flock-core 0.4.0b46__tar.gz → 0.4.0b49__tar.gz

flock_core-0.4.0b49/.github/ISSUE_TEMPLATE/bug_report.yaml

@@ -0,0 +1,27 @@
+name: 🪲 Bug Report
+description: File a Bug Report to help us improve
+title: "🪲 [BUG] <title>"
+labels: ["bug"]
+body:
+- type: textarea
+  attributes:
+    label: Describe the bug
+    description: A clear and concise description of what the bug is.
+  validations:
+    required: true
+- type: textarea
+  attributes:
+    label: To Reproduce
+    description: Describe how to reproduce the bug.
+  validations:
+    required: true
+- type: textarea
+  attributes:
+    label: Expected behavior 
+    description: A clear and concise description of what you expected to happen.
+  validations:
+    required: true
+- type: textarea
+  attributes:
+    label: Screenshots/Logs
+    description: If applicable, add screenshots and console output to help explain your problem. (You can drag and drop files here).

flock_core-0.4.0b49/.github/ISSUE_TEMPLATE/feature_request.yaml

@@ -0,0 +1,29 @@
+name: 🆕 Feature Request
+description: Suggest an idea for this project
+title: "🆕 [FEATURE] <title>"
+labels: ["enhancement"]
+body:
+- tpye: textarea
+  attributes:
+    label: Is your feature request related to a problem?
+    description: Please describe. A clear and concise description of what the problem is.
+  validations:
+    required: false
+- type: textarea
+  attributes:
+    label: Describe the solution you want to see
+    description: A clear and concise description of what you want to happen.
+  validations:
+    required: false
+- type: textarea
+  attributes:
+    label: Describe alternatives you have considered
+    description: A clear and concise description of any alternative solutions or features you have considered.
+  validations:
+    required: false
+- type: textarea
+  attributes:
+    label: Additional context
+    description: Add any other context or code snipptes about the feature request here.
+  validations:
+    required: false

flock_core-0.4.0b49/.github/PULL_REQUEST_TEMPLATE.yaml

@@ -0,0 +1,45 @@
+name: 🔄 Pull Request
+description: Create a Pull Request
+title: "🔄 [PR] <title>"
+labels: [""]
+body:
+- type: textarea
+  attributes:
+    label: Describe your Pull Request
+    description: A clear and concise description of what the Pull Request is about.
+  validations:
+    required: true
+- type: checkboxes
+  attributes:
+    label: is there an existing issue for this?
+    description: Check to see if there is an issue for this PR (optional)
+    options:
+    - label: Existing Issue
+      required: false
+    validations:
+      required: false
+- type: textarea
+  attributes:
+    label: Label of the exisiting Issue (if applicable)
+    description: Label/Name of the existing issue, if applicable 
+  validations:
+    required: false
+- type: checkboxes
+  attributes:
+    label: Please select relevant options:
+    description: Please select the options, which best describe this pull request.
+    options:
+    - label: Bug fix (non-breaking change which fixes an issue)
+    - label: New feature (non-breaking change which adds functionality)
+    - label: Breaking change (fix or feature that would cause existing functionality to not work as expected)
+  validations:
+    required: true
+- type: checkboxes
+  attributes:
+    label: Checklist
+    description: Please make sure you have followed these steps
+    options:
+    - label: My code follows the style guidelines of this project
+    - label: I have made corresponding changes to the documentation
+    - label: I have added tests that prove my fix is effective or that my feature works
+    - label: I have checked my code and corrected any misspellings

{flock_core-0.4.0b46 → flock_core-0.4.0b49}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: flock-core
-Version: 0.4.0b46
+Version: 0.4.0b49
 Summary: Declarative LLM Orchestration at Scale
 Author-email: Andre Ratzenberger <andre.ratzenberger@whiteduck.de>
 License-File: LICENSE
@@ -8,22 +8,23 @@ Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3
 Requires-Python: >=3.10
-Requires-Dist: azure-search-documents>=11.5.2
+Requires-Dist: aiosqlite>=0.21.0
 Requires-Dist: chromadb>=0.6.3
 Requires-Dist: cloudpickle>=3.1.1
 Requires-Dist: datasets>=3.2.0
 Requires-Dist: devtools>=0.12.2
-Requires-Dist: dspy==2.6.16
-Requires-Dist: duckduckgo-search>=7.3.2
+Requires-Dist: dspy==2.6.23
 Requires-Dist: fastapi>=0.115.8
 Requires-Dist: httpx>=0.28.1
 Requires-Dist: inspect-ai>=0.3.88
-Requires-Dist: litellm==1.63.7
+Requires-Dist: litellm==1.69.3
 Requires-Dist: loguru>=0.7.3
+Requires-Dist: markdown2>=2.5.3
 Requires-Dist: matplotlib>=3.10.0
+Requires-Dist: mem0ai[graph]>=0.1.100
 Requires-Dist: msgpack>=1.1.0
-Requires-Dist: nltk>=3.9.1
 Requires-Dist: notion-client>=2.3.0
+Requires-Dist: openai==1.75.0
 Requires-Dist: opentelemetry-api>=1.30.0
 Requires-Dist: opentelemetry-exporter-jaeger-proto-grpc>=1.21.0
 Requires-Dist: opentelemetry-exporter-jaeger>=1.21.0
@@ -49,14 +50,26 @@ Requires-Dist: tiktoken>=0.8.0
 Requires-Dist: toml>=0.10.2
 Requires-Dist: tqdm>=4.67.1
 Requires-Dist: uvicorn>=0.34.0
-Requires-Dist: zep-python>=2.0.2
-Provides-Extra: all
-Requires-Dist: docling>=2.18.0; extra == 'all'
-Requires-Dist: markdownify>=0.14.1; extra == 'all'
-Requires-Dist: tavily-python>=0.5.0; extra == 'all'
-Provides-Extra: tools
-Requires-Dist: markdownify>=0.14.1; extra == 'tools'
-Requires-Dist: tavily-python>=0.5.0; extra == 'tools'
+Provides-Extra: all-tools
+Requires-Dist: azure-identity>=1.23.0; extra == 'all-tools'
+Requires-Dist: azure-search-documents>=11.5.2; extra == 'all-tools'
+Requires-Dist: azure-storage-blob>=12.25.1; extra == 'all-tools'
+Requires-Dist: docling>=2.18.0; extra == 'all-tools'
+Requires-Dist: duckduckgo-search>=7.3.2; extra == 'all-tools'
+Requires-Dist: markdownify>=0.14.1; extra == 'all-tools'
+Requires-Dist: nltk>=3.9.1; extra == 'all-tools'
+Requires-Dist: tavily-python>=0.5.0; extra == 'all-tools'
+Provides-Extra: azure-tools
+Requires-Dist: azure-identity>=1.23.0; extra == 'azure-tools'
+Requires-Dist: azure-search-documents>=11.5.2; extra == 'azure-tools'
+Requires-Dist: azure-storage-blob>=12.25.1; extra == 'azure-tools'
+Provides-Extra: basic-tools
+Requires-Dist: docling>=2.18.0; extra == 'basic-tools'
+Requires-Dist: duckduckgo-search>=7.3.2; extra == 'basic-tools'
+Requires-Dist: markdownify>=0.14.1; extra == 'basic-tools'
+Requires-Dist: tavily-python>=0.5.0; extra == 'basic-tools'
+Provides-Extra: llm-tools
+Requires-Dist: nltk>=3.9.1; extra == 'llm-tools'
 Description-Content-Type: text/markdown
 
 <p align="center">

flock_core-0.4.0b49/SECURITY.md

@@ -0,0 +1,42 @@
+# Security Policy
+
+## Supported Versions
+
+The following versions of flock are currently actively receiving security updates:
+
+| Version | Supported          |
+| ------- | ------------------ |
+| 0.4.x  | :white_check_mark:  |
+
+## Reporting a Vulnerability
+
+We take every report of a vulnerability very seriously, 
+and actively encourage our community-members to report 
+any issues with the Flock framework as soon as possible
+in order to mitigate any potential risks.
+
+To report a vulnerability, you have serveral options available:
+
+### 1. Creating a Vulnerability Report (preferred):
+
+We prefer to keep our community informed about potential risks,
+and therefore would prefer you open a new [Vulnerability Report](https://github.com/whiteducksoftware/flock/issues/new/choose)
+on our Project Board.
+
+This helps keep our community informed about existing risks and allows us to react quickly to a potential
+vulnerability in the Flock framework.
+
+### 2. Contacting the Maintainers:
+
+If you prefer to report a vulnerability in private, feel free to contact
+any one of the core maintainers of Flock:
+
+- @tilman-sattler
+- @AndreRatzenberger
+- @swolg
+
+
+### 3. Official email: (slow)
+
+Alternatively, you can report any vulnerability you encounter to info@whiteduck.de.
+Keep in mind however, that this might be slower than Options 1 and 2.

flock_core-0.4.0b49/main.md

@@ -0,0 +1,54 @@
+\
+```python
+# Add these imports to the appropriate sections of src/flock/webapp/app/main.py
+
+# Standard library imports (if not already present)
+import uuid
+
+# Pydantic and FastAPI imports (ensure these are grouped with similar imports)
+from pydantic import BaseModel
+from fastapi import Depends, HTTPException 
+
+# Imports from your project for the sharing feature
+from flock.webapp.app.services.sharing_models import SharedLinkConfig
+from flock.webapp.app.services.sharing_store import SharedLinkStoreInterface
+from flock.webapp.app.dependencies import get_shared_link_store
+```
+
+```python
+# Add this code block in src/flock/webapp/app/main.py
+# A good place is after the app.include_router(...) lines 
+# and before other utility function definitions or page route definitions.
+
+# --- Share Link API Models and Endpoint ---
+class CreateShareLinkRequest(BaseModel):
+    agent_name: str
+
+class CreateShareLinkResponse(BaseModel):
+    share_url: str
+
+@app.post("/api/v1/share/link", response_model=CreateShareLinkResponse, tags=["UI Sharing"])
+async def create_share_link(
+    request_data: CreateShareLinkRequest,
+    store: SharedLinkStoreInterface = Depends(get_shared_link_store)
+):
+    \"\"\"Creates a new shareable link for an agent.\"\"\"
+    share_id = uuid.uuid4().hex
+    agent_name = request_data.agent_name
+
+    if not agent_name: # Basic validation
+        raise HTTPException(status_code=400, detail="Agent name cannot be empty.")
+
+    config = SharedLinkConfig(share_id=share_id, agent_name=agent_name)
+    try:
+        await store.save_config(config)
+        share_url = f"/ui/shared-run/{share_id}\" # Relative URL
+        # Ensure 'logger' is available in this scope, e.g., defined globally in main.py
+        logger.info(f"Created share link for agent \'{agent_name}\' with ID \'{share_id}\'. URL: {share_url}")
+        return CreateShareLinkResponse(share_url=share_url)
+    except Exception as e:
+        logger.error(f"Failed to create share link for agent \'{agent_name}\': {e}", exc_info=True)
+        # The f-string formatting for the exception detail was corrected here.
+        raise HTTPException(status_code=500, detail=f"Failed to create share link: {str(e)}")
+
+# --- End Share Link API --- 

{flock_core-0.4.0b46 → flock_core-0.4.0b49}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "flock-core"
-version = "0.4.0b-46"
+version = "0.4.0b-49"
 description = "Declarative LLM Orchestration at Scale"
 readme = "README.md"
 authors = [
@@ -13,21 +13,18 @@ classifiers = [
 ]
 requires-python = ">=3.10"
 dependencies = [
-    "azure-search-documents>=11.5.2",
     "chromadb>=0.6.3",
     "cloudpickle>=3.1.1",
     "datasets>=3.2.0",
     "devtools>=0.12.2",
-    "dspy==2.6.16",
-    "duckduckgo-search>=7.3.2",
+    "dspy==2.6.23",
     "fastapi>=0.115.8",
     "httpx>=0.28.1",
     "inspect-ai>=0.3.88",
-    "litellm==1.63.7",
+    "litellm==1.69.3",
     "loguru>=0.7.3",
     "matplotlib>=3.10.0",
     "msgpack>=1.1.0",
-    "nltk>=3.9.1",
     "notion-client>=2.3.0",
     "opentelemetry-api>=1.30.0",
     "opentelemetry-exporter-jaeger>=1.21.0",
@@ -35,6 +32,7 @@ dependencies = [
     "opentelemetry-exporter-otlp>=1.30.0",
     "opentelemetry-instrumentation-logging>=0.51b0",
     "opentelemetry-sdk>=1.30.0",
+    "openai==1.75.0",
     "pandas>=2.2.3",
     "pillow>=10.4.0",
     "prometheus-client>=0.21.1",
@@ -54,13 +52,36 @@ dependencies = [
     "toml>=0.10.2",
     "tqdm>=4.67.1",
     "uvicorn>=0.34.0",
-    "zep-python>=2.0.2",
+    "mem0ai[graph]>=0.1.100",
+    "aiosqlite>=0.21.0",
+    "markdown2>=2.5.3",
 ]
 
 [project.optional-dependencies]
-all = ["docling>=2.18.0", "tavily-python>=0.5.0", "markdownify>=0.14.1"]
-tools = ["tavily-python>=0.5.0", "markdownify>=0.14.1"]
-
+basic-tools = [
+    "docling>=2.18.0",
+    "tavily-python>=0.5.0",
+    "markdownify>=0.14.1",
+    "duckduckgo-search>=7.3.2",
+]
+azure-tools = [
+    "azure-identity>=1.23.0",
+    "azure-storage-blob>=12.25.1",
+    "azure-search-documents>=11.5.2",
+]
+llm-tools = [
+    "nltk>=3.9.1",
+]
+all-tools = [
+    "docling>=2.18.0",
+    "tavily-python>=0.5.0",
+    "markdownify>=0.14.1",
+    "duckduckgo-search>=7.3.2",
+    "azure-identity>=1.23.0",
+    "azure-storage-blob>=12.25.1",
+    "azure-search-documents>=11.5.2",
+    "nltk>=3.9.1",
+]
 
 [dependency-groups]
 dev = [

{flock_core-0.4.0b46 → flock_core-0.4.0b49}/src/flock/__init__.py

@@ -16,6 +16,11 @@ def main():
         action="store_true",
         help="Start the web interface instead of the CLI",
     )
+    parser.add_argument(
+        "--chat",
+        action="store_true",
+        help="Start a chat interface. If --web is also used, enables chat within the web app; otherwise, starts standalone chat.",
+    )
     parser.add_argument(
         "--theme",
         type=str,
@@ -30,7 +35,7 @@ def main():
             # Set environment variable for theme if provided
             if args.theme:
                 print(
-                    f"Setting FLOCK_WEB_THEME environment variable to: {args.theme}"
+                    f"INFO: Setting FLOCK_WEB_THEME environment variable to: {args.theme}"
                 )
                 os.environ["FLOCK_WEB_THEME"] = args.theme
             else:
@@ -38,10 +43,20 @@ def main():
                 if "FLOCK_WEB_THEME" in os.environ:
                     del os.environ["FLOCK_WEB_THEME"]
 
+            if args.chat: # --web --chat
+                print("INFO: Starting web application with chat feature enabled.")
+                os.environ["FLOCK_CHAT_ENABLED"] = "true"
+            else: # Just --web
+                print("INFO: Starting web application.")
+                if "FLOCK_CHAT_ENABLED" in os.environ:
+                    del os.environ["FLOCK_CHAT_ENABLED"]
+
+            # Ensure standalone chat mode is not active
+            if "FLOCK_START_MODE" in os.environ:
+                 del os.environ["FLOCK_START_MODE"]
+
             # Import and run the standalone webapp main function
-            # It will now read the theme from the environment variable
             from flock.webapp.run import main as run_webapp_main
-
             run_webapp_main()
 
         except ImportError:
@@ -55,6 +70,33 @@ def main():
             sys.exit(1)
         return
 
+    elif args.chat: # Standalone chat mode (args.web is false)
+        try:
+            print("INFO: Starting standalone chat application.")
+            os.environ["FLOCK_START_MODE"] = "chat"
+
+            # Clear web-specific env vars that might conflict or not apply
+            if "FLOCK_WEB_THEME" in os.environ:
+                del os.environ["FLOCK_WEB_THEME"]
+            if "FLOCK_CHAT_ENABLED" in os.environ:
+                del os.environ["FLOCK_CHAT_ENABLED"]
+
+            # Handle --theme if passed with --chat only
+            if args.theme:
+                 print(f"INFO: Standalone chat mode started with --theme '{args.theme}'. FLOCK_WEB_THEME will be set.")
+                 os.environ["FLOCK_WEB_THEME"] = args.theme
+
+            from flock.webapp.run import main as run_webapp_main
+            run_webapp_main() # The webapp main needs to interpret FLOCK_START_MODE="chat"
+
+        except ImportError:
+            print("Error: Could not import webapp components for chat. Ensure web dependencies are installed.", file=sys.stderr)
+            sys.exit(1)
+        except Exception as e:
+            print(f"Error starting standalone chat application: {e}", file=sys.stderr)
+            sys.exit(1)
+        return
+
     # Otherwise, run the CLI interface
     import questionary
     from rich.console import Console

{flock_core-0.4.0b46 → flock_core-0.4.0b49}/src/flock/core/flock.py

@@ -4,6 +4,7 @@
 from __future__ import annotations  # Ensure forward references work
 
 import asyncio
+import concurrent.futures  # Added import
 import os
 import uuid
 from collections.abc import Callable, Sequence
@@ -115,6 +116,19 @@ class Flock(BaseModel, Serializable):
         default=True,
         description="If True (default) and enable_temporal=True, start a temporary in-process worker for development/testing convenience. Set to False when using dedicated workers.",
     )
+
+    benchmark_agent_name: str | None = Field(
+        default=None,
+        description="The name of the agent to use for the benchmark.",
+    )
+    benchmark_eval_field: str | None = Field(
+        default=None,
+        description="The output field to use for the benchmark.",
+    )
+    benchmark_input_field: str | None = Field(
+        default=None,
+        description="The input field to use for the benchmark.",
+    )
     # Internal agent storage - not part of the Pydantic model for direct serialization
     # Marked with underscore to indicate it's managed internally and accessed via property
     _agents: dict[str, FlockAgent]
@@ -190,6 +204,8 @@ class Flock(BaseModel, Serializable):
         # Ensure session ID exists in baggage
         self._ensure_session_id()
 
+        FlockRegistry.discover_and_register_components()
+
         logger.info(
             "Flock instance initialized",
             name=self.name,
@@ -197,6 +213,46 @@ class Flock(BaseModel, Serializable):
             enable_temporal=self.enable_temporal,
         )
 
+    def prepare_benchmark(self, agent: FlockAgent | str | None = None, input_field: str | None = None, eval_field: str | None = None):
+        """Prepare a benchmark for the Flock instance."""
+        from flock.core.flock_agent import FlockAgent as ConcreteFlockAgent
+
+        logger.info(f"Preparing benchmark for Flock instance '{self.name}' with agent '{agent}'.")
+
+        name = agent.name if isinstance(agent, ConcreteFlockAgent) else agent
+
+        if self._agents.get(name) is None:
+            raise ValueError(f"Agent '{name}' not found in Flock instance '{self.name}'.")
+
+        self.benchmark_agent_name = name
+        self.benchmark_eval_field = eval_field
+        self.benchmark_input_field = input_field
+
+
+
+    def inspect(self):
+        """Inspect the Flock instance."""
+        logger.info(f"Inspecting Flock instance '{self.name}' with start agent '{self.benchmark_agent_name}' and input '{input}'.")
+
+        async def run(input: dict[str, Any])-> dict[str, Any]:
+            """Inspect the Flock instance."""
+            logger.info(f"Inspecting Flock instance '{self.name}' with start agent '{self.benchmark_agent_name}' and input '{input}'.")
+            msg_content = input.get("messages")[0].get("content")
+
+            agent_input = {
+                self.benchmark_input_field: msg_content
+            }
+
+            result = await self.run_async(start_agent=self.benchmark_agent_name, input=agent_input, box_result=False)
+
+            agent_output = result.get(self.benchmark_eval_field, "No answer found")
+
+            return {
+                "output": agent_output,
+            }
+
+        return run
+
     def _configure_logging(self, enable_logging_config: bool | list[str]):
         """Configure logging levels based on the enable_logging flag."""
         is_enabled_globally = False
@@ -302,45 +358,37 @@ class Flock(BaseModel, Serializable):
         agents: list[FlockAgent] | None = None,
     ) -> Box | dict:
         """Entry point for running an agent system synchronously."""
+        # Prepare the coroutine that needs to be run
+        coro = self.run_async(
+            start_agent=start_agent,
+            input=input,
+            context=context,
+            run_id=run_id,
+            box_result=box_result,  # run_async handles boxing
+            agents=agents,
+        )
+
         try:
+            # Check if an event loop is already running in the current thread
             loop = asyncio.get_running_loop()
-            if loop.is_closed():
-                raise RuntimeError("Event loop is closed")
-        except RuntimeError:  # No running loop
-            loop = asyncio.new_event_loop()
-            asyncio.set_event_loop(loop)
-
-        # Ensure the loop runs the task and handles closure if we created it
-        if asyncio.get_event_loop() is loop and not loop.is_running():
-            result = loop.run_until_complete(
-                self.run_async(
-                    start_agent=start_agent,
-                    input=input,
-                    context=context,
-                    run_id=run_id,
-                    box_result=box_result,
-                    agents=agents,
-                )
-            )
-            return result
+        except RuntimeError:
+            # No event loop is running in the current thread.
+            # We can safely use asyncio.run() to create a new loop,
+            # run the coroutine, and close the loop.
+            return asyncio.run(coro)
         else:
-            # If called from an already running loop (e.g. Jupyter, FastAPI endpoint)
-            # Create a future and run it to completion. This is tricky and
-            # ideally, one should use `await self.run_async` directly in async contexts.
-            # This simple run_until_complete on a future might block the existing loop.
-            # For truly non-blocking execution in an existing loop, one might need
-            # to schedule the coroutine differently or advise users to use `await`.
-            future = asyncio.ensure_future(
-                self.run_async(
-                    start_agent=start_agent,
-                    input=input,
-                    context=context,
-                    run_id=run_id,
-                    box_result=box_result,
-                    agents=agents,
-                )
+            # An event loop is already running in the current thread.
+            # Calling loop.run_until_complete() or asyncio.run() here would raise an error.
+            # To run the async code and wait for its result synchronously,
+            # we execute it in a separate thread with its own event loop.
+            logger.debug(
+                "Flock.run called in a context with an existing event loop. "
+                "Running async task in a separate thread to avoid event loop conflict."
             )
-            return loop.run_until_complete(future)
+            with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
+                future = executor.submit(asyncio.run, coro)
+                # Block and wait for the result from the other thread
+                return future.result()
 
 
     async def run_async(
@@ -555,14 +603,6 @@ class Flock(BaseModel, Serializable):
         delimiter: str = ",",
     ) -> list[Box | dict | None | Exception]:
         """Synchronous wrapper for run_batch_async."""
-        try:
-            loop = asyncio.get_running_loop()
-            if loop.is_closed():
-                raise RuntimeError("Event loop is closed")
-        except RuntimeError:  # No running loop
-            loop = asyncio.new_event_loop()
-            asyncio.set_event_loop(loop)
-
         coro = self.run_batch_async(
             start_agent=start_agent,
             batch_inputs=batch_inputs,
@@ -579,12 +619,18 @@ class Flock(BaseModel, Serializable):
             delimiter=delimiter,
         )
 
-        if asyncio.get_event_loop() is loop and not loop.is_running():
-            results = loop.run_until_complete(coro)
-            return results
+        try:
+            loop = asyncio.get_running_loop()
+        except RuntimeError:
+            return asyncio.run(coro)
         else:
-            future = asyncio.ensure_future(coro)
-            return loop.run_until_complete(future)
+            logger.debug(
+                "Flock.run_batch called in a context with an existing event loop. "
+                "Running async task in a separate thread."
+            )
+            with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
+                future = executor.submit(asyncio.run, coro)
+                return future.result()
 
     # --- Evaluation (Delegation) ---
     async def evaluate_async(
@@ -659,14 +705,6 @@ class Flock(BaseModel, Serializable):
         metadata_columns: list[str] | None = None,
     ) -> DataFrame | list[dict[str, Any]]: # type: ignore
         """Synchronous wrapper for evaluate_async."""
-        try:
-            loop = asyncio.get_running_loop()
-            if loop.is_closed():
-                raise RuntimeError("Event loop is closed")
-        except RuntimeError:  # No running loop
-            loop = asyncio.new_event_loop()
-            asyncio.set_event_loop(loop)
-
         coro = self.evaluate_async(
             dataset=dataset,
             start_agent=start_agent,
@@ -685,12 +723,18 @@ class Flock(BaseModel, Serializable):
             metadata_columns=metadata_columns,
         )
 
-        if asyncio.get_event_loop() is loop and not loop.is_running():
-            results = loop.run_until_complete(coro)
-            return results
+        try:
+            loop = asyncio.get_running_loop()
+        except RuntimeError:
+            return asyncio.run(coro)
         else:
-            future = asyncio.ensure_future(coro)
-            return loop.run_until_complete(future)
+            logger.debug(
+                "Flock.evaluate called in a context with an existing event loop. "
+                "Running async task in a separate thread."
+            )
+            with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
+                future = executor.submit(asyncio.run, coro)
+                return future.result()
 
     # --- Server & CLI Starters (Delegation) ---
     def start_api(