Parse backtick-enclosed json (#6607)

## Why are these changes needed?

Some models enclose json in markdown code blocks

## Related issue number

resolves #6599. , #6547 

## Checks

- [ ] I've included any doc changes needed for
<https://microsoft.github.io/autogen/>. See
<https://github.com/microsoft/autogen/blob/main/CONTRIBUTING.md> to
build and test documentation locally.
- [ ] I've added tests (if relevant) corresponding to the changes
introduced in this PR.
- [ ] I've made sure all auto checks have passed.

---------

Co-authored-by: Victor Dibia <victordibia@microsoft.com>

Author: peterychang

python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_magentic_one/_magentic_one_orchestrator.py

@@ -11,6 +11,7 @@
     LLMMessage,
     UserMessage,
 )
+from autogen_core.utils import extract_json_from_str
 
 from .... import TRACE_LOGGER_NAME
 from ....base import Response, TerminationCondition
@@ -325,7 +326,12 @@ async def _orchestrate_step(self, cancellation_token: CancellationToken) -> None
             ledger_str = response.content
             try:
                 assert isinstance(ledger_str, str)
-                progress_ledger = json.loads(ledger_str)
+                output_json = extract_json_from_str(ledger_str)
+                if len(output_json) != 1:
+                    raise ValueError(
+                        f"Progress ledger should contain a single JSON object, but found: {len(progress_ledger)}"
+                    )
+                progress_ledger = output_json[0]
 
                 # If the team consists of a single agent, deterministically set the next speaker
                 if len(self._participant_names) == 1:

python/packages/autogen-core/src/autogen_core/utils/__init__.py

@@ -1,3 +1,4 @@
 from ._json_to_pydantic import schema_to_pydantic_model
+from ._load_json import extract_json_from_str
 
-__all__ = ["schema_to_pydantic_model"]
+__all__ = ["schema_to_pydantic_model", "extract_json_from_str"]

python/packages/autogen-core/src/autogen_core/utils/_load_json.py

@@ -0,0 +1,20 @@
+import json
+import re
+from typing import Any, Dict, List
+
+
+def extract_json_from_str(content: str) -> List[Dict[str, Any]]:
+    """Extract JSON objects from a string. Supports backtick enclosed JSON objects"""
+    pattern = re.compile(r"```(?:\s*([\w\+\-]+))?\n([\s\S]*?)```")
+    matches = pattern.findall(content)
+    ret: List[Dict[str, Any]] = []
+    # If no matches found, assume the entire content is a JSON object
+    if not matches:
+        ret.append(json.loads(content))
+    for match in matches:
+        language = match[0].strip() if match[0] else None
+        if language and language.lower() != "json":
+            raise ValueError(f"Expected JSON object, but found language: {language}")
+        content = match[1]
+        ret.append(json.loads(content))
+    return ret

python/packages/autogen-core/tests/test_json_extraction.py

@@ -0,0 +1,85 @@
+import pytest
+from autogen_core.utils import extract_json_from_str
+
+
+def test_extract_json_from_str() -> None:
+    json_str = """
+  {
+    "name": "John",
+    "age": 30,
+    "city": "New York"
+  }
+  """
+    json_resp = [{"name": "John", "age": 30, "city": "New York"}]
+    resp = extract_json_from_str(json_str)
+    assert resp == json_resp
+
+    invalid_json_str = """
+  {
+    "name": "John",
+    "age": 30,
+    "city": "New York"
+  """
+    with pytest.raises(ValueError):
+        extract_json_from_str(invalid_json_str)
+
+
+def test_extract_json_from_str_codeblock() -> None:
+    code_block_lang_str = """
+  ```json
+  {
+    "name": "Alice",
+    "age": 28,
+    "city": "Seattle"
+  }
+  ```
+  """
+    code_block_no_lang_str = """
+  ```
+  {
+    "name": "Alice",
+    "age": 28,
+    "city": "Seattle"
+  }
+  ```
+  """
+    code_block_resp = [{"name": "Alice", "age": 28, "city": "Seattle"}]
+    multi_json_str = """
+  ```json
+  {
+    "name": "John",
+    "age": 30,
+    "city": "New York"
+  }
+  ```
+  ```json
+  {
+    "name": "Jane",
+    "age": 25,
+    "city": "Los Angeles"
+  }
+  ```
+  """
+    multi_json_resp = [
+        {"name": "John", "age": 30, "city": "New York"},
+        {"name": "Jane", "age": 25, "city": "Los Angeles"},
+    ]
+
+    lang_resp = extract_json_from_str(code_block_lang_str)
+    assert lang_resp == code_block_resp
+    no_lang_resp = extract_json_from_str(code_block_no_lang_str)
+    assert no_lang_resp == code_block_resp
+    multi_resp = extract_json_from_str(multi_json_str)
+    assert multi_resp == multi_json_resp
+
+    invalid_lang_code_block_str = """
+  ```notjson
+  {
+    "name": "Jane",
+    "age": 25,
+    "city": "Los Angeles"
+  }
+  ```
+  """
+    with pytest.raises(ValueError):
+        extract_json_from_str(invalid_lang_code_block_str)

python/packages/autogen-ext/src/autogen_ext/models/anthropic/_anthropic_client.py

@@ -61,6 +61,7 @@
     validate_model_info,
 )
 from autogen_core.tools import Tool, ToolSchema
+from autogen_core.utils import extract_json_from_str
 from pydantic import BaseModel, SecretStr
 from typing_extensions import Self, Unpack
 
@@ -220,7 +221,10 @@ def assistant_message_to_anthropic(message: AssistantMessage) -> MessageParam:
             args = __empty_content_to_whitespace(args)
             if isinstance(args, str):
                 try:
-                    args_dict = json.loads(args)
+                    json_objs = extract_json_from_str(args)
+                    if len(json_objs) != 1:
+                        raise ValueError(f"Expected a single JSON object, but found {len(json_objs)}")
+                    args_dict = json_objs[0]
                 except json.JSONDecodeError:
                     args_dict = {"text": args}
             else: