src/deriver/consumer.py (2)

39-60: Consider handling empty chat histories
If the session has no prior messages, this function returns an empty string. Ensure downstream steps (like inference) won't break or degrade when there's no history available.

---

95-193: Fix spelling in docstring and validate XML parse output

1. The docstring at line 106 has a minor typo ("Proces" → "Process").
2. If parse_xml_content encounters malformed or missing tags, the returned strings could be empty, silently affecting downstream logic. Consider adding checks or logs to handle or surface these cases.

src/deriver/tom/single_prompt.py (1)

1-188: Validate custom model name and consider externalizing prompts

1. The hard-coded model "claude-3-5-haiku-20241022" may be temporary or environment-specific. Confirm it's correct and stable for your deployment.
2. The large system prompts could be extracted to a separate file or template for easier maintenance and updates, reduce inline clutter, and improve readability.

src/deriver/tom/__init__.py (1)

18-30: Maintain clarity in logs
This function closely mirrors get_tom_inference; consider logging the chosen method (method argument) inside the function for improved debugging and tracing.

src/main.py (1)

96-99: Add version prefix and response model to the health check endpoint.

The health check endpoint should follow the same versioning pattern as other endpoints and have a documented response model.

Apply this diff to improve the endpoint:

-@app.get("/health")
-async def healthcheck():
-    return "OK"
+@app.get("/v1/health", response_model=str, description="Health check endpoint")
+async def healthcheck() -> str:
+    return "OK"

.vscode/tasks.json (1)

7-7: Verify the uv command and consider error handling.

The command has been updated to use uv instead of Poetry. While this aligns with the documentation changes, consider adding error handling.

Apply this diff to improve error handling:

-      "command": "uv sync && uv run fastapi dev src/main.py",
+      "command": "uv sync || exit 1 && uv run fastapi dev src/main.py",

Please verify:

1. Is uv installed in the development environment?
2. Have you tested the command on different platforms (Windows, Linux, macOS)?

.env.template (1)

14-14: Introduction of SECRET_KEY without a default value.
A new configuration SECRET_KEY= has been added without a provided value. Ensure that production setups securely assign a value and that documentation highlights its importance.

docs/README.md (1)

12-15: Fenced code block language specification recommended.
For improved markdown lint compliance, consider adding a language identifier (e.g., bash) to the fenced code block starting at line 12 (and similarly for other code blocks).

Proposed diff:

-```
+```bash

🧰 Tools

🪛 markdownlint-cli2 (0.17.2)

13-13: Fenced code blocks should have a language specified
null

(MD040, fenced-code-language)

docs/getting-started/introduction.mdx (1)

10-17: Enhanced API capabilities narrative.
The expanded explanation covering the Storage and Insights APIs now provides detailed context. Consider revisiting phrases like "social cognition and deeper understanding" for stylistic refinement if needed.

🧰 Tools

🪛 LanguageTool

[style] ~10-~10: Consider an alternative adjective to strengthen your wording.
Context: ...ers AI agents with social cognition and deeper understanding of their users. The API ...

(DEEP_PROFOUND)

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

src/deriver/consumer.py (4)

23-23: Address the FIXME comment.

The comment indicates potential safety concerns that need to be addressed.

Would you like me to help analyze the safety implications of the add_metamessage function and propose improvements?

---

40-60: Document the message limit and consider pagination.

The function retrieves the last 10 messages, but this limit isn't documented. Consider:

1. Adding a docstring explaining the limit
2. Making the limit configurable
3. Adding pagination for larger history retrieval

Here's a suggested improvement:

 async def get_chat_history(db, session_id, message_id) -> str:
+    """Retrieve chat history for a session.
+    
+    Args:
+        db: Database session
+        session_id: Session identifier
+        message_id: Current message ID
+    
+    Returns:
+        str: Last 10 messages formatted as 'human: message' or 'ai: message'
+    """
+    HISTORY_LIMIT = 10  # Consider making this configurable
     subquery = (
         select(models.Message.id)
         .where(models.Message.public_id == message_id)
         .scalar_subquery()
     )
     messages_stmt = (
         select(models.Message)
         .where(models.Message.session_id == session_id)
         .order_by(models.Message.id.desc())
         .where(models.Message.id < subquery)
-        .limit(10)
+        .limit(HISTORY_LIMIT)
     )

---

34-37: Enhance XML parsing robustness.

The current XML parsing implementation could be more robust:

1. It doesn't handle malformed XML
2. It uses regex for parsing XML which can be unreliable

Consider using a proper XML parser:

+from xml.etree import ElementTree
+from io import StringIO
+
 def parse_xml_content(text, tag):
+    """Parse XML content safely.
+    
+    Args:
+        text: XML text to parse
+        tag: Tag name to extract
+    
+    Returns:
+        str: Content within the tag or empty string if not found
+    """
+    try:
+        # Wrap content in root element to handle multiple tags
+        wrapped = f"<root>{text}</root>"
+        tree = ElementTree.parse(StringIO(wrapped))
+        element = tree.find(tag)
+        return element.text.strip() if element is not None else ""
+    except (ElementTree.ParseError, AttributeError):
+        return ""
-    pattern = f"<{tag}>(.*?)</{tag}>"
-    match = re.search(pattern, text, re.DOTALL)
-    return match.group(1).strip() if match else ""

---

130-147: Optimize user representation query performance.

The complex join query for fetching user representation could benefit from an index to improve performance.

Consider adding an index on the following columns:

1. metamessage_type in the metamessage table
2. public_id in the user, session, and app tables

Example SQL for adding indexes:

CREATE INDEX idx_metamessage_type ON metamessage(metamessage_type);
CREATE INDEX idx_user_public_id ON user(public_id);
CREATE INDEX idx_session_public_id ON session(public_id);
CREATE INDEX idx_app_public_id ON app(public_id);

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

CHANGELOG.md (2)

8-9: Suggestion: Add Release Date for Consistency

The new version header ## [0.0.16] is correctly added, but notice that previous versions (e.g., ## [0.0.14] — 2024-11-14) include a release date. For consistency and better historical tracking, consider appending the release date here as well.

-## [0.0.16]
+## [0.0.16] — YYYY-MM-DD

---

10-16: Review: Clarify 'Changed' Section Descriptions

The "Changed" section now lists:

• A new cognitive architecture for the Deriver that only updates on user messages and enhances confidence scores for known facts.
• An increase of the Dialectic API token cutoff from 150 to 300 tokens.

While these bullet points capture the changes, you might consider expanding the descriptions or adding contextual details (such as the rationale or expected impact) to help readers quickly understand the significance of these changes.

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro