refactor: simplify code structure and reduce complexity

src/embeddings/backend.py

@@ -21,17 +21,18 @@ def _validate_api(url: str, api_key: str, payload: dict) -> bool:
     """Validate an embedding API by making a test request."""
     if not api_key:
         return False
+
     try:
         response = httpx.post(
             url,
             headers={"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"},
             json=payload,
             timeout=10.0,
         )
-        if response.status_code != 200:
-            logger.warning(f"API at {url} returned {response.status_code}")
-            return False
-        return True
+        if response.status_code == 200:
+            return True
+        logger.warning(f"API at {url} returned {response.status_code}")
+        return False
     except Exception as e:
         logger.warning(f"API validation failed for {url}: {e}")
         return False

src/indexer/service.py

@@ -70,27 +70,28 @@ def resolve_repo_name(self, path: str | Path) -> str:
         base = self.path_to_repo_name(resolved_str)
         digest = hashlib.sha1(resolved_str.encode("utf-8")).hexdigest()
 
-        def get_repo_safe(name: str):
-            try:
-                return self._repo_manager.get_repo(name)
-            except Exception:
-                return None
-
         # Check existing registrations
         for name in [base] + [f"{base}_{digest[:n]}" for n in (8, 12, 16, 40)]:
-            existing = get_repo_safe(name)
+            existing = self._get_repo_safe(name)
             if existing and existing.path == resolved_str:
                 return name
 
         # Find available name
-        if get_repo_safe(base) is None:
+        if self._get_repo_safe(base) is None:
             return base
         for n in (8, 12, 16, 40):
             hashed = f"{base}_{digest[:n]}"
-            if get_repo_safe(hashed) is None:
+            if self._get_repo_safe(hashed) is None:
                 return hashed
         return f"{base}_{digest}"
 
+    def _get_repo_safe(self, name: str):
+        """Safely get repo, returning None on error."""
+        try:
+            return self._repo_manager.get_repo(name)
+        except Exception:
+            return None
+
     def _validate_path(self, path: str) -> Path:
         resolved = Path(path).resolve()
         if not resolved.exists():
@@ -145,24 +146,25 @@ def _has_files_changed(self, repo_name: str, repo_path: str) -> bool:
         last_ts = self._datetime_to_timestamp(repo.last_indexed)
         now = time.monotonic()
 
-        # Cache for 1 second to avoid redundant checks while keeping changes responsive
+        # Check cache (1 second TTL)
         cached = self._change_check_cache.get(repo_name)
         if cached and cached[0] == last_ts and (now - cached[1]) < 1.0:
             return cached[2]
 
-        changed = False
+        changed = self._check_files_modified(repo_path, last_ts)
+        self._change_check_cache[repo_name] = (last_ts, now, changed)
+        return changed
+
+    def _check_files_modified(self, repo_path: str, last_ts: float) -> bool:
+        """Check if any relevant files have been modified since last_ts."""
         for full_path, _ in self._iter_relevant_files(repo_path):
             try:
                 if full_path.stat().st_mtime > last_ts:
-                    changed = True
-                    break
+                    return True
             except OSError:
                 # File access error - assume changed to trigger re-index
-                changed = True
-                break
-
-        self._change_check_cache[repo_name] = (last_ts, now, changed)
-        return changed
+                return True
+        return False
 
     @staticmethod
     def _datetime_to_timestamp(value: datetime) -> float:

src/retrieval/bm25_search.py

@@ -85,8 +85,10 @@ def _has_bm25_index(table_name: str) -> bool:
 
 def _search_pg_search(table_name: str, tokens: list[str], top_k: int) -> list[SearchResult]:
     """Search using ParadeDB pg_search extension."""
-    if not tokens or not _has_bm25_index(table_name):
-        return _search_native_fts(table_name, tokens, top_k) if tokens else []
+    if not tokens:
+        return []
+    if not _has_bm25_index(table_name):
+        return _search_native_fts(table_name, tokens, top_k)
 
     with get_connection() as conn:
         with conn.cursor() as cur:
@@ -102,8 +104,10 @@ def _search_pg_search(table_name: str, tokens: list[str], top_k: int) -> list[Se
 
 def _search_pg_textsearch(table_name: str, tokens: list[str], top_k: int) -> list[SearchResult]:
     """Search using pg_textsearch extension."""
-    if not tokens or not _has_bm25_index(table_name):
-        return _search_native_fts(table_name, tokens, top_k) if tokens else []
+    if not tokens:
+        return []
+    if not _has_bm25_index(table_name):
+        return _search_native_fts(table_name, tokens, top_k)
 
     search_query = " ".join(tokens)
     with get_connection() as conn:

src/retrieval/hybrid.py

@@ -140,14 +140,19 @@ def _run_searches_parallel(
     with ThreadPoolExecutor(max_workers=len(searches)) as executor:
         futures = {name: executor.submit(fn) for name, fn in searches}
         for name, future in futures.items():
-            try:
-                outcomes[name] = SearchOutcome(results=future.result(timeout=30))
-            except Exception as e:
-                logger.warning(f"{name} search failed: {e}")
-                outcomes[name] = SearchOutcome(failed=True)
+            outcomes[name] = _await_search_result(name, future)
     return outcomes
 
 
+def _await_search_result(name: str, future) -> SearchOutcome:
+    """Wait for a search future and return its outcome."""
+    try:
+        return SearchOutcome(results=future.result(timeout=30))
+    except Exception as e:
+        logger.warning(f"{name} search failed: {e}")
+        return SearchOutcome(failed=True)
+
+
 def hybrid_search(
     repo_name: str,
     query: str,

src/retrieval/service.py

@@ -261,23 +261,38 @@ def _mmr_diversify(
         category_counts: dict[str, int] = defaultdict(int)
 
         while len(selected) < k and remaining:
-            best_file, best_score = None, float('-inf')
-
-            for candidate in remaining:
-                divisor = max(len(selected), 1)
-                penalty = category_counts[categories[candidate]] / divisor
-                mmr = lambda_param * scores[candidate] - (1 - lambda_param) * penalty
-
-                if mmr > best_score:
-                    best_score, best_file = mmr, candidate
-
+            best_file = self._select_best_candidate(
+                remaining, selected, scores, categories, category_counts, lambda_param
+            )
             if best_file:
                 selected.append(best_file)
                 remaining.remove(best_file)
                 category_counts[categories[best_file]] += 1
 
         return selected
 
+    def _select_best_candidate(
+        self,
+        candidates: list[str],
+        selected: list[str],
+        scores: dict[str, float],
+        categories: dict[str, str],
+        category_counts: dict[str, int],
+        lambda_param: float,
+    ) -> str | None:
+        """Select the best candidate using MMR scoring."""
+        best_file, best_score = None, float('-inf')
+        divisor = max(len(selected), 1)
+
+        for candidate in candidates:
+            penalty = category_counts[categories[candidate]] / divisor
+            mmr = lambda_param * scores[candidate] - (1 - lambda_param) * penalty
+
+            if mmr > best_score:
+                best_score, best_file = mmr, candidate
+
+        return best_file
+
     def _build_full_snippet(
         self, filename: str, chunks: list, score: float, repo_path: str | None,
         file_cache: dict[str, bytes], newline_cache: dict[str, list[int]],