feat: allowing for tools message in full prompt

[themaherkhalil]

No description provided.

[github-actions]

@CodiumAI-Agent /describe

[QodoAI-Agent]

Title

feat: allowing for tools message in full prompt

---

PR Type

Enhancement

---

Description

• Add TODO comments for single-tool handling

• Refactor full_prompt validation in Python client

• Support tool_calls and tool_call_id keys

• Rename variables to snake_case

---

Changes walkthrough 📝

	Relevant files
Documentation	AskModelEngineResponse.java Add TODOs for single-tool handling                                              src/prerna/engine/impl/model/responses/AskModelEngineResponse.java Inserted multiple TODOs about single-tool limitation Highlighted need to handle more than one tool +6/-0
Enhancement	chat.py Enhance full_prompt validation for tools                                  py/genai_client/text_generation/openai_clients/operations/chat.py Renamed variables to snake_case Extended valid key sets for full_prompt dicts Improved validation logic for OpenAI payload Removed extraneous whitespace +14/-6

---

Need help?

Type /help how to ... in the comments thread for any questions about PR-Agent usage.

Check out the documentation for more information.

[github-actions]

@CodiumAI-Agent /review

[QodoAI-Agent]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review TODO Duplication The added TODO comments are duplicated multiple times; consolidate or clarify them to avoid clutter. //TODO: why are we grabbing only 1 tool??? //TODO: why are we grabbing only 1 tool??? //TODO: why are we grabbing only 1 tool??? //TODO: why are we grabbing only 1 tool??? //TODO: why are we grabbing only 1 tool??? //TODO: why are we grabbing only 1 tool??? Single Tool Limitation The code only processes the first element of the tool response list; consider handling multiple tools or document this limitation. List<?> responseList = (List<?>) response; // Handle one tool object to process if (!responseList.isEmpty() && responseList.get(0) instanceof Map) { return new AskToolModelEngineResponse((Map<String, Object>) responseList.get(0), tokensInPrompt, tokensInResponse); Full Prompt Validation The validation logic uses a set of booleans and complex key‐set checks; simplify using all(isinstance(...)) and ensure key requirements cover all intended payloads. list_of_dicts = set([isinstance(x, dict) for x in full_prompt]) == {True} if not list_of_dicts: raise ValueError("The provided payload is not valid") # now we have to check the key value pairs are valid valid_key_sets = [ {"content", "role"}, {"role", "tool_calls"}, {"role", "tool_call_id"}, ] all_keys_valid = all( # if any in valid_keys are a subset of the full prompt keys any(valid_keys <= d.keys() for valid_keys in valid_key_sets) for d in full_prompt )

[github-actions]

@CodiumAI-Agent /improve

[QodoAI-Agent]

PR Code Suggestions ✨

Category	Suggestion	Impact
Possible issue	Enforce exact key matching Use exact key-set matching to prevent dictionaries with extra unexpected keys from being accepted. Change the subset check to an equality check against the allowed key sets. py/genai_client/text_generation/openai_clients/operations/chat.py [160-169] valid_key_sets = [ {"content", "role"}, {"role", "tool_calls"}, {"role", "tool_call_id"}, ] -all_keys_valid = all( - any(valid_keys <= d.keys() for valid_keys in valid_key_sets) - for d in full_prompt -) +all_keys_valid = all(set(d.keys()) in valid_key_sets for d in full_prompt) Suggestion importance[1-10]: 7 __ Why: Changing the subset check to set(d.keys()) in valid_key_sets tightens validation by rejecting dictionaries with unexpected keys, improving input correctness.	Medium
General	Remove redundant TODOs Consolidate or remove duplicated TODO comments to keep the code clean and maintainable. Keep a single comment that explains the intended behavior or tasks to address. src/prerna/engine/impl/model/responses/AskModelEngineResponse.java [78-83] -//TODO: why are we grabbing only 1 tool??? -//TODO: why are we grabbing only 1 tool??? -//TODO: why are we grabbing only 1 tool??? -//TODO: why are we grabbing only 1 tool??? -//TODO: why are we grabbing only 1 tool??? -//TODO: why are we grabbing only 1 tool??? +// TODO: Review tool response handling and support multiple entries Suggestion importance[1-10]: 3 __ Why: Consolidating six identical //TODO comments into one cleans up the code, but it is purely stylistic with minimal functional impact.	Low

[github-actions]

@CodiumAI-Agent /update_changelog