Change LANGCHAIN_REVISION_ID to revision_id; use git describe as ba…

…ckup; don't overwrite metadata from env variables (#365)

python/langsmith/client.py

@@ -744,7 +744,9 @@ def create_run(
         metadata: dict = run_extra.setdefault("metadata", {})
         runtime_env = ls_env.get_runtime_and_metrics()
         langchain_metadata = ls_env.get_langchain_env_var_metadata()
-        metadata.update(**langchain_metadata)
+        metadata.update(
+            {k: v for k, v in langchain_metadata.items() if k not in metadata}
+        )
         run_extra["runtime"] = {**runtime_env, **runtime}
         headers = {
             **self._headers,

python/langsmith/env/_git.py

@@ -12,16 +12,12 @@
 T = TypeVar("T")
 
 
-def _exec_git(command: List[str]) -> Optional[str]:
+def exec_git(command: List[str]) -> Optional[str]:
     try:
         return subprocess.check_output(
             ["git"] + command, encoding="utf-8", stderr=subprocess.DEVNULL
         ).strip()
-    except FileNotFoundError:
-        logger.warning("git is not installed, or cannot be found in PATH")
-        return None
-    except subprocess.CalledProcessError as e:
-        logger.debug(f"Error running git command: {e}")
+    except BaseException:
         return None
 
 
@@ -41,18 +37,18 @@ class GitInfo(TypedDict, total=False):
 def get_git_info(remote: str = "origin") -> Optional[GitInfo]:
     """Get information about the git repository."""
 
-    if not _exec_git(["rev-parse", "--is-inside-work-tree"]):
+    if not exec_git(["rev-parse", "--is-inside-work-tree"]):
         return None
 
     return {
-        "remote_url": _exec_git(["remote", "get-url", remote]),
-        "commit": _exec_git(["rev-parse", "HEAD"]),
-        "commit_time": _exec_git(["log", "-1", "--format=%ct"]),
-        "branch": _exec_git(["rev-parse", "--abbrev-ref", "HEAD"]),
-        "tags": _exec_git(
+        "remote_url": exec_git(["remote", "get-url", remote]),
+        "commit": exec_git(["rev-parse", "HEAD"]),
+        "commit_time": exec_git(["log", "-1", "--format=%ct"]),
+        "branch": exec_git(["rev-parse", "--abbrev-ref", "HEAD"]),
+        "tags": exec_git(
             ["describe", "--tags", "--exact-match", "--always", "--dirty"]
         ),
-        "dirty": _exec_git(["status", "--porcelain"]) != "",
-        "author_name": _exec_git(["log", "-1", "--format=%an"]),
-        "author_email": _exec_git(["log", "-1", "--format=%ae"]),
+        "dirty": exec_git(["status", "--porcelain"]) != "",
+        "author_name": exec_git(["log", "-1", "--format=%an"]),
+        "author_email": exec_git(["log", "-1", "--format=%ae"]),
     }

python/langsmith/env/_runtime_env.py

@@ -7,6 +7,7 @@
 from typing import Dict, List, Optional, Union
 
 from langsmith.utils import get_docker_compose_command
+from langsmith.env._git import exec_git
 
 try:
     # psutil is an optional dependency
@@ -164,11 +165,31 @@ def get_langchain_env_var_metadata() -> dict:
         "LANGCHAIN_PROJECT",
         "LANGCHAIN_SESSION",
     }
-    return {
+    langchain_metadata = {
         k: v
         for k, v in os.environ.items()
-        if k.startswith("LANGCHAIN_") and k not in excluded and "key" not in k.lower()
+        if k.startswith("LANGCHAIN_")
+        and k not in excluded
+        and "key" not in k.lower()
+        and "secret" not in k.lower()
+        and "token" not in k.lower()
     }
+    env_revision_id = langchain_metadata.pop("LANGCHAIN_REVISION_ID", None)
+    if env_revision_id:
+        langchain_metadata["revision_id"] = env_revision_id
+    elif default_revision_id := _get_default_revision_id():
+        langchain_metadata["revision_id"] = default_revision_id
+
+    return langchain_metadata
+
+
+@functools.lru_cache(maxsize=1)
+def _get_default_revision_id() -> Optional[str]:
+    """Get the default revision ID based on `git describe`."""
+    try:
+        return exec_git(["describe", "--tags", "--dirty"])
+    except BaseException:
+        return None
 
 
 @functools.lru_cache(maxsize=1)