Improve on agent codes; add auto-running experiments code; Fix some examples

mm_agents/gpt_4v_agent.py

@@ -63,6 +63,8 @@ def tag_screenshot(screenshot, accessibility_tree):
 
 
 def parse_actions_from_string(input_string):
+    if input_string.strip() in ['WAIT', 'DONE', 'FAIL']:
+        return [input_string.strip()]
     # Search for a JSON string within the input string
     actions = []
     matches = re.findall(r'```json\s+(.*?)\s+```', input_string, re.DOTALL)
@@ -95,6 +97,9 @@ def parse_actions_from_string(input_string):
 
 
 def parse_code_from_string(input_string):
+    if input_string.strip() in ['WAIT', 'DONE', 'FAIL']:
+        return [input_string.strip()]
+
     # This regular expression will match both ```code``` and ```python code```
     # and capture the `code` part. It uses a non-greedy match for the content inside.
     pattern = r"```(?:\w+\s+)?(.*?)```"