypi.nvidia.com"
 # priority = "supplemental"
 # url = "https://pypi.nvidia.com"
 
-[tool.ruff]
-exclude = ["images"]
-
 
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+
+# for formatting and linting
+[tool.black]
+line-length = 88
+target-version = ["py311"]
+
+[tool.ruff]
+exclude = ["images"]
+lint.extend-ignore = [
+    "E402",  # Ignore module-level import issues
+    "E731",
+    "UP007", # Wants | over Union, which breaks 3.8
+]