Streamlit web app for nft game #590
Conversation
WalkthroughThe pull request introduces a new enumeration class Changes
Possibly related PRs
Suggested reviewers
📜 Recent review detailsConfiguration used: CodeRabbit UI 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 7
🧹 Outside diff range and nitpick comments (10)
prediction_market_agent/agents/identifiers.py (2)
4-19: Consider organizing identifiers into logical groupsThe enum contains multiple related identifiers that could benefit from better organization. Consider grouping them using either:
- Separate enums for each category (THINK_THOROUGHLY, MICROCHAIN, NFT_GAME)
- Comments to visually separate groups
- Consistent prefix/suffix patterns
This would improve maintainability as more identifiers are added.
Example organization:
class AgentIdentifier(str, Enum): # Think Thoroughly Agents THINK_THOROUGHLY = "think-thoroughly-agent" THINK_THOROUGHLY_PROPHET = "think-thoroughly-prophet-research-agent" # Microchain Agents MICROCHAIN_AGENT_OMEN = "microchain-agent-deployment-omen" MICROCHAIN_AGENT_OMEN_TEST = "microchain-agent-deployment-omen_test" MICROCHAIN_AGENT_OMEN_LEARNING_0 = "general-agent-0" # ... other microchain agents ... # NFT Game Agents NFT_GAME_AGENT_0 = "nft-game-agent-0" NFT_GAME_AGENT_1 = "nft-game-agent-1" # ... other NFT game agents ...
8-8: Inconsistent naming in string valueThe test agent identifier uses an underscore in its string value while others use hyphens:
"microchain-agent-deployment-omen_test"vs"microchain-agent-deployment-omen"This inconsistency could cause issues with string matching or filtering.
- MICROCHAIN_AGENT_OMEN_TEST = "microchain-agent-deployment-omen_test" + MICROCHAIN_AGENT_OMEN_TEST = "microchain-agent-deployment-omen-test"scripts/delete_agent_db_entries.py (1)
Line range hint
12-17: Add type hints and improve function documentationThe main function's documentation could be more descriptive and include type hints.
def main( session_id: AgentIdentifier, delete_memories: bool = True, delete_prompts: bool = True, delete_goals: bool = True, ) -> None: - """ - Delete all memories and prompts for a given agent, defined by the session_id. - """ + """ + Delete database entries for a specified agent. + + Args: + session_id (AgentIdentifier): The identifier of the agent + delete_memories (bool, optional): Delete memory entries. Defaults to True. + delete_prompts (bool, optional): Delete prompt entries. Defaults to True. + delete_goals (bool, optional): Delete goal entries. Defaults to True. + + Raises: + Exception: If any entries fail to delete + """prediction_market_agent/agents/microchain_agent/agents_nft_game.py (2)
19-70: Consider moving agent configurations to a separate fileThe hardcoded agent configurations might be better placed in a separate configuration file for easier maintenance and updates.
Consider creating a new file
agent_configs.pyand moving the configurations there:from typing import List from prediction_market_agent.agents.microchain_agent.agents_nft_game import NFTAgent AGENTS_IN_THE_GAME: List[NFTAgent] = [ # ... agent configurations ... ]
71-76: Consider environment variables for blockchain addressesHardcoding blockchain addresses directly in the source code might make it difficult to manage different environments (testnet vs mainnet).
Consider using environment variables or a configuration system that supports different environments:
from prediction_market_agent_tooling.config import get_config TREASURY_SAFE_ADDRESS = Web3.to_checksum_address( get_config("TREASURY_SAFE_ADDRESS") ) NFT_KEYS_ADDRESS = Web3.to_checksum_address( get_config("NFT_KEYS_ADDRESS") )prediction_market_agent/agents/microchain_agent/app_nft_game.py (5)
1-5: Enhance documentation with additional setup details.Consider adding more details to the documentation:
- Required environment variables
- Dependencies installation instructions
- Expected behavior when running the app
34-44: Add error handling for database connection failures.The cached handler functions should include try-except blocks to gracefully handle potential database connection failures.
@st.cache_resource def long_term_memory_table_handler( identifier: AgentIdentifier, ) -> LongTermMemoryTableHandler: + try: return LongTermMemoryTableHandler.from_agent_identifier(identifier) + except Exception as e: + st.error(f"Failed to connect to memory database: {str(e)}") + return None
90-90: Remove unnecessary f-string.The f-string doesn't contain any placeholders.
- st.markdown(f"""### Agent's actions""") + st.markdown("""### Agent's actions""")🧰 Tools
🪛 Ruff (0.8.0)
90-90: f-string without any placeholders
Remove extraneous
fprefix(F541)
89-107: Implement pagination for better performance.Loading all actions at once could be slow with a large history. Consider implementing pagination.
def show_function_calls_part(nft_agent: NFTAgent) -> None: st.markdown("""### Agent's actions""") + page_size = 10 + page = st.number_input("Page", min_value=1, value=1) with st.spinner("Loading agent's actions..."): - calls = long_term_memory_table_handler(nft_agent.identifier).search() + calls = long_term_memory_table_handler(nft_agent.identifier).search( + limit=page_size, + offset=(page - 1) * page_size + )🧰 Tools
🪛 Ruff (0.8.0)
90-90: f-string without any placeholders
Remove extraneous
fprefix(F541)
142-147: Add error handling for treasury balance display.The treasury balance fetching should handle potential failures gracefully.
-treasury_xdai_balance = get_balances(TREASURY_SAFE_ADDRESS).xdai +try: + treasury_xdai_balance = get_balances(TREASURY_SAFE_ADDRESS).xdai +except Exception as e: + st.sidebar.error(f"Failed to fetch treasury balance: {str(e)}") + treasury_xdai_balance = 0
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (13)
prediction_market_agent/agents/identifiers.py(1 hunks)prediction_market_agent/agents/microchain_agent/agents_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/app.py(4 hunks)prediction_market_agent/agents/microchain_agent/app_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/deploy.py(5 hunks)prediction_market_agent/agents/social_media_agent/deploy.py(2 hunks)prediction_market_agent/agents/think_thoroughly_agent/think_thoroughly_agent.py(5 hunks)prediction_market_agent/agents/utils.py(0 hunks)prediction_market_agent/db/long_term_memory_table_handler.py(2 hunks)prediction_market_agent/db/models.py(2 hunks)prediction_market_agent/db/prompt_table_handler.py(1 hunks)scripts/delete_agent_db_entries.py(2 hunks)scripts/deployed_general_agent_viewer.py(4 hunks)
💤 Files with no reviewable changes (1)
- prediction_market_agent/agents/utils.py
🧰 Additional context used
🪛 Ruff (0.8.0)
prediction_market_agent/agents/microchain_agent/app_nft_game.py
90-90: f-string without any placeholders
Remove extraneous f prefix
(F541)
🔇 Additional comments (11)
prediction_market_agent/db/prompt_table_handler.py (2)
20-24: LGTM: Factory method enhances code organization
The new factory method provides a clean way to create handlers from agent identifiers, improving code organization and type safety.
14-14: Breaking change: session_identifier is now required
The constructor parameter session_identifier has been changed from optional to required. This is a breaking change that might affect existing code.
✅ Verification successful
Let me check the test file to see how the constructor is being used and if there are any instances where the session_identifier is omitted.
Let me check if there are any other usages of PromptTableHandler in the codebase, particularly focusing on the actual implementation to see the before/after state of the constructor.
Let me try to see the current implementation of the file to understand the constructor's definition.
Based on the evidence gathered:
- The only usages of
PromptTableHandlerconstructor are in tests and through the new factory methodfrom_agent_identifier - The tests already provide the required
session_identifierparameter - The new factory method ensures the parameter is always provided
- There are no other direct instantiations of
PromptTableHandlerin the codebase
The session_identifier parameter change is safe
The change to make session_identifier required is not a breaking change as all existing usages already provide this parameter, and new usages are guided through the factory method from_agent_identifier.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Search for existing instantiations that might be affected
rg "PromptTableHandler\(" --type py
Length of output: 238
Script:
#!/bin/bash
# Check the test file content
cat tests/db/test_prompt_table_handler.py
Length of output: 1330
Script:
#!/bin/bash
# Check the implementation file to see constructor changes
git diff prediction_market_agent/db/prompt_table_handler.py
Length of output: 61
Script:
#!/bin/bash
# Check the current implementation
cat prediction_market_agent/db/prompt_table_handler.py
Length of output: 2025
prediction_market_agent/agents/microchain_agent/agents_nft_game.py (1)
8-16: LGTM: Well-structured agent model
The NFTAgent class is well-designed using Pydantic for type safety and validation. The url property provides a clean way to generate URL-friendly names.
prediction_market_agent/db/long_term_memory_table_handler.py (1)
22-26: LGTM: Consistent factory method implementation
The new factory method aligns well with the similar implementation in PromptTableHandler, maintaining consistency across the codebase.
prediction_market_agent/agents/social_media_agent/deploy.py (1)
53-55: LGTM! Clean transition to AgentIdentifier.
The change from direct instantiation to using from_agent_identifier is clean and consistent with the codebase-wide refactoring.
scripts/deployed_general_agent_viewer.py (2)
30-30: LGTM: Import statement updated correctly.
The import statement has been updated to use the new location for AgentIdentifier, aligning with the codebase's restructuring.
Line range hint 128-130: Verify initialization of LongTermMemoryTableHandler.
The initialization has been updated to use the new from_agent_identifier factory method. Let's verify that all agent identifiers are properly handled.
✅ Verification successful
All agent identifiers are properly handled through the enum class
Based on the search results, I can verify that:
- The
AgentIdentifieris properly defined as an enum class inprediction_market_agent/agents/identifiers.py - The
from_agent_identifierfactory method inLongTermMemoryTableHandlercorrectly accepts anAgentIdentifierenum and uses its value as the task description - All usages of
from_agent_identifieracross the codebase consistently pass validAgentIdentifierenum values - The specific usage in
scripts/deployed_general_agent_viewer.pyis correct as it uses a valid agent identifier selected from the available enum values
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify that all agent identifiers used in the codebase are valid enum values
# Search for all agent identifiers used with LongTermMemoryTableHandler
ast-grep --pattern 'LongTermMemoryTableHandler.from_agent_identifier($identifier)'
# Search for all AgentIdentifier enum values
ast-grep --pattern 'class AgentIdentifier'
Length of output: 129
Script:
#!/bin/bash
# Let's try a different approach to find the usage and definition
# Search for LongTermMemoryTableHandler usage
rg "LongTermMemoryTableHandler" -A 3
# Search for AgentIdentifier usage and definition
rg "AgentIdentifier" -A 3
# Search for from_agent_identifier usage
rg "from_agent_identifier" -A 3
Length of output: 37850
prediction_market_agent/agents/microchain_agent/app.py (2)
33-33: LGTM: Import statement added correctly.
The AgentIdentifier import has been properly added to support the enum-based agent identification.
146-146: Verify PromptTableHandler initialization.
The initialization has been updated to use the new from_agent_identifier factory method. This change aligns with the new pattern of using enumerated identifiers.
✅ Verification successful
Initialization of PromptTableHandler is consistent with the new pattern
The verification shows that all usages of PromptTableHandler across the codebase consistently use the from_agent_identifier factory method with AgentIdentifier enums, including:
prediction_market_agent/agents/microchain_agent/app.pyprediction_market_agent/agents/microchain_agent/deploy.pyprediction_market_agent/agents/microchain_agent/app_nft_game.pyscripts/delete_agent_db_entries.py
The implementation aligns with the intended pattern of using enumerated identifiers.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify that PromptTableHandler initialization is consistent across the codebase
# Search for all PromptTableHandler initializations
ast-grep --pattern 'PromptTableHandler.from_agent_identifier($identifier)'
Length of output: 76
Script:
#!/bin/bash
# Let's try a broader search for PromptTableHandler usage
rg "PromptTableHandler" -A 2 -B 2
# Also search for the specific method
rg "from_agent_identifier" -A 2 -B 2
# And look for the actual class definition
ast-grep --pattern 'class PromptTableHandler {
$$$
}'
Length of output: 12979
prediction_market_agent/agents/think_thoroughly_agent/think_thoroughly_agent.py (2)
107-107: LGTM: Type annotation updated correctly.
The identifier type has been properly updated to use the new AgentIdentifier enum type, improving type safety.
117-119: LGTM: Memory handler initialization updated.
The long-term memory handler initialization has been properly updated to use the new from_agent_identifier factory method with appropriate null handling.
prediction_market_agent/db/models.py
Outdated
| @property | ||
| def metadata_dict(self) -> dict[str, Any] | None: | ||
| out: dict[str, Any] | None = ( | ||
| json.loads(self.metadata_) if self.metadata_ else None | ||
| ) | ||
| return out |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Add error handling and documentation for metadata_dict property
The property lacks error handling for JSON parsing and documentation explaining its purpose.
Consider this implementation:
@property
def metadata_dict(self) -> dict[str, Any] | None:
+ """
+ Converts the metadata_ JSON string to a dictionary.
+
+ Returns:
+ dict[str, Any] | None: Parsed metadata dictionary or None if metadata_ is not set
+
+ Raises:
+ json.JSONDecodeError: If metadata_ contains invalid JSON
+ """
out: dict[str, Any] | None = (
- json.loads(self.metadata_) if self.metadata_ else None
+ None if not self.metadata_ else json.loads(self.metadata_)
)
return out📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| @property | |
| def metadata_dict(self) -> dict[str, Any] | None: | |
| out: dict[str, Any] | None = ( | |
| json.loads(self.metadata_) if self.metadata_ else None | |
| ) | |
| return out | |
| @property | |
| def metadata_dict(self) -> dict[str, Any] | None: | |
| """ | |
| Converts the metadata_ JSON string to a dictionary. | |
| Returns: | |
| dict[str, Any] | None: Parsed metadata dictionary or None if metadata_ is not set | |
| Raises: | |
| json.JSONDecodeError: If metadata_ contains invalid JSON | |
| """ | |
| out: dict[str, Any] | None = ( | |
| None if not self.metadata_ else json.loads(self.metadata_) | |
| ) | |
| return out |
There was a problem hiding this comment.
Good point @RabbitAI, we should add error handling for JSON parsing since DB can hold any string in that column
There was a problem hiding this comment.
Seems like the humans are having a chat. I'll hop back into my burrow for now. If you need me again, just tag @coderabbitai in a new comment, and I'll come hopping out!
There was a problem hiding this comment.
@gabrielfior What do you want to handle there? If it's not JSON it will error out and that's exactly what we want it to do, because it means a bug.
There was a problem hiding this comment.
I was expecting a more graceful error handling, instead of what json.loads gives us.
scripts/delete_agent_db_entries.py
Outdated
| prompt_handler = PromptTableHandler.from_agent_identifier(session_id) | ||
| prompt_handler.delete_all_prompts() | ||
| if prompt_handler.fetch_latest_prompt() is not None: | ||
| raise Exception("Prompt entries were not deleted.") | ||
| else: |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Standardize handler initialization and improve error messages
The script shows inconsistent patterns in handler initialization and could benefit from more informative messages.
Consider these improvements:
- Standardize handler initialization:
- evaluated_goal_table_handler = EvaluatedGoalTableHandler(agent_id=session_id)
+ evaluated_goal_table_handler = EvaluatedGoalTableHandler.from_agent_identifier(session_id)- Add more informative messages:
- raise Exception("Prompt entries were not deleted.")
+ raise Exception(f"Failed to delete prompt entries for agent {session_id.value}")
else:
- print("Prompt entries successfully deleted.")
+ print(f"Successfully deleted all prompt entries for agent {session_id.value}")- Consider adding a count of deleted entries in success messages:
deleted_count = len(prompt_handler.fetch_all_prompts())
prompt_handler.delete_all_prompts()
print(f"Successfully deleted {deleted_count} prompt entries for agent {session_id.value}")Also applies to: 31-35, 39-44
| if role == "assistant": | ||
| parsed_function = message.split("(")[0] | ||
| parsed_body = message.split("(")[1].rsplit(")")[0] |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Improve function parsing robustness.
Current string splitting approach is fragile and could break with complex messages containing parentheses.
Consider using regex or a proper parser:
- parsed_function = message.split("(")[0]
- parsed_body = message.split("(")[1].rsplit(")")[0]
+ import re
+ if match := re.match(r"(\w+)\((.*)\)$", message.strip()):
+ parsed_function, parsed_body = match.groups()
+ else:
+ parsed_function = "Unknown"
+ parsed_body = message📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| if role == "assistant": | |
| parsed_function = message.split("(")[0] | |
| parsed_body = message.split("(")[1].rsplit(")")[0] | |
| if role == "assistant": | |
| import re | |
| if match := re.match(r"(\w+)\((.*)\)$", message.strip()): | |
| parsed_function, parsed_body = match.groups() | |
| else: | |
| parsed_function = "Unknown" | |
| parsed_body = message |
| def show_about_agent_part(nft_agent: NFTAgent) -> None: | ||
| system_prompt = prompt_table_handler(nft_agent.identifier).fetch_latest_prompt() | ||
| xdai_balance = get_balances(nft_agent.wallet_address).xdai | ||
| st.markdown( | ||
| f"""### {nft_agent.name} | ||
|
|
||
| Currently holds <span style='font-size: 1.1em;'><strong>{xdai_balance:.2f} xDAI</strong></span>. | ||
|
|
||
| --- | ||
| """, | ||
| unsafe_allow_html=True, | ||
| ) | ||
| st.text_area( | ||
| f"{nft_agent.name}'s system prompt", | ||
| value=system_prompt.prompt if system_prompt else "No system prompt yet.", | ||
| disabled=True, | ||
| ) | ||
|
|
There was a problem hiding this comment.
Improve security and error handling.
- Using
unsafe_allow_html=Truecould expose XSS vulnerabilities - No error handling for balance fetching
Consider these improvements:
def show_about_agent_part(nft_agent: NFTAgent) -> None:
system_prompt = prompt_table_handler(nft_agent.identifier).fetch_latest_prompt()
- xdai_balance = get_balances(nft_agent.wallet_address).xdai
+ try:
+ xdai_balance = get_balances(nft_agent.wallet_address).xdai
+ except Exception as e:
+ st.error(f"Failed to fetch balance: {str(e)}")
+ xdai_balance = 0
+
+ # Use st.markdown's native formatting instead of HTML
st.markdown(
- f"""### {nft_agent.name}
-
-Currently holds <span style='font-size: 1.1em;'><strong>{xdai_balance:.2f} xDAI</strong></span>.
-
----
-""",
- unsafe_allow_html=True,
+ f"""### {nft_agent.name}\n\nCurrently holds **{xdai_balance:.2f} xDAI**.\n\n---"""
)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| def show_about_agent_part(nft_agent: NFTAgent) -> None: | |
| system_prompt = prompt_table_handler(nft_agent.identifier).fetch_latest_prompt() | |
| xdai_balance = get_balances(nft_agent.wallet_address).xdai | |
| st.markdown( | |
| f"""### {nft_agent.name} | |
| Currently holds <span style='font-size: 1.1em;'><strong>{xdai_balance:.2f} xDAI</strong></span>. | |
| --- | |
| """, | |
| unsafe_allow_html=True, | |
| ) | |
| st.text_area( | |
| f"{nft_agent.name}'s system prompt", | |
| value=system_prompt.prompt if system_prompt else "No system prompt yet.", | |
| disabled=True, | |
| ) | |
| def show_about_agent_part(nft_agent: NFTAgent) -> None: | |
| system_prompt = prompt_table_handler(nft_agent.identifier).fetch_latest_prompt() | |
| try: | |
| xdai_balance = get_balances(nft_agent.wallet_address).xdai | |
| except Exception as e: | |
| st.error(f"Failed to fetch balance: {str(e)}") | |
| xdai_balance = 0 | |
| # Use st.markdown's native formatting instead of HTML | |
| st.markdown( | |
| f"""### {nft_agent.name}\n\nCurrently holds **{xdai_balance:.2f} xDAI**.\n\n---""" | |
| ) | |
| st.text_area( | |
| f"{nft_agent.name}'s system prompt", | |
| value=system_prompt.prompt if system_prompt else "No system prompt yet.", | |
| disabled=True, | |
| ) |
| def send_message_part(nft_agent: NFTAgent) -> None: | ||
| message = st.text_area("Write a message to the agent") | ||
|
|
||
| if st.button("Send message", disabled=not message): | ||
| # TODO: Don't do this manually with deployment private key, use the user's wallet! | ||
| SendPaidMessageToAnotherAgent()(nft_agent.wallet_address, message) | ||
| st.success("Message sent and will be processed soon!") | ||
|
|
There was a problem hiding this comment.
Implement input validation and secure wallet integration.
Current implementation has several concerns:
- No input validation for the message content
- Manual wallet handling (as noted in TODO)
- No rate limiting for message sending
Consider implementing:
def send_message_part(nft_agent: NFTAgent) -> None:
message = st.text_area("Write a message to the agent")
+ if message and len(message) > 1000: # Example limit
+ st.error("Message too long. Maximum 1000 characters allowed.")
+ return
if st.button("Send message", disabled=not message):
+ if not message.strip():
+ st.error("Message cannot be empty")
+ return
# TODO: Don't do this manually with deployment private key, use the user's wallet!
SendPaidMessageToAnotherAgent()(nft_agent.wallet_address, message)
st.success("Message sent and will be processed soon!")📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| def send_message_part(nft_agent: NFTAgent) -> None: | |
| message = st.text_area("Write a message to the agent") | |
| if st.button("Send message", disabled=not message): | |
| # TODO: Don't do this manually with deployment private key, use the user's wallet! | |
| SendPaidMessageToAnotherAgent()(nft_agent.wallet_address, message) | |
| st.success("Message sent and will be processed soon!") | |
| def send_message_part(nft_agent: NFTAgent) -> None: | |
| message = st.text_area("Write a message to the agent") | |
| if message and len(message) > 1000: # Example limit | |
| st.error("Message too long. Maximum 1000 characters allowed.") | |
| return | |
| if st.button("Send message", disabled=not message): | |
| if not message.strip(): | |
| st.error("Message cannot be empty") | |
| return | |
| # TODO: Don't do this manually with deployment private key, use the user's wallet! | |
| SendPaidMessageToAnotherAgent()(nft_agent.wallet_address, message) | |
| st.success("Message sent and will be processed soon!") |
| @@ -38,7 +38,7 @@ class DeployableMicrochainAgent(DeployableAgent): | |||
| n_iterations = 50 | |||
| load_historical_prompt: bool = False | |||
| system_prompt_choice: SystemPromptChoice = SystemPromptChoice.TRADING_AGENT | |||
| task_description = AgentIdentifier.MICROCHAIN_AGENT_OMEN | |||
| identifier: AgentIdentifier = AgentIdentifier.MICROCHAIN_AGENT_OMEN | |||
There was a problem hiding this comment.
💡 Codebase verification
The transition from task_description to identifier is not complete
The codebase still contains multiple references to task_description in the following locations:
prediction_market_agent/db/models.py: Model definition still uses task_descriptionprediction_market_agent/db/long_term_memory_table_handler.py: Multiple occurrences in class methods and database operations- Several test files still use task_description:
tests/test_chat_history.pytests/agents/microchain/test_functions.pytests/db/test_long_term_memory_table_handler.py
🔗 Analysis chain
Verify the complete transition from task_description to identifier.
The changes look good. Let's verify that all occurrences of task_description have been replaced.
Also applies to: 66-66, 69-72, 142-142, 149-149, 156-156, 163-163, 171-171, 181-181
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify no remaining task_description references
# Search for any remaining references to task_description
rg "task_description" --type python
Length of output: 66
Script:
#!/bin/bash
# Search for any remaining references to task_description without file type restriction
rg "task_description"
Length of output: 1745
| @@ -373,7 +376,7 @@ def answer_binary_market( | |||
|
|
|||
|
|
|||
| class ThinkThoroughlyWithItsOwnResearch(ThinkThoroughlyBase): | |||
| identifier = "think-thoroughly-agent" | |||
| identifier = AgentIdentifier.THINK_THOROUGHLY | |||
There was a problem hiding this comment.
💡 Codebase verification
Model versions need to be updated to match supported versions
The codebase shows inconsistencies in model versioning:
gpt-4-turbo-2024-04-09is not a valid OpenAI model versiongpt-4o-2024-08-06appears to be a custom alias defined inSupportedModelenum, but is used directly in some files
Suggested changes:
- Update ThinkThoroughlyWithItsOwnResearch to use
gpt-4-turbo-previeworgpt-4-0125-preview - Update ThinkThoroughlyWithPredictionProphetResearch to use the enum value
SupportedModel.gpt_4o.valueinstead of hardcoding the version - Update other occurrences of these model versions throughout the codebase
🔗 Analysis chain
Verify model versions and compatibility.
The agent classes specify different model versions:
- ThinkThoroughlyWithItsOwnResearch: gpt-4-turbo-2024-04-09
- ThinkThoroughlyWithPredictionProphetResearch: gpt-4-turbo-2024-04-09 and gpt-4o-2024-08-06
Let's verify these model versions are valid and supported.
Also applies to: 446-446
Let me gather more information about model validation and configuration.
Based on the search results, I found the SupportedModel enum in microchain_agent/microchain_agent.py that defines valid model versions. Let me check if the models used in ThinkThoroughlyAgent are consistent with OpenAI's model lineup.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify model versions used across the codebase
# Search for all model version strings
rg -A 1 'model.*=.*"gpt-'
# Search for model version validation logic
ast-grep --pattern 'class SupportedModel'
Length of output: 3230
Script:
#!/bin/bash
# Search for model validation or configuration
rg -A 3 'class.*Model'
rg -A 3 'SUPPORTED_MODELS'
rg -A 3 'MODEL_CONFIG'
Length of output: 13200
Script:
#!/bin/bash
# Get more context around model definitions and validations
rg -A 5 'gpt-4-turbo-2024-04-09|gpt-4o-2024-08-06' --context 5
Length of output: 7825
| from enum import Enum | ||
|
|
||
|
|
||
| class AgentIdentifier(str, Enum): |
There was a problem hiding this comment.
Moved to a new file because of circular imports.
| return self.name.lower().replace(" ", "-") | ||
|
|
||
|
|
||
| AGENTS_IN_THE_GAME = [ |
There was a problem hiding this comment.
I know it doesn't look like the best thing to have everything hard-coded here, but that's what PMA repository currently expects -- that each agent is defined in the code.
|
|
||
| @property | ||
| def url(self) -> str: | ||
| return self.name.lower().replace(" ", "-") |
There was a problem hiding this comment.
Will work as https://final-streamlit-url/agent-name
| @@ -124,7 +125,9 @@ def long_term_memory_is_initialized() -> bool: | |||
| def maybe_initialize_long_term_memory() -> None: | |||
| # Initialize the db storage | |||
| if not long_term_memory_is_initialized(): | |||
| st.session_state.long_term_memory = LongTermMemoryTableHandler(AGENT_IDENTIFIER) | |||
| st.session_state.long_term_memory = ( | |||
| LongTermMemoryTableHandler.from_agent_identifier(AGENT_IDENTIFIER) | |||
There was a problem hiding this comment.
We have a little identity issue with how to identify agents.
AgentIdentifier class exists, but in LongTermMemoryTableHandler it was passed into task_description argument, in PromptTableHandler it was passed to session_identifier argument., Somewhere it's used as string, somewhere as enum. Trying to consolidate it at least a little.
| message = st.text_area("Write a message to the agent") | ||
|
|
||
| if st.button("Send message", disabled=not message): | ||
| # TODO: Don't do this manually with deployment private key, use the user's wallet! |
There was a problem hiding this comment.
Created issue for this here #591
Turns out https://github.com/AlgoveraAI/streamlit-metamask is meant only for ERC-20 transfers (and didn't work on first try anyway 😄 ).
But I don't want to block the showcasing of this because of that, we can use this to show it off.
There was a problem hiding this comment.
I was hoping the streamlit-metamask extension was gonna work like a charm :(
Anyways, what do you think about adding 1 text input (RPC_URL) so that we are able to send transactions on a local chain, instead of burning xDAI on Mainnet during the demo?
There was a problem hiding this comment.
We are able to change the GNOSIS_RPC_URL variable in the deployment's environment. It doesn't help to do it only here if all the agents aren't deployed on the same local chain.
There was a problem hiding this comment.
GNOSIS_RPC_URL is probably good enough for now - I want to avoid doing demos, etc on the main chain, but I have other ideas. This can be closed.
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Outside diff range and nitpick comments (1)
prediction_market_agent/agents/microchain_agent/app_nft_game.py (1)
103-103: Remove unnecessary f-string prefixThe f-string on line 103 does not contain any placeholders, so the
fprefix is unnecessary.Apply this diff to remove the unnecessary prefix:
-st.markdown(f"""### Agent's actions""") +st.markdown("""### Agent's actions""")🧰 Tools
🪛 Ruff (0.8.0)
103-103: f-string without any placeholders
Remove extraneous
fprefix(F541)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (2)
prediction_market_agent/agents/microchain_agent/agents_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/app_nft_game.py(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (1)
- prediction_market_agent/agents/microchain_agent/agents_nft_game.py
🧰 Additional context used
🪛 Ruff (0.8.0)
prediction_market_agent/agents/microchain_agent/app_nft_game.py
103-103: f-string without any placeholders
Remove extraneous f prefix
(F541)
🔇 Additional comments (4)
prediction_market_agent/agents/microchain_agent/app_nft_game.py (4)
46-53: Implement input validation and secure wallet integration
The current implementation has several issues:
- No input validation for the message content.
- Manual wallet handling as noted in the TODO comment.
Consider implementing input validation to ensure the message is appropriate and integrating the user's wallet securely for sending messages.
61-62: Improve function parsing robustness
The current string splitting approach in parse_function_and_body may fail with complex messages containing parentheses.
Consider using regular expressions or a proper parser for more reliable parsing.
124-133: Improve security and error handling
- Using
unsafe_allow_html=Truemay expose XSS vulnerabilities. - No error handling for balance fetching.
Consider these improvements:
155-160:
Add error handling for balance fetching and avoid unsafe_allow_html
The code does not handle potential exceptions when fetching the treasury balance, and using unsafe_allow_html=True can lead to security risks.
Apply this diff to improve error handling and enhance security:
-# Fetch the treasury balance without error handling
-treasury_xdai_balance = get_balances(TREASURY_SAFE_ADDRESS).xdai
+try:
+ treasury_xdai_balance = get_balances(TREASURY_SAFE_ADDRESS).xdai
+except Exception as e:
+ st.error(f"Failed to fetch treasury balance: {e}")
+ treasury_xdai_balance = 0
-# Display the balance using unsafe HTML
st.sidebar.markdown(
- f"""### Treasury
-Currently holds <span style='font-size: 1.1em;'><strong>{treasury_xdai_balance:.2f} xDAI</strong></span>.""",
- unsafe_allow_html=True,
+ f"""### Treasury\nCurrently holds **{treasury_xdai_balance:.2f} xDAI**."""
)Likely invalid or redundant comment.
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Outside diff range and nitpick comments (2)
prediction_market_agent/agents/microchain_agent/agents_nft_game.py (2)
8-16: Consider enhancing the URL property implementation.While the current implementation works for basic cases, consider making it more robust:
- Handle special characters beyond spaces
- Add length validation for the name field
- Consider URL-encoding the result
@property def url(self) -> str: - return self.name.lower().replace(" ", "-") + import re + from urllib.parse import quote + # Remove special chars, convert spaces to hyphens, ensure single hyphens + clean_name = re.sub(r'[^\w\s-]', '', self.name.lower()) + clean_name = re.sub(r'[-\s]+', '-', clean_name).strip('-') + return quote(clean_name)
71-76: Add documentation for critical addresses.The
TREASURY_SAFE_ADDRESSandNFT_KEYS_ADDRESSappear to be critical system addresses. Consider adding docstrings or comments explaining:
- The purpose of each address
- The expected behavior/interactions
- Any security implications
+# Safe address holding treasury funds for the NFT game TREASURY_SAFE_ADDRESS = Web3.to_checksum_address( "0xd1A54FD7a200C2ca76B6D06437795d660d37FE28" ) +# Contract address for NFT key management NFT_KEYS_ADDRESS = Web3.to_checksum_address( "0x0D7C0Bd4169D090038c6F41CFd066958fe7619D0" )
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (1)
prediction_market_agent/agents/microchain_agent/agents_nft_game.py(1 hunks)
🔇 Additional comments (2)
prediction_market_agent/agents/microchain_agent/agents_nft_game.py (2)
1-6: LGTM! Imports are well-organized and appropriate.
The imports are clean, necessary, and well-structured with standard library imports separated from local imports.
19-70: Verify Ethereum addresses validity and uniqueness.
While the addresses are properly checksummed, let's verify:
- All addresses are valid Ethereum addresses
- No address reuse across agents
- Addresses match expected format
✅ Verification successful
Let me modify the script to correctly check only the addresses in the NFT game agents file and fix the path issue:
Ethereum addresses are valid and properly implemented
The verification confirms:
- All addresses are unique (no duplicates found)
- Each address follows the correct Ethereum format (0x + 40 hex characters)
- Total of 12 addresses (6 pairs of wallet/mech addresses for 6 agents) matches the expected count
- All addresses are properly checksummed
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify Ethereum addresses in the agents list
# Extract all addresses and verify format
echo "Verifying address format and uniqueness..."
rg -o '0x[a-fA-F0-9]{40}' | sort | uniq -c | while read count addr; do
if [ "$count" -gt 1 ]; then
echo "WARNING: Address $addr is used $count times"
fi
# Verify address length
if [ ${#addr} -ne 42 ]; then
echo "ERROR: Invalid address length for $addr"
fi
done
Length of output: 11596
Script:
#!/bin/bash
# Description: Verify Ethereum addresses specifically in agents_nft_game.py
echo "Verifying addresses in agents_nft_game.py..."
rg -o '0x[a-fA-F0-9]{40}' prediction_market_agent/agents/microchain_agent/agents_nft_game.py | sort | uniq -c | while read count addr; do
if [ "$count" -gt 1 ]; then
echo "WARNING: Address $addr is used $count times"
fi
done
# Count total unique addresses to ensure we have the right number
echo -e "\nTotal unique addresses:"
rg -o '0x[a-fA-F0-9]{40}' prediction_market_agent/agents/microchain_agent/agents_nft_game.py | sort -u | wc -l
Length of output: 839
prediction_market_agent/agents/microchain_agent/agents_nft_game.py
Outdated
Show resolved
Hide resolved
| st.markdown( | ||
| f"""### {nft_agent.name} | ||
|
|
||
| Currently holds <span style='font-size: 1.1em;'><strong>{xdai_balance:.2f} xDAI</strong></span>. | ||
|
|
||
| --- | ||
| """, |
There was a problem hiding this comment.
You can do this for bold on markdown
**bold** not bold
There was a problem hiding this comment.
Bold yes, but bigger font size, no.
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Outside diff range and nitpick comments (4)
tests/db/test_long_term_memory_table_handler.py (1)
Line range hint
1-48: Consider adding more test cases for better coverageWhile the current test cases cover basic functionality, consider adding:
- Edge cases (empty metadata, large metadata)
- Error scenarios (invalid metadata format)
- Boundary conditions for timestamp filtering
Example test cases to add:
def test_empty_metadata(memory_long_term_memory_handler): empty_item = {} memory_long_term_memory_handler.save_history([empty_item]) results = memory_long_term_memory_handler.search() assert results[0].metadata_dict == empty_item def test_timestamp_boundary_conditions(memory_long_term_memory_handler): item = {"test": "data"} timestamp = utcnow() memory_long_term_memory_handler.save_history([item]) # Test exact timestamp match results = memory_long_term_memory_handler.search(from_=timestamp, to_=timestamp) assert len(results) == 0 # Should be empty as item was saved after timestampprediction_market_agent/agents/microchain_agent/agents_nft_game.py (2)
14-16: Ensure URL safety in theurlpropertyThe current implementation may not handle special or non-ASCII characters in agent names, which could lead to invalid URLs. Consider using URL encoding to ensure the generated URLs are valid.
Apply this diff to enhance URL safety:
+ from urllib.parse import quote + @property def url(self) -> str: - return self.name.lower().replace(" ", "-") + return quote(self.name.lower().replace(" ", "-"))
19-70: Consider externalizing agent configurationsHardcoding agent data within the codebase reduces maintainability and flexibility. Loading agent configurations from an external file or environment variables can make updates easier and improve separation of data from code.
Here's how you might modify the code to load agents from a JSON file:
+ import json + AGENTS_IN_THE_GAME = [ - NFTAgent( - name="Sneaky Pete", - identifier=AgentIdentifier.NFT_GAME_AGENT_0, - wallet_address=Web3.to_checksum_address( - "0xd845A24014B3BD96212A21f602a4F16A7dA518A4" - ), - mech_address=Web3.to_checksum_address( - "0xDDe0780F744B84b505E344931F37cEDEaD8B6163" - ), - ), - # ... other agents ... - ] + NFTAgent(**agent_data) + for agent_data in json.load(open('agents_config.json')) + ]And create an
agents_config.jsonfile with the agent data:[ { "name": "Sneaky Pete", "identifier": "NFT_GAME_AGENT_0", "wallet_address": "0xd845A24014B3BD96212A21f602a4F16A7dA518A4", "mech_address": "0xDDe0780F744B84b505E344931F37cEDEaD8B6163" }, // ... other agents ... ]prediction_market_agent/agents/microchain_agent/app_nft_game.py (1)
108-108: Remove unnecessary f-stringThe string does not contain any placeholders; therefore, the
fprefix is unnecessary.Apply this diff to remove the extraneous
fprefix:- st.markdown(f"""### Agent's actions""") + st.markdown("""### Agent's actions""")🧰 Tools
🪛 Ruff (0.8.0)
108-108: f-string without any placeholders
Remove extraneous
fprefix(F541)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (4)
prediction_market_agent/agents/microchain_agent/agents_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/app_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/memory.py(2 hunks)tests/db/test_long_term_memory_table_handler.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.8.0)
prediction_market_agent/agents/microchain_agent/app_nft_game.py
108-108: f-string without any placeholders
Remove extraneous f prefix
(F541)
🔇 Additional comments (7)
tests/db/test_long_term_memory_table_handler.py (1)
41-41: LGTM! Improved assertion readability
The change to use metadata_dict instead of JSON parsing makes the test more maintainable and efficient.
Also applies to: 45-45
prediction_market_agent/agents/microchain_agent/memory.py (2)
37-37: LGTM!
Using metadata_dict with check_not_none improves data retrieval by directly accessing the dictionary, ensuring that metadata is present.
57-58: LGTM!
Updating to use model_validate with metadata_dict enhances metadata validation and aligns with the updated data handling approach.
prediction_market_agent/agents/microchain_agent/app_nft_game.py (4)
62-63: Improve function parsing robustness
The current string splitting approach may fail with unexpected message formats or nested parentheses. Consider using regular expressions to parse the function and arguments more reliably.
Apply this diff to enhance parsing:
def parse_function_and_body(
role: t.Literal["user", "assistant", "system"], message: str
) -> t.Tuple[str | None, str | None]:
message = message.strip()
- if role == "assistant":
- parsed_function = message.split("(")[0]
- parsed_body = message.split("(")[1].rsplit(")")[0]
+ import re
+ if role == "assistant":
+ match = re.match(r"(\w+)\((.*)\)$", message)
+ if match:
+ parsed_function, parsed_body = match.groups()
+ else:
+ parsed_function = "Unknown"
+ parsed_body = message
elif role == "user":
parsed_function = "Response"
parsed_body = message49-51:
Implement wallet integration and input validation
The code currently lacks input validation and uses manual wallet handling, as noted in the TODO comment. Integrating the user's wallet securely and validating inputs are essential to prevent potential security issues.
Would you like assistance in implementing secure wallet integration and input validation?
130-138: 🛠️ Refactor suggestion
Avoid using unsafe_allow_html=True to prevent XSS vulnerabilities
Using unsafe_allow_html=True can expose the application to XSS attacks. Consider using Markdown formatting to achieve the desired styling without enabling unsafe HTML.
You can modify the code as follows:
xdai_balance = get_balances(nft_agent.wallet_address).xdai
st.markdown(
- f"""### {nft_agent.name}
-
- Currently holds <span style='font-size: 1.1em;'><strong>{xdai_balance:.2f} xDAI</strong></span>.
-
- ---
- """,
- unsafe_allow_html=True,
+ f"""### {nft_agent.name}
+
+ Currently holds **{xdai_balance:.2f} xDAI**.
+
+ ---""",
)161-165: 🛠️ Refactor suggestion
Avoid using unsafe_allow_html=True to prevent XSS vulnerabilities
Using unsafe_allow_html=True in the sidebar can also pose XSS risks. Use Markdown formatting to display the balance safely.
Modify the code as follows:
st.sidebar.markdown(
- f"""### Treasury
- Currently holds <span style='font-size: 1.1em;'><strong>{treasury_xdai_balance:.2f} xDAI</strong></span>.""",
- unsafe_allow_html=True,
+ f"""### Treasury
+ Currently holds **{treasury_xdai_balance:.2f} xDAI**.""",
)
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Outside diff range and nitpick comments (3)
prediction_market_agent/agents/microchain_agent/app_nft_game.py (3)
1-5: Enhance docstring with setup prerequisites.The running instructions could be more comprehensive by including:
- Required environment variables
- Dependencies installation steps
- Expected behavior when running locally
Would you like me to help generate a more detailed docstring?
110-110: Remove unnecessary f-string.The string doesn't contain any placeholders, making the f-string unnecessary.
- st.markdown(f"""### Agent's actions""") + st.markdown("""### Agent's actions""")🧰 Tools
🪛 Ruff (0.8.0)
110-110: f-string without any placeholders
Remove extraneous
fprefix(F541)
108-169: Consider implementing rate limiting for auto-refresh.While the auto-refresh implementation using
@st.fragmentis good, frequent refreshes every 5 seconds might cause unnecessary load on the backend services.Consider:
- Increasing the refresh interval to 10-15 seconds
- Adding exponential backoff if no changes are detected
- Implementing a websocket-based approach for real-time updates
Would you like me to help implement any of these improvements?
🧰 Tools
🪛 Ruff (0.8.0)
110-110: f-string without any placeholders
Remove extraneous
fprefix(F541)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
⛔ Files ignored due to path filters (2)
poetry.lockis excluded by!**/*.lock,!**/*.lockpyproject.tomlis excluded by!**/*.toml
📒 Files selected for processing (3)
prediction_market_agent/agents/microchain_agent/app_nft_game.py(1 hunks)prediction_market_agent/agents/microchain_agent/memory.py(2 hunks)tests/db/test_long_term_memory_table_handler.py(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- tests/db/test_long_term_memory_table_handler.py
- prediction_market_agent/agents/microchain_agent/memory.py
🧰 Additional context used
🪛 Ruff (0.8.0)
prediction_market_agent/agents/microchain_agent/app_nft_game.py
110-110: f-string without any placeholders
Remove extraneous f prefix
(F541)
🔇 Additional comments (2)
prediction_market_agent/agents/microchain_agent/app_nft_game.py (2)
56-77:
Improve message parsing robustness.
The current string splitting approach is fragile and could break with complex messages containing nested parentheses.
Consider using a proper parser or regex as suggested in the previous review.
149-156:
Avoid using unsafe HTML in markdown.
Using unsafe_allow_html=True could expose XSS vulnerabilities.
Consider using Streamlit's native markdown formatting instead.
| @st.cache_resource | ||
| def long_term_memory_table_handler( | ||
| identifier: AgentIdentifier, | ||
| ) -> LongTermMemoryTableHandler: | ||
| return LongTermMemoryTableHandler.from_agent_identifier(identifier) | ||
|
|
||
|
|
||
| @st.cache_resource | ||
| def prompt_table_handler(identifier: AgentIdentifier) -> PromptTableHandler: | ||
| return PromptTableHandler.from_agent_identifier(identifier) | ||
|
|
There was a problem hiding this comment.
🛠️ Refactor suggestion
Add error handling for handler initialization.
The cached handlers could fail silently if initialization fails. Consider adding try-catch blocks and appropriate error messages.
@st.cache_resource
def long_term_memory_table_handler(
identifier: AgentIdentifier,
) -> LongTermMemoryTableHandler:
+ try:
return LongTermMemoryTableHandler.from_agent_identifier(identifier)
+ except Exception as e:
+ st.error(f"Failed to initialize memory handler: {str(e)}")
+ raise
@st.cache_resource
def prompt_table_handler(identifier: AgentIdentifier) -> PromptTableHandler:
+ try:
return PromptTableHandler.from_agent_identifier(identifier)
+ except Exception as e:
+ st.error(f"Failed to initialize prompt handler: {str(e)}")
+ raise📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| @st.cache_resource | |
| def long_term_memory_table_handler( | |
| identifier: AgentIdentifier, | |
| ) -> LongTermMemoryTableHandler: | |
| return LongTermMemoryTableHandler.from_agent_identifier(identifier) | |
| @st.cache_resource | |
| def prompt_table_handler(identifier: AgentIdentifier) -> PromptTableHandler: | |
| return PromptTableHandler.from_agent_identifier(identifier) | |
| @st.cache_resource | |
| def long_term_memory_table_handler( | |
| identifier: AgentIdentifier, | |
| ) -> LongTermMemoryTableHandler: | |
| try: | |
| return LongTermMemoryTableHandler.from_agent_identifier(identifier) | |
| except Exception as e: | |
| st.error(f"Failed to initialize memory handler: {str(e)}") | |
| raise | |
| @st.cache_resource | |
| def prompt_table_handler(identifier: AgentIdentifier) -> PromptTableHandler: | |
| try: | |
| return PromptTableHandler.from_agent_identifier(identifier) | |
| except Exception as e: | |
| st.error(f"Failed to initialize prompt handler: {str(e)}") | |
| raise |
build please