fix: support commas in custom exclude patterns

Author: RubenFricke

README.md

@@ -130,7 +130,7 @@ Use the cocoindex-code MCP server for semantic code search when:
 | `COCOINDEX_CODE_EMBEDDING_MODEL` | Embedding model (see below) | `sbert/sentence-transformers/all-MiniLM-L6-v2` |
 | `COCOINDEX_CODE_BATCH_SIZE` | Max batch size for local embedding model | `16` |
 | `COCOINDEX_CODE_EXTRA_EXTENSIONS` | Additional file extensions to index (comma-separated, e.g. `"inc:php,yaml,toml"` — use `ext:lang` to override language detection) | _(none)_ |
-| `COCOINDEX_CODE_EXCLUDED_PATTERNS` | Additional glob patterns to exclude from indexing (comma-separated, e.g. `"**/migration.sql,**/*.d.ts"`) | _(none)_ |
+| `COCOINDEX_CODE_EXCLUDED_PATTERNS` | Additional glob patterns to exclude from indexing as a JSON array (e.g. `'["**/migration.sql", "{**/*.md,**/*.txt}"]'`) | _(none)_ |
 
 
 ### Root Path Discovery

src/cocoindex_code/config.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import json
 import os
 from dataclasses import dataclass
 from pathlib import Path
@@ -42,6 +43,33 @@ def _discover_codebase_root() -> Path:
     return root if root is not None else cwd
 
 
+def _parse_json_string_list_env(var_name: str) -> list[str]:
+    """Parse an environment variable as a JSON array of strings."""
+    raw_value = os.environ.get(var_name, "")
+    if not raw_value.strip():
+        return []
+
+    try:
+        parsed = json.loads(raw_value)
+    except json.JSONDecodeError as exc:
+        raise ValueError(
+            f"{var_name} must be a JSON array of strings, got invalid JSON"
+        ) from exc
+
+    if not isinstance(parsed, list):
+        raise ValueError(f"{var_name} must be a JSON array of strings")
+
+    result: list[str] = []
+    for item in parsed:
+        if not isinstance(item, str):
+            raise ValueError(f"{var_name} must be a JSON array of strings")
+        item = item.strip()
+        if item:
+            result.append(item)
+
+    return result
+
+
 @dataclass
 class Config:
     """Configuration loaded from environment variables."""
@@ -101,13 +129,9 @@ def from_env(cls) -> Config:
                 extra_extensions[f".{token}"] = None
 
         # Excluded file glob patterns
-        raw_excluded_patterns = os.environ.get("COCOINDEX_CODE_EXCLUDED_PATTERNS", "")
-        excluded_patterns: list[str] = []
-        for pattern in raw_excluded_patterns.split(","):
-            pattern = pattern.strip()
-            if not pattern:
-                continue
-            excluded_patterns.append(pattern)
+        excluded_patterns = _parse_json_string_list_env(
+            "COCOINDEX_CODE_EXCLUDED_PATTERNS"
+        )
 
         return cls(
             codebase_root_path=root,

tests/test_config.py

@@ -6,6 +6,8 @@
 from pathlib import Path
 from unittest.mock import patch
 
+import pytest
+
 from cocoindex_code.config import Config
 
 
@@ -168,23 +170,34 @@ def test_empty_by_default(self, tmp_path: Path) -> None:
             config = Config.from_env()
             assert config.excluded_patterns == []
 
-    def test_parses_comma_separated(self, tmp_path: Path) -> None:
+    def test_parses_json_array(self, tmp_path: Path) -> None:
         with patch.dict(
             os.environ,
             {
                 "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
-                "COCOINDEX_CODE_EXCLUDED_PATTERNS": "**/migration.sql,**/*.d.ts",
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": '["**/migration.sql", "**/*.d.ts"]',
             },
         ):
             config = Config.from_env()
             assert config.excluded_patterns == ["**/migration.sql", "**/*.d.ts"]
 
-    def test_trims_whitespace(self, tmp_path: Path) -> None:
+    def test_preserves_commas_inside_globs(self, tmp_path: Path) -> None:
         with patch.dict(
             os.environ,
             {
                 "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
-                "COCOINDEX_CODE_EXCLUDED_PATTERNS": " **/migration.sql , **/*.d.ts , ",
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": '["{**/*.md,**/*.txt}"]',
+            },
+        ):
+            config = Config.from_env()
+            assert config.excluded_patterns == ["{**/*.md,**/*.txt}"]
+
+    def test_trims_whitespace_and_ignores_empty_entries(self, tmp_path: Path) -> None:
+        with patch.dict(
+            os.environ,
+            {
+                "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": '[" **/migration.sql ", " ", "**/*.d.ts"]',
             },
         ):
             config = Config.from_env()
@@ -200,3 +213,48 @@ def test_empty_string_gives_empty_list(self, tmp_path: Path) -> None:
         ):
             config = Config.from_env()
             assert config.excluded_patterns == []
+
+    def test_rejects_invalid_json(self, tmp_path: Path) -> None:
+        with patch.dict(
+            os.environ,
+            {
+                "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": "**/migration.sql,**/*.d.ts",
+            },
+        ):
+            with pytest.raises(
+                ValueError,
+                match=(
+                    "COCOINDEX_CODE_EXCLUDED_PATTERNS must be a JSON array of strings, "
+                    "got invalid JSON"
+                ),
+            ):
+                Config.from_env()
+
+    def test_rejects_valid_json_non_list(self, tmp_path: Path) -> None:
+        with patch.dict(
+            os.environ,
+            {
+                "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": "{}",
+            },
+        ):
+            with pytest.raises(
+                ValueError,
+                match="COCOINDEX_CODE_EXCLUDED_PATTERNS must be a JSON array of strings",
+            ):
+                Config.from_env()
+
+    def test_rejects_non_string_entries(self, tmp_path: Path) -> None:
+        with patch.dict(
+            os.environ,
+            {
+                "COCOINDEX_CODE_ROOT_PATH": str(tmp_path),
+                "COCOINDEX_CODE_EXCLUDED_PATTERNS": '["**/*.py", 1]',
+            },
+        ):
+            with pytest.raises(
+                ValueError,
+                match="COCOINDEX_CODE_EXCLUDED_PATTERNS must be a JSON array of strings",
+            ):
+                Config.from_env()