Fix tool calls parsing for definitions containing arrays

src/promptflow-core/CHANGELOG.md

@@ -4,6 +4,7 @@
 ### Bugs fixed
 - Fix ChatUI can't work in docker container when running image build with `pf flow build`.
 - Fix [#3355](https://github.com/microsoft/promptflow/issues/3355) that IndexError is raised when generator is used in a flow and the flow is called inside another flow.
+- Fix [#3388](https://github.com/microsoft/promptflow/issues/3388) that tool calls definitions with internal arrays failed parsing
 
 ## v1.11.0 (2024.05.17)
 

src/promptflow-core/promptflow/core/_prompty_utils.py

@@ -653,11 +653,17 @@ def try_parse_tool_call_id_and_content(role_prompt):
 def try_parse_tool_calls(role_prompt):
     # customer can add ## in front of tool_calls for markdown highlight.
     # and we still support tool_calls without ## prefix for backward compatibility.
-    pattern = r"\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*(\[.*?\])"
+
+    # Previously used pattern (commented) cannot parse tool_calls with internal square brackets
+    # common with function call arguments. Updated pattern matches content before tool_calls array
+    # and removes it before parsing the array.
+    # pattern = r"\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*(\[.*?\])"
+    pattern = r"(\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*)\["
     match = re.search(pattern, role_prompt, re.DOTALL)
     if match:
         try:
-            parsed_array = eval(match.group(1))
+            stripped_prompt = role_prompt.replace(match.group(1), "")
+            parsed_array = eval(stripped_prompt)
             return parsed_array
         except Exception:
             None

src/promptflow-tools/CHANGELOG.md

@@ -7,6 +7,9 @@
 - Avoid unintended parsing by role process to user flow inputs in prompt templates.
 - Introduce universal contract LLM tool and combine "Azure OpenAI GPT-4 Turbo with Vision" and "OpenAI GPT-4V" tools to "LLM-Vision" tool.
 
+### Bugs Fixed
+- Fix [#3388](https://github.com/microsoft/promptflow/issues/3388) that tool calls definitions with internal arrays failed parsing
+
 ## 1.4.0 (2024.03.26)
 
 ### Features Added

src/promptflow-tools/promptflow/tools/common.py

@@ -241,11 +241,35 @@ def try_parse_tool_call_id_and_content(role_prompt):
 def try_parse_tool_calls(role_prompt):
     # customer can add ## in front of tool_calls for markdown highlight.
     # and we still support tool_calls without ## prefix for backward compatibility.
-    pattern = r"\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*(\[.*?\])"
+
+    # Previously used pattern (commented) cannot parse tool_calls with internal square brackets
+    # common with function call arguments. Updated pattern matches content before tool_calls array
+    # and removes it before parsing the array.
+    # pattern = r"\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*(\[.*?\])"
+    # match = re.search(pattern, role_prompt, re.DOTALL)
+    pattern = r"(\s*\n*#{0,2}\s*tool_calls\s*:\s*\n+\s*)\["
     match = re.search(pattern, role_prompt, re.DOTALL)
     if match:
         try:
-            parsed_array = eval(match.group(1))
+            stripped_prompt = role_prompt.replace(match.group(1), "")
+            # Find outer array brackets starting from [;
+            # Required for incomplete chunk parsing due to role naming errors
+            open_brackets = 1
+            last_open = 0
+            last_close = -1
+            # Loop until finding closed outer brackets or no more brackets found
+            while open_brackets != 0 and not (last_close == -1 and last_open == -1):
+                next_open = stripped_prompt.find("[", last_open + 1)
+                next_close = stripped_prompt.find("]", last_close + 1)
+                if next_open == -1 or next_open > next_close:
+                    open_brackets -= 1
+                    last_close = next_close
+                else:
+                    open_brackets += 1
+                    last_open = next_open
+            stripped_prompt = stripped_prompt[:last_close+1]
+
+            parsed_array = eval(stripped_prompt)
             return parsed_array
         except Exception:
             None

src/promptflow-tools/tests/test_common.py

@@ -249,6 +249,22 @@ def test_try_parse_chat_with_tools(self, example_prompt_template_with_tool, pars
          ("tool_calls:\r\n[", None),
          ("tool_calls:\r\n[{'id': 'tool_call_id', 'type': 'function', 'function': {'name': 'func1', 'arguments': ''}}]",
           [{'id': 'tool_call_id', 'type': 'function', 'function': {'name': 'func1', 'arguments': ''}}]),
+         ("tool_calls:\r\n[{'id': 'tool_call_id', 'type': 'function',"
+          " 'function': {'name': 'func1', 'arguments': '{\"arg1\": []}'}}]",
+          [{'id': 'tool_call_id', 'type': 'function', 'function': {'name': 'func1', 'arguments': '{"arg1": []}'}}]),
+         ("tool_calls:\r\n[{'id': 'tool_call_id', 'type': 'function',"
+          " 'function': {'name': 'func1', 'arguments': '{\"arg1\": [{\"nested\": []}]}'}}]",
+          [{'id': 'tool_call_id', 'type': 'function', 'function':
+              {'name': 'func1', 'arguments': '{"arg1": [{"nested": []}]}'}}]),
+         ("tool_calls:\r\n[{'id': 'tool_call_id', 'type': 'function',"
+          " 'function': {'name': 'func1', 'arguments': '{\"arg1\": [{\"nested\": []}, {\"nested_2\": [[]]}]}'}}]",
+          [{'id': 'tool_call_id', 'type': 'function', 'function':
+              {'name': 'func1', 'arguments': '{"arg1": [{"nested": []}, {"nested_2": [[]]}]}'}}]),
+         ("tool_calls:\r\n[{'id': 'tool_call_id', 'type': 'function',"
+          " 'function': {'name': 'func1', 'arguments': '{\"arg1\": [{\"nested\": []}, {\"nested_2\": [[]]}]}'}}]"
+          "\n   #tool",
+          [{'id': 'tool_call_id', 'type': 'function', 'function':
+              {'name': 'func1', 'arguments': '{"arg1": [{"nested": []}, {"nested_2": [[]]}]}'}}]),
          ("tool_calls:\n[{'id': 'tool_call_id', 'type': 'function', 'function': {'name': 'func1', 'arguments': ''}}]",
           [{'id': 'tool_call_id', 'type': 'function', 'function': {'name': 'func1', 'arguments': ''}}])])
     def test_try_parse_tool_calls(self, role_prompt, expected_result):