ocp_pipeline/opm.py (5)

52-52: Use os.path.join for constructing file paths

To ensure cross-platform compatibility and handle path separators correctly, it's recommended to use os.path.join instead of string concatenation when constructing file paths.

Apply this diff to improve the code:

-    intent_cache = f"{xdg_data_home()}/{get_xdg_base()}/intent_cache"
+    intent_cache = os.path.join(xdg_data_home(), get_xdg_base(), "intent_cache")

---

320-321: Simplify empty dictionary check

Instead of using not len(self.skill_aliases), you can directly use if not self.skill_aliases to check if the dictionary is empty. This is more Pythonic and slightly more efficient.

Apply this diff to refine the code:

-            if not len(self.skill_aliases):  # skill_id registered when skills load
+            if not self.skill_aliases:  # skill_id registered when skills load

---

332-338: Refactor duplicate match normalization code

The logic for normalizing the match object is duplicated. Consider refactoring this into a separate method to improve maintainability and reduce code duplication.

Apply this diff to extract the normalization logic:

+    def _normalize_match(self, match):
+        if hasattr(match, "name"):  # padatious
+            return {
+                "name": match.name,
+                "conf": match.conf,
+                "entities": match.matches
+            }
+        return match

Then update the code to use this method:

-            if hasattr(match, "name"):  # padatious
-                match = {
-                    "name": match.name,
-                    "conf": match.conf,
-                    "entities": match.matches
-                }
+            match = self._normalize_match(match)

---

342-344: Make confidence threshold configurable

Instead of hardcoding the confidence threshold to 0.7, consider making it a configurable parameter. This allows flexibility to adjust the threshold based on different requirements or environments.

Apply this diff to enhance configurability:

-            if match.get("conf", 1.0) < 0.7:
+            threshold = self.config.get("intent_match_confidence_threshold", 0.7)
+            if match.get("conf", 1.0) < threshold:

---

1162-1167: Refactor duplicate match normalization code

The logic for normalizing the match object is duplicated here as well. Refactoring this into a shared method improves code maintainability.

Use the previously defined _normalize_match method:

-            if hasattr(match, "name"):  # padatious
-                match = {
-                    "name": match.name,
-                    "conf": match.conf,
-                    "entities": match.matches
-                }
+            match = self._normalize_match(match)

Configuration used: CodeRabbit UI
Review profile: CHILL

tests/test_ocp.py (1)

Line range hint 164-182: Consider un-skipping and fixing the classifier test

The skipped test test_predict_prob_with_unknown_entity contains a TODO comment about classifier retraining. Since this PR involves changes to the matching system, it would be valuable to:

1. Retrain the classifiers to fix the noted issue
2. Un-skip this test to ensure robust validation of the matching system

Would you like me to help create a GitHub issue to track the classifier retraining task?

Configuration used: CodeRabbit UI
Review profile: CHILL