cocoindex-code 0.2.10__tar.gz → 0.2.22__tar.gz

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cocoindex-code
-Version: 0.2.10
+Version: 0.2.22
 Summary: MCP server for indexing and querying codebases using CocoIndex
 Project-URL: Homepage, https://github.com/cocoindex-io/cocoindex-code
 Project-URL: Repository, https://github.com/cocoindex-io/cocoindex-code
@@ -17,7 +17,7 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.11
-Requires-Dist: cocoindex[litellm]==1.0.0a38
+Requires-Dist: cocoindex[litellm]==1.0.0a43
 Requires-Dist: einops>=0.8.2
 Requires-Dist: mcp>=1.0.0
 Requires-Dist: msgspec>=0.19.0
@@ -345,6 +345,7 @@ embedding:
   provider: sentence-transformers                    # or "litellm"
   model: sentence-transformers/all-MiniLM-L6-v2
   device: mps                                        # optional: cpu, cuda, mps (auto-detected if omitted)
+  min_interval_ms: 300                               # optional: pace LiteLLM embedding requests to reduce 429s; defaults to 5 for LiteLLM
 
 envs:                                                # extra environment variables for the daemon
   OPENAI_API_KEY: your-key                           # only needed if not already in your shell environment
@@ -436,6 +437,7 @@ Set `OLLAMA_API_BASE` in `envs:` if your Ollama server is not at `http://localho
 ```yaml
 embedding:
   model: text-embedding-3-small
+  min_interval_ms: 300                               # optional: override the 5ms LiteLLM default
 envs:
   OPENAI_API_KEY: your-api-key
 ```

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/README.md

@@ -306,6 +306,7 @@ embedding:
   provider: sentence-transformers                    # or "litellm"
   model: sentence-transformers/all-MiniLM-L6-v2
   device: mps                                        # optional: cpu, cuda, mps (auto-detected if omitted)
+  min_interval_ms: 300                               # optional: pace LiteLLM embedding requests to reduce 429s; defaults to 5 for LiteLLM
 
 envs:                                                # extra environment variables for the daemon
   OPENAI_API_KEY: your-key                           # only needed if not already in your shell environment
@@ -397,6 +398,7 @@ Set `OLLAMA_API_BASE` in `envs:` if your Ollama server is not at `http://localho
 ```yaml
 embedding:
   model: text-embedding-3-small
+  min_interval_ms: 300                               # optional: override the 5ms LiteLLM default
 envs:
   OPENAI_API_KEY: your-api-key
 ```

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
 
 dependencies = [
     "mcp>=1.0.0",
-    "cocoindex[litellm]==1.0.0a38",
+    "cocoindex[litellm]==1.0.0a43",
     "sentence-transformers>=2.2.0",
     "sqlite-vec>=0.1.0",
     "pydantic>=2.0.0",

cocoindex_code-0.2.22/src/cocoindex_code/_daemon_paths.py

@@ -0,0 +1,44 @@
+"""Daemon filesystem paths and connection helpers.
+
+Lightweight module with no cocoindex dependency so that the CLI client
+can import these without pulling in the full daemon stack.
+"""
+
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+from .settings import user_settings_dir
+
+
+def daemon_dir() -> Path:
+    """Return the daemon directory (``~/.cocoindex_code/``)."""
+    return user_settings_dir()
+
+
+def connection_family() -> str:
+    """Return the multiprocessing connection family for this platform."""
+    return "AF_PIPE" if sys.platform == "win32" else "AF_UNIX"
+
+
+def daemon_socket_path() -> str:
+    """Return the daemon socket/pipe address."""
+    if sys.platform == "win32":
+        import hashlib
+
+        # Hash the daemon dir so COCOINDEX_CODE_DIR overrides create unique pipe names,
+        # preventing conflicts between different daemon instances (tests, users, etc.)
+        dir_hash = hashlib.md5(str(daemon_dir()).encode()).hexdigest()[:12]
+        return rf"\\.\pipe\cocoindex_code_{dir_hash}"
+    return str(daemon_dir() / "daemon.sock")
+
+
+def daemon_pid_path() -> Path:
+    """Return the path for the daemon's PID file."""
+    return daemon_dir() / "daemon.pid"
+
+
+def daemon_log_path() -> Path:
+    """Return the path for the daemon's log file."""
+    return daemon_dir() / "daemon.log"

cocoindex_code-0.2.22/src/cocoindex_code/_version.py

@@ -0,0 +1,24 @@
+# file generated by vcs-versioning
+# don't change, don't track in version control
+from __future__ import annotations
+
+__all__ = [
+    "__version__",
+    "__version_tuple__",
+    "version",
+    "version_tuple",
+    "__commit_id__",
+    "commit_id",
+]
+
+version: str
+__version__: str
+__version_tuple__: tuple[int | str, ...]
+version_tuple: tuple[int | str, ...]
+commit_id: str | None
+__commit_id__: str | None
+
+__version__ = version = '0.2.22'
+__version_tuple__ = version_tuple = (0, 2, 22)
+
+__commit_id__ = commit_id = None

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/cli.py

@@ -599,7 +599,7 @@ def doctor() -> None:
 
     # --- 8. Log files ---
     _print_section("Log Files")
-    from .daemon import daemon_log_path as _daemon_log_path
+    from ._daemon_paths import daemon_log_path as _daemon_log_path
 
     _typer.echo(f"  Daemon logs: {_daemon_log_path()}")
     _typer.echo("  Check logs above for further troubleshooting.")
@@ -675,8 +675,8 @@ def daemon_restart() -> None:
 @daemon_app.command("stop")
 def daemon_stop() -> None:
     """Stop the daemon."""
+    from ._daemon_paths import daemon_pid_path
     from .client import is_daemon_running, stop_daemon
-    from .daemon import daemon_pid_path
 
     pid_path = daemon_pid_path()
     if not pid_path.exists() and not is_daemon_running():

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/client.py

@@ -17,8 +17,14 @@ from collections.abc import Callable
 from multiprocessing.connection import Client, Connection
 from pathlib import Path
 
+from ._daemon_paths import (
+    connection_family,
+    daemon_dir,
+    daemon_log_path,
+    daemon_pid_path,
+    daemon_socket_path,
+)
 from ._version import __version__
-from .daemon import _connection_family, daemon_log_path, daemon_pid_path, daemon_socket_path
 from .protocol import (
     DaemonEnvRequest,
     DaemonEnvResponse,
@@ -105,7 +111,7 @@ def _raw_connect_and_handshake() -> Connection:
     if sys.platform != "win32" and not os.path.exists(sock):
         raise ConnectionRefusedError(f"Daemon socket not found: {sock}")
     try:
-        conn = Client(sock, family=_connection_family())
+        conn = Client(sock, family=connection_family())
     except (ConnectionRefusedError, FileNotFoundError, OSError) as e:
         raise ConnectionRefusedError(f"Cannot connect to daemon: {e}") from e
 
@@ -329,7 +335,7 @@ def is_daemon_running() -> bool:
     """Check if the daemon is running."""
     if sys.platform == "win32":
         try:
-            conn = Client(daemon_socket_path(), family=_connection_family())
+            conn = Client(daemon_socket_path(), family=connection_family())
             conn.close()
             return True
         except (ConnectionRefusedError, OSError):
@@ -343,8 +349,6 @@ def start_daemon() -> subprocess.Popen[bytes]:
     Returns the ``Popen`` object so callers can detect early process death
     (via ``proc.poll()``) instead of waiting for a full timeout.
     """
-    from .daemon import daemon_dir, daemon_log_path
-
     daemon_dir().mkdir(parents=True, exist_ok=True)
     log_path = daemon_log_path()
 
@@ -518,7 +522,7 @@ def _wait_for_daemon(
 
         if sys.platform == "win32":
             try:
-                conn = Client(sock_path, family=_connection_family())
+                conn = Client(sock_path, family=connection_family())
                 conn.close()
                 return
             except (ConnectionRefusedError, OSError):

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/daemon.py

@@ -15,6 +15,13 @@ from multiprocessing.connection import Connection, Listener
 from pathlib import Path
 from typing import Any
 
+from ._daemon_paths import (
+    connection_family,
+    daemon_dir,
+    daemon_log_path,
+    daemon_pid_path,
+    daemon_socket_path,
+)
 from ._version import __version__
 from .chunking import ChunkerFn as _ChunkerFn
 from .project import Project
@@ -53,7 +60,6 @@ from .settings import (
     load_project_settings,
     load_user_settings,
     target_sqlite_db_path,
-    user_settings_dir,
 )
 from .shared import Embedder, create_embedder
 
@@ -79,43 +85,6 @@ def _resolve_chunker_registry(mappings: list[ChunkerMapping]) -> dict[str, _Chun
     return registry
 
 
-# ---------------------------------------------------------------------------
-# Daemon paths
-# ---------------------------------------------------------------------------
-
-
-def daemon_dir() -> Path:
-    """Return the daemon directory (``~/.cocoindex_code/``)."""
-    return user_settings_dir()
-
-
-def _connection_family() -> str:
-    """Return the multiprocessing connection family for this platform."""
-    return "AF_PIPE" if sys.platform == "win32" else "AF_UNIX"
-
-
-def daemon_socket_path() -> str:
-    """Return the daemon socket/pipe address."""
-    if sys.platform == "win32":
-        import hashlib
-
-        # Hash the daemon dir so COCOINDEX_CODE_DIR overrides create unique pipe names,
-        # preventing conflicts between different daemon instances (tests, users, etc.)
-        dir_hash = hashlib.md5(str(daemon_dir()).encode()).hexdigest()[:12]
-        return rf"\\.\pipe\cocoindex_code_{dir_hash}"
-    return str(daemon_dir() / "daemon.sock")
-
-
-def daemon_pid_path() -> Path:
-    """Return the path for the daemon's PID file."""
-    return daemon_dir() / "daemon.pid"
-
-
-def daemon_log_path() -> Path:
-    """Return the path for the daemon's log file."""
-    return daemon_dir() / "daemon.log"
-
-
 # ---------------------------------------------------------------------------
 # Project Registry
 # ---------------------------------------------------------------------------
@@ -540,7 +509,7 @@ def run_daemon() -> None:
         except Exception:
             pass
 
-    listener = Listener(sock_path, family=_connection_family())
+    listener = Listener(sock_path, family=connection_family())
     logger.info("Listening on %s", sock_path)
 
     loop = asyncio.new_event_loop()

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/indexer.py

@@ -224,5 +224,6 @@ async def indexer_main() -> None:
         path_matcher=matcher,
     )
 
-    with coco.component_subpath(coco.Symbol("process_file")):
-        await coco.mount_each(process_file, files.items(), table)
+    await coco.mount_each(
+        coco.component_subpath(coco.Symbol("process_file")), process_file, files.items(), table
+    )

cocoindex_code-0.2.22/src/cocoindex_code/litellm_embedder.py

@@ -0,0 +1,126 @@
+"""LiteLLM embedder with optional request pacing and rate-limit retries."""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+import re
+import time
+from typing import Any
+
+import cocoindex as coco
+import numpy as np
+from cocoindex.ops.litellm import LiteLLMEmbedder, litellm
+from numpy.typing import NDArray
+
+logger = logging.getLogger(__name__)
+
+_RATE_LIMIT_DELAY_RE = re.compile(r"Please try again in ([0-9.]+)(ms|s)", re.IGNORECASE)
+_MAX_RATE_LIMIT_RETRIES = 6
+
+
+def _get_rate_limit_delay(exc: Exception, attempt: int) -> float | None:
+    message = str(exc)
+    if "rate limit" not in message.lower():
+        return None
+
+    match = _RATE_LIMIT_DELAY_RE.search(message)
+    if match is not None:
+        value = float(match.group(1))
+        unit = match.group(2).lower()
+        delay = value / 1000.0 if unit == "ms" else value
+    else:
+        delay = min(0.5 * (2**attempt), 10.0)
+
+    return min(delay + 0.1, 10.0)
+
+
+class PacedLiteLLMEmbedder(LiteLLMEmbedder):
+    """LiteLLM embedder that serializes requests and paces them when configured."""
+
+    def __init__(self, model: str, *, min_interval_ms: int | None = None, **kwargs: Any) -> None:
+        super().__init__(model, **kwargs)
+        self._min_request_interval_seconds = max(0.0, float(min_interval_ms or 0) / 1000.0)
+        self._request_lock: asyncio.Lock | None = None
+        self._next_request_at: float = 0.0
+
+    def _get_request_lock(self) -> asyncio.Lock:
+        if self._request_lock is None:
+            self._request_lock = asyncio.Lock()
+        return self._request_lock
+
+    async def _aembedding_with_rate_limit_retries(
+        self, *, model: str, input: list[str], **kwargs: Any
+    ) -> Any:
+        last_exc: Exception | None = None
+
+        for attempt in range(_MAX_RATE_LIMIT_RETRIES):
+            try:
+                return await litellm.aembedding(model=model, input=input, **kwargs)
+            except Exception as exc:  # noqa: BLE001
+                delay = _get_rate_limit_delay(exc, attempt)
+                last_exc = exc
+                if delay is None or attempt == _MAX_RATE_LIMIT_RETRIES - 1:
+                    raise
+
+                logger.warning(
+                    "Embedding rate limited for model %s, retrying in %.3fs (attempt %d/%d)",
+                    model,
+                    delay,
+                    attempt + 1,
+                    _MAX_RATE_LIMIT_RETRIES,
+                )
+                await asyncio.sleep(delay)
+
+        assert last_exc is not None
+        raise last_exc
+
+    async def run_embedding_request(self, *, input: list[str], **kwargs: Any) -> Any:
+        lock = self._get_request_lock()
+        async with lock:
+            now = time.monotonic()
+            if self._next_request_at > now:
+                await asyncio.sleep(self._next_request_at - now)
+
+            response = await self._aembedding_with_rate_limit_retries(
+                model=self._model,
+                input=input,
+                **kwargs,
+            )
+
+            now = time.monotonic()
+            if self._min_request_interval_seconds > 0:
+                self._next_request_at = now + self._min_request_interval_seconds
+            else:
+                self._next_request_at = now
+
+            return response
+
+    async def _get_dim(self) -> int:
+        if self._dim is not None:
+            return self._dim
+        async with self._get_lock():
+            if self._dim is not None:
+                return self._dim
+            response = await self.run_embedding_request(input=["hello"], **self._kwargs)
+            embedding = response.data[0]["embedding"]
+            self._dim = len(embedding)
+            return self._dim
+
+    @coco.fn.as_async(
+        batching=True,
+        max_batch_size=64,
+        memo=True,
+        version=1,
+        logic_tracking="self",
+    )
+    async def embed(
+        self,
+        texts: list[str],
+        input_type: str | None = None,
+    ) -> list[NDArray[np.float32]]:
+        kwargs = dict(self._kwargs)
+        if input_type is not None:
+            kwargs["input_type"] = input_type
+        response = await self.run_embedding_request(input=texts, **kwargs)
+        return [np.array(item["embedding"], dtype=np.float32) for item in response.data]

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/settings.py

@@ -92,6 +92,7 @@ class EmbeddingSettings:
     model: str
     provider: str = "litellm"
     device: str | None = None
+    min_interval_ms: int | None = None
 
 
 @dataclass
@@ -351,6 +352,8 @@ def _user_settings_to_dict(settings: UserSettings) -> dict[str, Any]:
     }
     if settings.embedding.device is not None:
         emb["device"] = settings.embedding.device
+    if settings.embedding.min_interval_ms is not None:
+        emb["min_interval_ms"] = settings.embedding.min_interval_ms
     d["embedding"] = emb
     if settings.envs:
         d["envs"] = dict(settings.envs)
@@ -367,6 +370,8 @@ def _user_settings_from_dict(d: dict[str, Any]) -> UserSettings:
         emb_kwargs["provider"] = emb_dict["provider"]
     if "device" in emb_dict:
         emb_kwargs["device"] = emb_dict["device"]
+    if "min_interval_ms" in emb_dict:
+        emb_kwargs["min_interval_ms"] = emb_dict["min_interval_ms"]
     embedding = EmbeddingSettings(**emb_kwargs)
     envs = d.get("envs", {})
     return UserSettings(embedding=embedding, envs=envs)

{cocoindex_code-0.2.10 → cocoindex_code-0.2.22}/src/cocoindex_code/shared.py

@@ -21,6 +21,7 @@ from .settings import EmbeddingSettings
 logger = logging.getLogger(__name__)
 
 SBERT_PREFIX = "sbert/"
+DEFAULT_LITELLM_MIN_INTERVAL_MS = 5
 
 # Models that define a "query" prompt for asymmetric retrieval.
 _QUERY_PROMPT_MODELS = {"nomic-ai/nomic-embed-code", "nomic-ai/CodeRankEmbed"}
@@ -63,11 +64,23 @@ def create_embedder(settings: EmbeddingSettings) -> Embedder:
         )
         logger.info("Embedding model: %s | device: %s", settings.model, settings.device)
     else:
-        from cocoindex.ops.litellm import LiteLLMEmbedder
+        from .litellm_embedder import PacedLiteLLMEmbedder
 
-        instance = LiteLLMEmbedder(settings.model)
+        min_interval_ms = (
+            settings.min_interval_ms
+            if settings.min_interval_ms is not None
+            else DEFAULT_LITELLM_MIN_INTERVAL_MS
+        )
+        instance = PacedLiteLLMEmbedder(
+            settings.model,
+            min_interval_ms=min_interval_ms,
+        )
         query_prompt_name = None
-        logger.info("Embedding model (LiteLLM): %s", settings.model)
+        logger.info(
+            "Embedding model (LiteLLM): %s | min_interval_ms: %s",
+            settings.model,
+            min_interval_ms,
+        )
 
     embedder = instance
     return instance

cocoindex_code-0.2.10/src/cocoindex_code/_version.py

@@ -1,34 +0,0 @@
-# file generated by setuptools-scm
-# don't change, don't track in version control
-
-__all__ = [
-    "__version__",
-    "__version_tuple__",
-    "version",
-    "version_tuple",
-    "__commit_id__",
-    "commit_id",
-]
-
-TYPE_CHECKING = False
-if TYPE_CHECKING:
-    from typing import Tuple
-    from typing import Union
-
-    VERSION_TUPLE = Tuple[Union[int, str], ...]
-    COMMIT_ID = Union[str, None]
-else:
-    VERSION_TUPLE = object
-    COMMIT_ID = object
-
-version: str
-__version__: str
-__version_tuple__: VERSION_TUPLE
-version_tuple: VERSION_TUPLE
-commit_id: COMMIT_ID
-__commit_id__: COMMIT_ID
-
-__version__ = version = '0.2.10'
-__version_tuple__ = version_tuple = (0, 2, 10)
-
-__commit_id__ = commit_id = None