| Name | pylangdb JSON |
| Version |
0.3.6
JSON |
| download |
| home_page | None |
| Summary | Comprehensive LangDB integration package with enhanced agent support, LLM integration, and distributed tracing. |
| upload_time | 2025-07-11 04:53:18 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.10 |
| license | None |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# LangDB - Multi-Framework AI Agent Tracing and Client
[](https://badge.fury.io/py/pylangdb)
[](https://opensource.org/licenses/MIT)
## Key Features
LangDB exposes **two complementary capabilities**:
1. **Chat Completions Client** – Call LLMs using the `LangDb` Python client. This works as a drop-in replacement for `openai.ChatCompletion` while adding automatic usage, cost and latency reporting.
2. **Agent Tracing** – Instrument your existing AI framework (ADK, LangChain, CrewAI, etc.) with a single `init()` call. All calls are routed through the LangDB collector and are enriched with additional metadata regarding the framework is visible on the LangDB dashboard.
---
## ⚡ Quick Start (Chat Completions)
```bash
pip install pylangdb[client]
```
```python
from pylangdb.client import LangDb
# Initialize LangDB client
client = LangDb(api_key="your_api_key", project_id="your_project_id")
# Simple chat completion
resp = client.chat.completions.create(
model="openai/gpt-4o-mini",
messages=[{"role": "user", "content": "Hello!"}]
)
print(resp.choices[0].message.content)
```
---
## 🚀 Agent Tracing Quick Start
```bash
# Install the package with Google ADK support
pip install pylangdb[adk]
```
```python
# Import and initialize LangDB tracing
# First initialize LangDB before defining any agents
from pylangdb.adk import init
init()
import datetime
from zoneinfo import ZoneInfo
from google.adk.agents import Agent
def get_weather(city: str) -> dict:
if city.lower() != "new york":
return {"status": "error", "error_message": f"Weather information for '{city}' is not available."}
return {"status": "success", "report": "The weather in New York is sunny with a temperature of 25 degrees Celsius (77 degrees Fahrenheit)."}
def get_current_time(city: str) -> dict:
if city.lower() != "new york":
return {"status": "error", "error_message": f"Sorry, I don't have timezone information for {city}."}
tz = ZoneInfo("America/New_York")
now = datetime.datetime.now(tz)
return {"status": "success", "report": f'The current time in {city} is {now.strftime("%Y-%m-%d %H:%M:%S %Z%z")}'}
root_agent = Agent(
name="weather_time_agent",
model="gemini-2.0-flash",
description=("Agent to answer questions about the time and weather in a city." ),
instruction=("You are a helpful agent who can answer user questions about the time and weather in a city."),
tools=[get_weather, get_current_time],
)
```
> **Note:** Always initialize LangDB **before** importing any framework-specific classes to ensure proper instrumentation.
#### Example Trace Screenshot
## 🛠️ Supported Frameworks (Tracing)
| Framework | Installation | Import Pattern | Key Features |
| --- | --- | --- | --- |
| Google ADK | `pip install pylangdb[adk]` | `from pylangdb.adk import init` | Automatic sub-agent discovery |
| OpenAI | `pip install pylangdb[openai]` | `from pylangdb.openai import init` | Custom model provider support |
| LangChain | `pip install pylangdb[langchain]` | `from pylangdb.langchain import init` | Automatic chain tracing |
| CrewAI | `pip install pylangdb[crewai]` | `from pylangdb.crewai import init` | Multi-agent crew tracing |
| Agno | `pip install pylangdb[agno]` | `from pylangdb.agno import init` | Tool usage tracing, model interactions |
## 🔧 How It Works
LangDB uses intelligent monkey patching to instrument your AI frameworks at runtime:
<details>
<summary><b>👉 Click to see technical details for each framework</b></summary>
### Google ADK
- Patches `Agent.__init__` to inject callbacks
- Tracks agent hierarchies and tool usage
- Maintains thread context across invocations
### OpenAI
- Intercepts HTTP requests via `AsyncOpenAI.post`
- Propagates trace context via headers
- Correlates spans across agent interactions
### LangChain
- Modifies `httpx.Client.send` for request tracing
- Automatically tracks chains and agents
- Injects trace headers into all requests
### CrewAI
- Intercepts `litellm.completion` for LLM calls
- Tracks crew members and task delegation
- Propagates context through LiteLLM headers
### Agno
- Patches `LangDB.invoke` and client parameters
- Traces workflows and model interactions
- Maintains consistent session context
</details>
## 📦 Installation
```bash
# For client library functionality (chat completions, analytics, etc.)
pip install pylangdb[client]
# For framework tracing - install specific framework extras
pip install pylangdb[adk] # Google ADK tracing
pip install pylangdb[openai] # OpenAI agents tracing
pip install pylangdb[langchain] # LangChain tracing
pip install pylangdb[crewai] # CrewAI tracing
pip install pylangdb[agno] # Agno tracing
```
## 🔑 Configuration
Set your credentials (or pass them directly to the `init()` function):
```bash
export LANGDB_API_KEY="your-api-key"
export LANGDB_PROJECT_ID="your-project-id"
```
## Client Usage (Chat Completions)
### Initialize LangDb Client
```python
from pylangdb import LangDb
# Initialize with API key and project ID
client = LangDb(api_key="your_api_key", project_id="your_project_id")
```
### Chat Completions
```python
messages = [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say hello!"}
]
response = client.completion(
model="gemini-1.5-pro-latest",
messages=messages,
temperature=0.7,
max_tokens=100
)
```
### Thread Operations
#### Get Messages
Retrieve messages from a specific thread:
```python
messages = client.get_messages(thread_id="your_thread_id")
# Access message details
for message in messages:
print(f"Type: {message.type}")
print(f"Content: {message.content}")
if message.tool_calls:
for tool_call in message.tool_calls:
print(f"Tool: {tool_call.function.name}")
```
#### Get Thread Cost
Get cost and token usage information for a thread:
```python
usage = client.get_usage(thread_id="your_thread_id")
print(f"Total cost: ${usage.total_cost:.4f}")
print(f"Input tokens: {usage.total_input_tokens}")
print(f"Output tokens: {usage.total_output_tokens}")
```
### Analytics
Get analytics data for specific tags:
```python
# Get raw analytics data
analytics = client.get_analytics(
tags="model1,model2",
start_time_us=None, # Optional: defaults to 24 hours ago
end_time_us=None # Optional: defaults to current time
)
# Get analytics as a pandas DataFrame
df = client.get_analytics_dataframe(
tags="model1,model2",
start_time_us=None,
end_time_us=None
)
```
### Evaluate Multiple Threads
```python
df = client.create_evaluation_df(thread_ids=["thread1", "thread2"])
print(df.head())
```
### List Available Models
```python
models = client.list_models()
print(models)
```
## 🧩 Framework-Specific Examples (Tracing)
### Google ADK
```python
from pylangdb.adk import init
# Monkey-patch the client for tracing
init()
# Import your agents after initializing tracing
from google.adk.agents import Agent
from travel_concierge.sub_agents.booking.agent import booking_agent
from travel_concierge.sub_agents.in_trip.agent import in_trip_agent
from travel_concierge.sub_agents.inspiration.agent import inspiration_agent
from travel_concierge.sub_agents.planning.agent import planning_agent
from travel_concierge.sub_agents.post_trip.agent import post_trip_agent
from travel_concierge.sub_agents.pre_trip.agent import pre_trip_agent
from travel_concierge.tools.memory import _load_precreated_itinerary
root_agent = Agent(
model="openai/gpt-4.1",
name="root_agent",
description="A Travel Conceirge using the services of multiple sub-agents",
instruction="Instruct the travel concierge to plan a trip for the user.",
sub_agents=[
inspiration_agent,
planning_agent,
booking_agent,
pre_trip_agent,
in_trip_agent,
post_trip_agent,
],
before_agent_callback=_load_precreated_itinerary,
)
```
### OpenAI
```python
import uuid
import os
# Import LangDB tracing
from pylangdb.openai import init
# Initialize tracing
init()
# Import agent components
from agents import (
Agent,
Runner,
set_default_openai_client,
RunConfig,
ModelProvider,
Model,
OpenAIChatCompletionsModel
)
# Configure OpenAI client with environment variables
from openai import AsyncOpenAI
client = AsyncOpenAI(
api_key=os.environ.get("LANGDB_API_KEY"),
base_url=os.environ.get("LANGDB_API_BASE_URL"),
default_headers={
"x-project-id": os.environ.get("LANGDB_PROJECT_ID")
}
)
set_default_openai_client(client)
# Create a custom model provider
class CustomModelProvider(ModelProvider):
def get_model(self, model_name: str | None) -> Model:
return OpenAIChatCompletionsModel(model=model_name, openai_client=client)
CUSTOM_MODEL_PROVIDER = CustomModelProvider()
agent = Agent(
name="Math Tutor",
model="gpt-4.1",
instruction="You are a math tutor who can help students with their math homework.",
)
group_id = str(uuid.uuid4())
# Use the model provider with a unique group_id for tracing
async def run_agent():
response = await Runner.run(
triage_agent,
input="Hello World",
run_config=RunConfig(
model_provider=CUSTOM_MODEL_PROVIDER, # Inject custom model provider
group_id=group_id # Link all steps to the same trace
)
)
print(response.final_output)
# Run the async function with asyncio
asyncio.run(run_agent())
```
### LangChain
```python
import os
from pylangdb.langchain import init
init()
# Get environment variables for configuration
api_base = os.getenv("LANGDB_API_BASE_URL")
api_key = os.getenv("LANGDB_API_KEY")
if not api_key:
raise ValueError("Please set the LANGDB_API_KEY environment variable")
project_id = os.getenv("LANGDB_PROJECT_ID")
# Default headers for API requests
default_headers: dict[str, str] = {
"x-project-id": project-id
}
# Your existing LangChain code works with proper configuration
from langchain.chat_models import ChatOpenAI
from langchain.schema import HumanMessage
# Initialize OpenAI LLM with proper configuration
llm = ChatOpenAI(
model_name="gpt-4",
temperature=0.3,
openai_api_base=api_base,
openai_api_key=api_key,
default_headers=default_headers,
)
result = llm.invoke([HumanMessage(content="Hello, LangChain!")])
```
### CrewAI
```python
import os
from crewai import Agent, Task, Crew, LLM
from dotenv import load_dotenv
load_dotenv()
# Import and initialize LangDB tracing
from pylangdb.crewai import init
# Initialize tracing before importing or creating any agents
init()
# Initialize API credentials
api_key = os.environ.get("LANGDB_API_KEY")
api_base = os.environ.get("LANGDB_API_BASE_URL")
project_id = os.environ.get("LANGDB_PROJECT_ID")
# Create LLM with proper headers
llm = LLM(
model="gpt-4",
api_key=api_key,
base_url=api_base,
extra_headers={
"x-project-id": project_id
}
)
# Create and use your CrewAI components as usual
# They will be automatically traced by LangDB
researcher = Agent(
role="researcher",
goal="Research the topic thoroughly",
backstory="You are an expert researcher",
llm=llm,
verbose=True
)
task = Task(
description="Research the given topic",
agent=researcher
)
crew = Crew(agents=[researcher], tasks=[task])
result = crew.kickoff()
```
### Agno
```python
import os
from agno.agent import Agent
from agno.tools.duckduckgo import DuckDuckGoTools
# Import and initialize LangDB tracing
from pylangdb.agno import init
init()
# Import LangDB model after initializing tracing
from agno.models.langdb import LangDB
# Create agent with LangDB model
agent = Agent(
name="Web Agent",
role="Search the web for information",
model=LangDB(
id="openai/gpt-4",
base_url=os.getenv("LANGDB_API_BASE_URL") + '/' + os.getenv("LANGDB_PROJECT_ID") + '/v1',
api_key=os.getenv("LANGDB_API_KEY"),
project_id=os.getenv("LANGDB_PROJECT_ID"),
),
tools=[DuckDuckGoTools()],
instructions="Answer questions using web search",
show_tool_calls=True,
markdown=True,
)
# Use the agent
response = agent.run("What is LangDB?")
```
## ⚙️ Advanced Configuration
### Environment Variables
| Variable | Description | Default |
|----------|-------------|---------|
| `LANGDB_API_KEY` | Your LangDB API key | Required |
| `LANGDB_PROJECT_ID` | Your LangDB project ID | Required |
| `LANGDB_API_BASE_URL` | LangDB API base URL | `https://api.us-east-1.langdb.ai` |
| `LANGDB_TRACING_BASE_URL` | Tracing collector endpoint | `https://api.us-east-1.langdb.ai:4317` |
| `LANGDB_TRACING` | Enable/disable tracing | `true` |
| `LANGDB_TRACING_EXPORTERS` | Comma-separated list of exporters | `otlp`, `console` |
### Custom Configuration
All `init()` functions accept the same optional parameters:
```python
from langdb.openai import init
init(
collector_endpoint='https://api.us-east-1.langdb.ai:4317',
api_key="langdb-api-key",
project_id="langdb-project-id"
)
```
## 🔬 Technical Details
### Session and Thread Management
- **Thread ID**: Maintains consistent session identifiers across agent calls
- **Run ID**: Unique identifier for each execution trace
- **Invocation Tracking**: Tracks the sequence of agent invocations
- **State Persistence**: Maintains context across callbacks and sub-agent interactions
### Distributed Tracing
- **OpenTelemetry Integration**: Uses OpenTelemetry for standardized tracing
- **Attribute Propagation**: Automatically propagates LangDB-specific attributes
- **Span Correlation**: Links related spans across different agents and frameworks
- **Custom Exporters**: Supports multiple export formats (OTLP, Console)
## API Reference
### Initialization Functions
Each framework has a simple `init()` function that handles all necessary setup:
- `langdb.adk.init()`: Patches Google ADK Agent class with LangDB callbacks
- `langdb.openai.init()`: Initializes OpenAI agents tracing
- `langdb.langchain.init()`: Initializes LangChain tracing
- `langdb.crewai.init()`: Initializes CrewAI tracing
- `langdb.agno.init()`: Initializes Agno tracing
All init functions accept optional parameters for custom configuration (collector_endpoint, api_key, project_id)
## 🛟 Troubleshooting
### Common Issues
1. **Missing API Key**: Ensure `LANGDB_API_KEY` and `LANGDB_PROJECT_ID` are set
2. **Tracing Not Working**: Check that initialization functions are called before creating agents
3. **Network Issues**: Verify collector endpoint is accessible
4. **Framework Conflicts**: Initialize LangDB integration before other instrumentation
### Debug Mode
Enable console output for debugging:
```bash
export LANGDB_TRACING_EXPORTERS="otlp,console"
```
Disable tracing entirely:
```bash
export LANGDB_TRACING="false"
```
## Development
### Setting up the environment
1. Clone the repository
2. Create a `.env` file with your credentials:
```bash
LANGDB_API_KEY="your_api_key"
LANGDB_PROJECT_ID="your_project_id"
```
### Running Tests
```bash
python -m unittest tests/client.py -v
```
## Publishing
```bash
poetry build
poetry publish
```
## 📋 Requirements
- Python >= 3.10
- Framework-specific dependencies (installed automatically)
- OpenTelemetry libraries (installed automatically)
## 📄 License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## 🤝 Support
- **GitHub Issues**: [Report bugs and feature requests](https://github.com/langdb/pylangdb/issues)
- **Documentation**: [LangDB Documentation](https://docs.langdb.ai)
Raw data
{
"_id": null,
"home_page": null,
"name": "pylangdb",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": null,
"keywords": null,
"author": null,
"author_email": null,
"download_url": "https://files.pythonhosted.org/packages/dd/9d/a97bc244ba95185d1e682485443d1d5e6b23caa45a8e9191a7fff78be61f/pylangdb-0.3.6.tar.gz",
"platform": null,
"description": "# LangDB - Multi-Framework AI Agent Tracing and Client\n\n[](https://badge.fury.io/py/pylangdb)\n[](https://opensource.org/licenses/MIT)\n\n## Key Features\n\nLangDB exposes **two complementary capabilities**:\n\n1. **Chat Completions Client** \u2013 Call LLMs using the `LangDb` Python client. This works as a drop-in replacement for `openai.ChatCompletion` while adding automatic usage, cost and latency reporting.\n2. **Agent Tracing** \u2013 Instrument your existing AI framework (ADK, LangChain, CrewAI, etc.) with a single `init()` call. All calls are routed through the LangDB collector and are enriched with additional metadata regarding the framework is visible on the LangDB dashboard.\n\n---\n\n## \u26a1 Quick Start (Chat Completions)\n\n```bash\npip install pylangdb[client]\n```\n\n```python\nfrom pylangdb.client import LangDb\n\n# Initialize LangDB client\nclient = LangDb(api_key=\"your_api_key\", project_id=\"your_project_id\")\n\n# Simple chat completion\nresp = client.chat.completions.create(\n model=\"openai/gpt-4o-mini\",\n messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\nprint(resp.choices[0].message.content)\n```\n\n---\n\n## \ud83d\ude80 Agent Tracing Quick Start\n\n```bash\n# Install the package with Google ADK support\npip install pylangdb[adk]\n```\n\n```python\n# Import and initialize LangDB tracing\n# First initialize LangDB before defining any agents\nfrom pylangdb.adk import init\ninit()\n\nimport datetime\nfrom zoneinfo import ZoneInfo\nfrom google.adk.agents import Agent\n\ndef get_weather(city: str) -> dict:\n if city.lower() != \"new york\":\n return {\"status\": \"error\", \"error_message\": f\"Weather information for '{city}' is not available.\"}\n return {\"status\": \"success\", \"report\": \"The weather in New York is sunny with a temperature of 25 degrees Celsius (77 degrees Fahrenheit).\"}\n\ndef get_current_time(city: str) -> dict:\n if city.lower() != \"new york\":\n return {\"status\": \"error\", \"error_message\": f\"Sorry, I don't have timezone information for {city}.\"}\n tz = ZoneInfo(\"America/New_York\")\n now = datetime.datetime.now(tz)\n return {\"status\": \"success\", \"report\": f'The current time in {city} is {now.strftime(\"%Y-%m-%d %H:%M:%S %Z%z\")}'}\n\nroot_agent = Agent(\n name=\"weather_time_agent\",\n model=\"gemini-2.0-flash\",\n description=(\"Agent to answer questions about the time and weather in a city.\" ),\n instruction=(\"You are a helpful agent who can answer user questions about the time and weather in a city.\"),\n tools=[get_weather, get_current_time],\n)\n```\n\n> **Note:** Always initialize LangDB **before** importing any framework-specific classes to ensure proper instrumentation.\n\n\n#### Example Trace Screenshot\n\n\n\n## \ud83d\udee0\ufe0f Supported Frameworks (Tracing)\n\n| Framework | Installation | Import Pattern | Key Features |\n| --- | --- | --- | --- |\n| Google ADK | `pip install pylangdb[adk]` | `from pylangdb.adk import init` | Automatic sub-agent discovery |\n| OpenAI | `pip install pylangdb[openai]` | `from pylangdb.openai import init` | Custom model provider support |\n| LangChain | `pip install pylangdb[langchain]` | `from pylangdb.langchain import init` | Automatic chain tracing |\n| CrewAI | `pip install pylangdb[crewai]` | `from pylangdb.crewai import init` | Multi-agent crew tracing |\n| Agno | `pip install pylangdb[agno]` | `from pylangdb.agno import init` | Tool usage tracing, model interactions |\n\n## \ud83d\udd27 How It Works\n\nLangDB uses intelligent monkey patching to instrument your AI frameworks at runtime:\n\n<details>\n<summary><b>\ud83d\udc49 Click to see technical details for each framework</b></summary>\n\n### Google ADK\n- Patches `Agent.__init__` to inject callbacks\n- Tracks agent hierarchies and tool usage\n- Maintains thread context across invocations\n\n### OpenAI\n- Intercepts HTTP requests via `AsyncOpenAI.post`\n- Propagates trace context via headers\n- Correlates spans across agent interactions\n\n### LangChain\n- Modifies `httpx.Client.send` for request tracing\n- Automatically tracks chains and agents\n- Injects trace headers into all requests\n\n### CrewAI\n- Intercepts `litellm.completion` for LLM calls\n- Tracks crew members and task delegation\n- Propagates context through LiteLLM headers\n\n### Agno\n- Patches `LangDB.invoke` and client parameters\n- Traces workflows and model interactions\n- Maintains consistent session context\n</details>\n\n## \ud83d\udce6 Installation\n\n```bash\n# For client library functionality (chat completions, analytics, etc.)\npip install pylangdb[client]\n\n# For framework tracing - install specific framework extras\npip install pylangdb[adk] # Google ADK tracing\npip install pylangdb[openai] # OpenAI agents tracing\npip install pylangdb[langchain] # LangChain tracing\npip install pylangdb[crewai] # CrewAI tracing\npip install pylangdb[agno] # Agno tracing\n```\n\n## \ud83d\udd11 Configuration\n\nSet your credentials (or pass them directly to the `init()` function):\n\n```bash\nexport LANGDB_API_KEY=\"your-api-key\"\nexport LANGDB_PROJECT_ID=\"your-project-id\"\n```\n\n## Client Usage (Chat Completions)\n\n### Initialize LangDb Client\n\n```python\nfrom pylangdb import LangDb\n\n# Initialize with API key and project ID\nclient = LangDb(api_key=\"your_api_key\", project_id=\"your_project_id\")\n```\n\n### Chat Completions\n\n```python\nmessages = [\n {\"role\": \"system\", \"content\": \"You are a helpful assistant.\"},\n {\"role\": \"user\", \"content\": \"Say hello!\"}\n]\n\nresponse = client.completion(\n model=\"gemini-1.5-pro-latest\",\n messages=messages,\n temperature=0.7,\n max_tokens=100\n)\n```\n\n### Thread Operations\n\n#### Get Messages\nRetrieve messages from a specific thread:\n\n```python\nmessages = client.get_messages(thread_id=\"your_thread_id\")\n\n# Access message details\nfor message in messages:\n print(f\"Type: {message.type}\")\n print(f\"Content: {message.content}\")\n if message.tool_calls:\n for tool_call in message.tool_calls:\n print(f\"Tool: {tool_call.function.name}\")\n```\n\n#### Get Thread Cost\nGet cost and token usage information for a thread:\n\n```python\nusage = client.get_usage(thread_id=\"your_thread_id\")\nprint(f\"Total cost: ${usage.total_cost:.4f}\")\nprint(f\"Input tokens: {usage.total_input_tokens}\")\nprint(f\"Output tokens: {usage.total_output_tokens}\")\n```\n\n### Analytics\n\nGet analytics data for specific tags:\n\n```python\n# Get raw analytics data\nanalytics = client.get_analytics(\n tags=\"model1,model2\",\n start_time_us=None, # Optional: defaults to 24 hours ago\n end_time_us=None # Optional: defaults to current time\n)\n\n# Get analytics as a pandas DataFrame\ndf = client.get_analytics_dataframe(\n tags=\"model1,model2\",\n start_time_us=None,\n end_time_us=None\n)\n```\n\n### Evaluate Multiple Threads\n```python\ndf = client.create_evaluation_df(thread_ids=[\"thread1\", \"thread2\"])\nprint(df.head())\n```\n\n### List Available Models\n```python\nmodels = client.list_models()\nprint(models)\n```\n\n## \ud83e\udde9 Framework-Specific Examples (Tracing)\n\n### Google ADK\n\n```python\nfrom pylangdb.adk import init\n\n# Monkey-patch the client for tracing\ninit()\n\n# Import your agents after initializing tracing\nfrom google.adk.agents import Agent\nfrom travel_concierge.sub_agents.booking.agent import booking_agent\nfrom travel_concierge.sub_agents.in_trip.agent import in_trip_agent\nfrom travel_concierge.sub_agents.inspiration.agent import inspiration_agent\nfrom travel_concierge.sub_agents.planning.agent import planning_agent\nfrom travel_concierge.sub_agents.post_trip.agent import post_trip_agent\nfrom travel_concierge.sub_agents.pre_trip.agent import pre_trip_agent\nfrom travel_concierge.tools.memory import _load_precreated_itinerary\n\n\nroot_agent = Agent(\n model=\"openai/gpt-4.1\",\n name=\"root_agent\",\n description=\"A Travel Conceirge using the services of multiple sub-agents\",\n instruction=\"Instruct the travel concierge to plan a trip for the user.\",\n sub_agents=[\n inspiration_agent,\n planning_agent,\n booking_agent,\n pre_trip_agent,\n in_trip_agent,\n post_trip_agent,\n ],\n before_agent_callback=_load_precreated_itinerary,\n)\n```\n\n### OpenAI\n\n```python\nimport uuid\nimport os\n\n# Import LangDB tracing\nfrom pylangdb.openai import init\n\n# Initialize tracing\ninit()\n\n# Import agent components\nfrom agents import (\n Agent,\n Runner,\n set_default_openai_client,\n RunConfig,\n ModelProvider,\n Model,\n OpenAIChatCompletionsModel\n)\n\n# Configure OpenAI client with environment variables\nfrom openai import AsyncOpenAI\n\nclient = AsyncOpenAI(\n api_key=os.environ.get(\"LANGDB_API_KEY\"),\n base_url=os.environ.get(\"LANGDB_API_BASE_URL\"),\n default_headers={\n \"x-project-id\": os.environ.get(\"LANGDB_PROJECT_ID\")\n }\n)\nset_default_openai_client(client)\n\n# Create a custom model provider\nclass CustomModelProvider(ModelProvider):\n def get_model(self, model_name: str | None) -> Model:\n return OpenAIChatCompletionsModel(model=model_name, openai_client=client)\n\nCUSTOM_MODEL_PROVIDER = CustomModelProvider()\n\nagent = Agent(\n name=\"Math Tutor\",\n model=\"gpt-4.1\",\n instruction=\"You are a math tutor who can help students with their math homework.\",\n)\n\ngroup_id = str(uuid.uuid4())\n# Use the model provider with a unique group_id for tracing\nasync def run_agent():\n response = await Runner.run(\n triage_agent,\n input=\"Hello World\",\n run_config=RunConfig(\n model_provider=CUSTOM_MODEL_PROVIDER, # Inject custom model provider\n group_id=group_id # Link all steps to the same trace\n )\n )\n print(response.final_output)\n\n# Run the async function with asyncio\nasyncio.run(run_agent())\n```\n\n### LangChain\n\n```python\nimport os\nfrom pylangdb.langchain import init\n\ninit()\n\n# Get environment variables for configuration\napi_base = os.getenv(\"LANGDB_API_BASE_URL\")\napi_key = os.getenv(\"LANGDB_API_KEY\")\nif not api_key:\n raise ValueError(\"Please set the LANGDB_API_KEY environment variable\")\n\nproject_id = os.getenv(\"LANGDB_PROJECT_ID\")\n\n# Default headers for API requests\ndefault_headers: dict[str, str] = {\n \"x-project-id\": project-id\n}\n\n# Your existing LangChain code works with proper configuration\nfrom langchain.chat_models import ChatOpenAI\nfrom langchain.schema import HumanMessage\n\n# Initialize OpenAI LLM with proper configuration\nllm = ChatOpenAI(\n model_name=\"gpt-4\",\n temperature=0.3,\n openai_api_base=api_base,\n openai_api_key=api_key,\n default_headers=default_headers,\n)\nresult = llm.invoke([HumanMessage(content=\"Hello, LangChain!\")])\n```\n\n### CrewAI\n\n```python\nimport os\nfrom crewai import Agent, Task, Crew, LLM\nfrom dotenv import load_dotenv\n\nload_dotenv()\n\n# Import and initialize LangDB tracing\nfrom pylangdb.crewai import init\n\n# Initialize tracing before importing or creating any agents\ninit()\n\n# Initialize API credentials\napi_key = os.environ.get(\"LANGDB_API_KEY\")\napi_base = os.environ.get(\"LANGDB_API_BASE_URL\")\nproject_id = os.environ.get(\"LANGDB_PROJECT_ID\")\n\n# Create LLM with proper headers\nllm = LLM(\n model=\"gpt-4\",\n api_key=api_key,\n base_url=api_base,\n extra_headers={\n \"x-project-id\": project_id\n }\n)\n\n# Create and use your CrewAI components as usual\n# They will be automatically traced by LangDB\nresearcher = Agent(\n role=\"researcher\",\n goal=\"Research the topic thoroughly\",\n backstory=\"You are an expert researcher\",\n llm=llm,\n verbose=True\n)\n\ntask = Task(\n description=\"Research the given topic\",\n agent=researcher\n)\n\ncrew = Crew(agents=[researcher], tasks=[task])\nresult = crew.kickoff()\n```\n\n### Agno\n\n```python\nimport os\nfrom agno.agent import Agent\nfrom agno.tools.duckduckgo import DuckDuckGoTools\n\n# Import and initialize LangDB tracing\nfrom pylangdb.agno import init\ninit()\n\n# Import LangDB model after initializing tracing\nfrom agno.models.langdb import LangDB\n\n# Create agent with LangDB model\nagent = Agent(\n name=\"Web Agent\",\n role=\"Search the web for information\",\n model=LangDB(\n id=\"openai/gpt-4\",\n base_url=os.getenv(\"LANGDB_API_BASE_URL\") + '/' + os.getenv(\"LANGDB_PROJECT_ID\") + '/v1',\n api_key=os.getenv(\"LANGDB_API_KEY\"),\n project_id=os.getenv(\"LANGDB_PROJECT_ID\"),\n ),\n tools=[DuckDuckGoTools()],\n instructions=\"Answer questions using web search\",\n show_tool_calls=True,\n markdown=True,\n)\n\n# Use the agent\nresponse = agent.run(\"What is LangDB?\")\n```\n\n## \u2699\ufe0f Advanced Configuration\n\n### Environment Variables\n\n| Variable | Description | Default |\n|----------|-------------|---------|\n| `LANGDB_API_KEY` | Your LangDB API key | Required |\n| `LANGDB_PROJECT_ID` | Your LangDB project ID | Required |\n| `LANGDB_API_BASE_URL` | LangDB API base URL | `https://api.us-east-1.langdb.ai` |\n| `LANGDB_TRACING_BASE_URL` | Tracing collector endpoint | `https://api.us-east-1.langdb.ai:4317` |\n| `LANGDB_TRACING` | Enable/disable tracing | `true` |\n| `LANGDB_TRACING_EXPORTERS` | Comma-separated list of exporters | `otlp`, `console` |\n\n### Custom Configuration\n\nAll `init()` functions accept the same optional parameters:\n\n```python\nfrom langdb.openai import init\n\ninit(\n collector_endpoint='https://api.us-east-1.langdb.ai:4317',\n api_key=\"langdb-api-key\",\n project_id=\"langdb-project-id\"\n)\n```\n\n## \ud83d\udd2c Technical Details\n\n### Session and Thread Management\n\n- **Thread ID**: Maintains consistent session identifiers across agent calls\n- **Run ID**: Unique identifier for each execution trace\n- **Invocation Tracking**: Tracks the sequence of agent invocations\n- **State Persistence**: Maintains context across callbacks and sub-agent interactions\n\n### Distributed Tracing\n\n- **OpenTelemetry Integration**: Uses OpenTelemetry for standardized tracing\n- **Attribute Propagation**: Automatically propagates LangDB-specific attributes\n- **Span Correlation**: Links related spans across different agents and frameworks\n- **Custom Exporters**: Supports multiple export formats (OTLP, Console)\n\n## API Reference\n\n### Initialization Functions\n\nEach framework has a simple `init()` function that handles all necessary setup:\n\n- `langdb.adk.init()`: Patches Google ADK Agent class with LangDB callbacks\n- `langdb.openai.init()`: Initializes OpenAI agents tracing\n- `langdb.langchain.init()`: Initializes LangChain tracing\n- `langdb.crewai.init()`: Initializes CrewAI tracing\n- `langdb.agno.init()`: Initializes Agno tracing\n\nAll init functions accept optional parameters for custom configuration (collector_endpoint, api_key, project_id)\n\n## \ud83d\udedf Troubleshooting\n\n### Common Issues\n\n1. **Missing API Key**: Ensure `LANGDB_API_KEY` and `LANGDB_PROJECT_ID` are set\n2. **Tracing Not Working**: Check that initialization functions are called before creating agents\n3. **Network Issues**: Verify collector endpoint is accessible\n4. **Framework Conflicts**: Initialize LangDB integration before other instrumentation\n\n### Debug Mode\n\nEnable console output for debugging:\n```bash\nexport LANGDB_TRACING_EXPORTERS=\"otlp,console\"\n```\n\nDisable tracing entirely:\n```bash\nexport LANGDB_TRACING=\"false\"\n```\n\n## Development\n\n### Setting up the environment\n\n1. Clone the repository\n2. Create a `.env` file with your credentials:\n```bash\nLANGDB_API_KEY=\"your_api_key\"\nLANGDB_PROJECT_ID=\"your_project_id\"\n```\n\n### Running Tests\n\n```bash\npython -m unittest tests/client.py -v\n```\n\n## Publishing\n\n```bash\npoetry build\npoetry publish\n```\n\n## \ud83d\udccb Requirements\n\n- Python >= 3.10\n- Framework-specific dependencies (installed automatically)\n- OpenTelemetry libraries (installed automatically)\n\n## \ud83d\udcc4 License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\n## \ud83e\udd1d Support\n\n- **GitHub Issues**: [Report bugs and feature requests](https://github.com/langdb/pylangdb/issues)\n- **Documentation**: [LangDB Documentation](https://docs.langdb.ai)\n",
"bugtrack_url": null,
"license": null,
"summary": "Comprehensive LangDB integration package with enhanced agent support, LLM integration, and distributed tracing.",
"version": "0.3.6",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "7b62538fedb5fbb96712f07edb7b504cb02e4110525845ac66896b612032d9e9",
"md5": "1e727698730517a51c74c53659d7e4fd",
"sha256": "db5c2eaa3c5b211b65c5e1d530557817049ebdc63b4315f0d778592026f4bc41"
},
"downloads": -1,
"filename": "pylangdb-0.3.6-py3-none-any.whl",
"has_sig": false,
"md5_digest": "1e727698730517a51c74c53659d7e4fd",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 23297,
"upload_time": "2025-07-11T04:53:15",
"upload_time_iso_8601": "2025-07-11T04:53:15.967026Z",
"url": "https://files.pythonhosted.org/packages/7b/62/538fedb5fbb96712f07edb7b504cb02e4110525845ac66896b612032d9e9/pylangdb-0.3.6-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "dd9da97bc244ba95185d1e682485443d1d5e6b23caa45a8e9191a7fff78be61f",
"md5": "09700c1bf9950a65ae3745eca776c285",
"sha256": "47d265c60e3d6f6d7d26c1cf38266145c157cf572251c1eb02d5402ee2bbb9f7"
},
"downloads": -1,
"filename": "pylangdb-0.3.6.tar.gz",
"has_sig": false,
"md5_digest": "09700c1bf9950a65ae3745eca776c285",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 25901,
"upload_time": "2025-07-11T04:53:18",
"upload_time_iso_8601": "2025-07-11T04:53:18.534526Z",
"url": "https://files.pythonhosted.org/packages/dd/9d/a97bc244ba95185d1e682485443d1d5e6b23caa45a8e9191a7fff78be61f/pylangdb-0.3.6.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-07-11 04:53:18",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "pylangdb"
}
