"""Tools for interacting with LangSmith prompts."""
from typing import Any, Dict
from langchain_core.load import dumpd
from langsmith import Client
def list_prompts_tool(client: Client, is_public: bool = False, limit: int = 20) -> Dict[str, Any]:
"""
Fetch prompts from LangSmith with optional filtering.
Args:
client: LangSmith Client instance
is_public (bool): Optional boolean to filter public/private prompts
limit (int): Optional limit to the number of prompts to return
Returns:
Dictionary containing the prompts and metadata
"""
try:
response = client.list_prompts(is_public=is_public, limit=limit)
# Convert list of tuples to a dictionary
response_dict = dict(response)
prompts_list = response_dict.get("repos", [])
total_count = response_dict.get("total", len(prompts_list))
formatted_prompts = []
for prompt in prompts_list:
prompt_dict = {}
for attr in [
"repo_handle",
"description",
"id",
"is_public",
"tags",
"owner",
"full_name",
"num_likes",
"num_downloads",
"num_views",
]:
prompt_dict[attr] = getattr(prompt, attr, None)
if hasattr(prompt, "created_at"):
prompt_dict["created_at"] = prompt.created_at.isoformat()
if hasattr(prompt, "updated_at"):
prompt_dict["updated_at"] = prompt.updated_at.isoformat()
formatted_prompts.append(prompt_dict)
return {"prompts": formatted_prompts, "total_count": total_count}
except Exception as e:
return {"error": f"Error fetching prompts: {str(e)}"}
def get_prompt_tool(client: Client, prompt_name: str = None, prompt_id: str = None) -> dict:
"""
Get a specific prompt (including model/tool bindings) by its name or ID, and return its full
string representation.
Args:
client: LangSmith Client instance
prompt_name: The full name of the prompt (e.g., 'owner/repo')
prompt_id: The UUID of the prompt
Returns:
String representation of the full prompt object (including tools and config)
"""
try:
if prompt_name:
prompt = client.pull_prompt(prompt_identifier=prompt_name)
elif prompt_id:
prompt = client.pull_prompt(prompt_identifier=prompt_id)
else:
return {"error": "Error: Either prompt_name or prompt_id must be provided."}
# Convert prompt object to JSON-serializable dictionary
try:
# Try using model_dump() if available (Pydantic v2)
if hasattr(prompt, "model_dump"):
return prompt.model_dump()
# Try using dict() method if available
elif hasattr(prompt, "dict"):
return prompt.dict()
# Fall back to dumpd() from langchain_core.load
else:
return dumpd(prompt)
except Exception as conv_error:
# If conversion fails, return a basic representation
return {
"type": type(prompt).__name__,
"repr": repr(prompt),
"error": f"Could not fully serialize prompt: {str(conv_error)}",
}
except Exception as e:
return {"error": f"Error fetching prompt: {str(e)}"}
