Merge branch 'master' into feat/huggingface-inference

.github/workflows/test.yaml

@@ -235,6 +235,7 @@ jobs:
             "trace",
             "trace_server",
             "anthropic",
+            "bedrock",
             "cerebras",
             "cohere",
             "dspy",

docs/docs/guides/integrations/bedrock.md

@@ -0,0 +1,137 @@
+# Amazon Bedrock
+
+Weave automatically tracks and logs LLM calls made via Amazon Bedrock, AWS's fully managed service that offers foundation models from leading AI companies through a unified API.
+
+## Traces
+
+Weave will automatically capture traces for Bedrock API calls. You can use the Bedrock client as usual after initializing Weave and patching the client:
+
+```python
+import weave
+import boto3
+import json
+from weave.integrations.bedrock.bedrock_sdk import patch_client
+
+weave.init("my_bedrock_app")
+
+# Create and patch the Bedrock client
+client = boto3.client("bedrock-runtime")
+patch_client(client)
+
+# Use the client as usual
+response = client.invoke_model(
+    modelId="anthropic.claude-3-5-sonnet-20240620-v1:0",
+    body=json.dumps({
+        "anthropic_version": "bedrock-2023-05-31",
+        "max_tokens": 100,
+        "messages": [
+            {"role": "user", "content": "What is the capital of France?"}
+        ]
+    }),
+    contentType='application/json',
+    accept='application/json'
+)
+response_dict = json.loads(response.get('body').read())
+print(response_dict["content"][0]["text"])
+```
+
+of using the `converse` API:
+
+```python
+messages = [{"role": "user", "content": [{"text": "What is the capital of France?"}]}]
+
+response = client.converse(
+    modelId="anthropic.claude-3-5-sonnet-20240620-v1:0",
+    system=[{"text": "You are a helpful AI assistant."}],
+    messages=messages,
+    inferenceConfig={"maxTokens": 100},
+)
+print(response["output"]["message"]["content"][0]["text"])
+
+```
+
+## Wrapping with your own ops
+
+You can create reusable operations using the `@weave.op()` decorator. Here's an example showing both the `invoke_model` and `converse` APIs:
+
+```python
+@weave.op
+def call_model_invoke(
+    model_id: str,
+    prompt: str,
+    max_tokens: int = 100,
+    temperature: float = 0.7
+) -> dict:
+    body = json.dumps({
+        "anthropic_version": "bedrock-2023-05-31",
+        "max_tokens": max_tokens,
+        "temperature": temperature,
+        "messages": [
+            {"role": "user", "content": prompt}
+        ]
+    })
+
+    response = client.invoke_model(
+        modelId=model_id,
+        body=body,
+        contentType='application/json',
+        accept='application/json'
+    )
+    return json.loads(response.get('body').read())
+
+@weave.op
+def call_model_converse(
+    model_id: str,
+    messages: str,
+    system_message: str,
+    max_tokens: int = 100,
+) -> dict:
+    response = client.converse(
+        modelId=model_id,
+        system=[{"text": system_message}],
+        messages=messages,
+        inferenceConfig={"maxTokens": max_tokens},
+    )
+    return response
+```
+
+![](./imgs/bedrock_converse.png)
+
+## Create a `Model` for easier experimentation
+
+You can create a Weave Model to better organize your experiments and capture parameters. Here's an example using the `converse` API:
+
+```python
+class BedrockLLM(weave.Model):
+    model_id: str
+    max_tokens: int = 100
+    system_message: str = "You are a helpful AI assistant."
+
+    @weave.op
+    def predict(self, prompt: str) -> str:
+        "Generate a response using Bedrock's converse API"
+
+        messages = [{
+            "role": "user",
+            "content": [{"text": prompt}]
+        }]
+
+        response = client.converse(
+            modelId=self.model_id,
+            system=[{"text": self.system_message}],
+            messages=messages,
+            inferenceConfig={"maxTokens": self.max_tokens},
+        )
+        return response["output"]["message"]["content"][0]["text"]
+
+# Create and use the model
+model = BedrockLLM(
+    model_id="anthropic.claude-3-5-sonnet-20240620-v1:0",
+    max_tokens=100,
+    system_message="You are an expert software engineer that knows a lot of programming. You prefer short answers."
+)
+result = model.predict("What is the best way to handle errors in Python?")
+print(result)
+```
+
+This approach allows you to version your experiments and easily track different configurations of your Bedrock-based application.

docs/docs/guides/tracking/objects.md

@@ -58,6 +58,28 @@ Saving an object with a name will create the first version of that object if it
   </TabItem>
 </Tabs>
 
+## Deleting an object
+
+<Tabs groupId="programming-language" queryString>
+  <TabItem value="python" label="Python" default>
+    To delete a version of an object, call `.delete()` on the object ref.
+
+    ```python
+    weave.init('intro-example')
+    cat_names_ref = weave.ref('cat-names:v1')
+    cat_names_ref.delete()
+    ```
+
+    Trying to access a deleted object will result in an error. Resolving an object that has a reference to a deleted object will return a `DeletedRef` object in place of the deleted object.
+
+  </TabItem>
+  <TabItem value="typescript" label="TypeScript">
+    ```plaintext
+    This feature is not available in TypeScript yet.  Stay tuned!
+    ```
+  </TabItem>
+</Tabs>
+
 ## Ref styles
 
 A fully qualified weave object ref uri looks like this:

docs/docs/guides/tracking/ops.md

@@ -156,3 +156,25 @@ If you want to suppress the printing of call links during logging, you can set t
 ```bash
 export WEAVE_PRINT_CALL_LINK=false
 ```
+
+## Deleting an op
+
+<Tabs groupId="programming-language" queryString>
+  <TabItem value="python" label="Python" default>
+    To delete a version of an op, call `.delete()` on the op ref.
+
+    ```python
+    weave.init('intro-example')
+    my_op_ref = weave.ref('track_me:v1')
+    my_op_ref.delete()
+    ```
+
+    Trying to access a deleted op will result in an error.
+
+  </TabItem>
+  <TabItem value="typescript" label="TypeScript">
+    ```plaintext
+    This feature is not available in TypeScript yet.  Stay tuned!
+    ```
+  </TabItem>
+</Tabs>

docs/docs/guides/tracking/tracing.mdx

@@ -5,6 +5,7 @@ import TracingCallsMacroImage from '@site/static/img/screenshots/calls_macro.png
 import TracingCallsFilterImage from '@site/static/img/screenshots/calls_filter.png';
 import BasicCallImage from '@site/static/img/screenshots/basic_call.png';
 
+
 # Calls
 
 <DesktopWindow 
@@ -699,17 +700,7 @@ The easiest way to get started is to construct a view in the UI, then learn more
 :::info[Comming Soon]
 ::: */}
 
-## Call FAQ
-
-{/* TODO:
-Common Questions / Variations:
-* Images
-* Ops
-* Cost?
-* General data model
-*/}
-
-#### Call Schema
+### Call Schema
 
 Please see the [schema](../../reference/python-sdk/weave/trace_server/weave.trace_server.trace_server_interface#class-callschema) for a complete list of fields.
 
@@ -748,4 +739,64 @@ Calculated Fields:
     * Duration
     * Status
 
+## FAQs
+
+### How do I disable tracing?
+
+**Environment Variable:**
+
+In situations where you want to unconditionally disable tracing for the entire program, you can set the environment variable `WEAVE_DISABLED=true`.
+
+**Client Initialization:**
+
+Sometimes, you may want to conditionally enable tracing for a specific initialization based on some condition. In this case, you can initialize the client with the `disabled` flag in init settings.
+
+```python
+import weave
+
+# Initialize the client
+client = weave.init(..., settings={"disabled": True})
+```
+
+**Context Manager:**
+
+Finally, you may want to conditionally disable tracing for a single function based on some application logic. In this case, you can use the context manager `with set_tracing_enabled(False)` which can be imported from `weave.trace.context.call_context`.
+
+```python
+import weave
+from weave.trace.context.call_context import set_tracing_enabled
+
+client = weave.init(...)
+
+@weave.op
+def my_op():
+    ...
+
+with set_tracing_enabled(False):
+    my_op()
+```
+
+### How do I capture information about a call?
+
+Typically you would call an op directly:
+
+```python
+@weave.op
+def my_op():
+    ...
+
+my_op()
+```
+
+However, you can also get access to the call object directly by invoking the `call` method on the op:
+
+```python
+@weave.op
+def my_op():
+    ...
+
+output, call = my_op.call()
+```
+
+From here, the `call` object will have all the information about the call, including the inputs, outputs, and other metadata.
 

docs/sidebars.ts

@@ -98,6 +98,7 @@ const sidebars: SidebarsConfig = {
             "guides/integrations/cohere",
             "guides/integrations/mistral",
             "guides/integrations/azure",
+            "guides/integrations/bedrock",
             "guides/integrations/google-gemini",
             "guides/integrations/together_ai",
             "guides/integrations/groq",

noxfile.py

@@ -12,6 +12,7 @@
     "litellm",
     "notdiamond",
     "google_ai_studio",
+    "bedrock",
     "scorers_tests",
 ]
 
@@ -50,6 +51,7 @@ def lint(session):
         "notdiamond",
         "openai",
         "vertexai",
+        "bedrock",
         "scorers_tests",
         "pandas-test",
         "huggingface",

pyproject.toml

@@ -89,6 +89,7 @@ pandas-test = ["pandas>=2.2.3"]
 modal = ["modal", "python-dotenv"]
 huggingface = ["huggingface-hub>=0.26.2"]
 vertexai = ["vertexai>=1.70.0"]
+bedrock = ["boto3", "moto[bedrock]>=5.0.24"]
 test = [
   "nox",
   "pytest>=8.2.0",
@@ -230,7 +231,7 @@ module = "weave_query.*"
 ignore_errors = true
 
 [tool.bumpversion]
-current_version = "0.51.29-dev0"
+current_version = "0.51.30-dev0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.