opentelemetry-instrumentation-botocore: fix handling of tool input chunked json in bedrock converse_stream (#3544)

* opentelemetry-instrumentation-botocore: fix handling of tool input chunked json in converse_stream

We need to accumulate all the tool input as string before decoding it
otherwise we may end up with invalid json.

* Add changelog

* Remove leftover print

* No need for too-many-statements disable

Author: xrmx

CHANGELOG.md

@@ -37,6 +37,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   ([#3520](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3520))
 - `opentelemetry-instrumentation-botocore` Ensure spans end on early stream closure for Bedrock Streaming APIs
   ([#3481](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3481))
+- `opentelemetry-instrumentation-botocore`: fix handling of tool input in Bedrock ConverseStream
+  ([#3544](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3544))
 - `opentelemetry-instrumentation-botocore` Add type check when extracting tool use from Bedrock request message content
   ([#3548](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3548))
 

instrumentation/opentelemetry-instrumentation-botocore/src/opentelemetry/instrumentation/botocore/extensions/bedrock_utils.py

@@ -64,6 +64,7 @@ def __init__(
         self._response = {}
         self._message = None
         self._content_block = {}
+        self._tool_json_input_buf = ""
         self._record_message = False
         self._ended = False
 
@@ -89,28 +90,40 @@ def _process_event(self, event):
             # {'contentBlockStart': {'start': {'toolUse': {'toolUseId': 'id', 'name': 'func_name'}}, 'contentBlockIndex': 1}}
             start = event["contentBlockStart"].get("start", {})
             if "toolUse" in start:
-                tool_use = _decode_tool_use(start["toolUse"])
-                self._content_block = {"toolUse": tool_use}
+                self._content_block = {"toolUse": start["toolUse"]}
             return
 
         if "contentBlockDelta" in event:
             # {'contentBlockDelta': {'delta': {'text': "Hello"}, 'contentBlockIndex': 0}}
             # {'contentBlockDelta': {'delta': {'toolUse': {'input': '{"location":"Seattle"}'}}, 'contentBlockIndex': 1}}
+            # {'contentBlockDelta': {'delta': {'toolUse': {'input': 'a", "Yok'}}, 'contentBlockIndex': 1}}
             if self._record_message:
                 delta = event["contentBlockDelta"].get("delta", {})
                 if "text" in delta:
                     self._content_block.setdefault("text", "")
                     self._content_block["text"] += delta["text"]
                 elif "toolUse" in delta:
-                    tool_use = _decode_tool_use(delta["toolUse"])
-                    self._content_block["toolUse"].update(tool_use)
+                    if (
+                        input_buf := delta["toolUse"].get("input")
+                    ) is not None:
+                        self._tool_json_input_buf += input_buf
             return
 
         if "contentBlockStop" in event:
             # {'contentBlockStop': {'contentBlockIndex': 0}}
             if self._record_message:
+                if self._tool_json_input_buf:
+                    try:
+                        self._content_block["toolUse"]["input"] = json.loads(
+                            self._tool_json_input_buf
+                        )
+                    except json.DecodeError:
+                        self._content_block["toolUse"]["input"] = (
+                            self._tool_json_input_buf
+                        )
                 self._message["content"].append(self._content_block)
                 self._content_block = {}
+                self._tool_json_input_buf = ""
             return
 
         if "messageStop" in event:

instrumentation/opentelemetry-instrumentation-botocore/tests/bedrock_utils.py

@@ -275,8 +275,6 @@ def remove_none_values(body):
             continue
         if isinstance(value, dict):
             result[key] = remove_none_values(value)
-        elif isinstance(value, list):
-            result[key] = [remove_none_values(i) for i in value]
         else:
             result[key] = value
     return result