# ms-public-sdk
## Quick Start Guide
### Using the Nstream SDK for Recommending Promotions
### Introduction
The Nstream SDK is designed for building and managing neural network pipelines that incorporate large language models (LLMs) and stream processing capabilities. This guide will walk you through creating a simple but powerful neural network pipeline for generating insights from streaming data using the Nstream SDK.
### Installation
To begin using the Nstream SDK in your projects, install it via pip with the following command:
```bash
pip install nstream
```
After installation, import the necessary modules and classes as follows:
```python
from nstreamai_nstream_ai.core.nsinit import NsInit
from nstreamai_nstream_ai.core.nsnode import (
NsNode, NsLink, NsProvider, NsProviderType, Nstream
)
from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM
from nstreamai_nstream_ai.core.nsgraph import NsGraph
from utils.logger import logger
import sys
```
## Core Module Imports
### NsInit
**Import**: `from nstreamai_nstream_ai.core.nsinit import NsInit`
**Description**: This import brings in the `NsInit` class responsible for initializing and configuring the connection settings for the Nstream SDK.
**Usage**: It is used to set up the initial connection with the Nstream API using credentials like API key, username, and password.
### NsNode, NsLink, NsProvider, NsProviderType, Nstream
**Import**: `from nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream`
**Description**: This import includes multiple classes that are fundamental to constructing nodes within an Nstream pipeline:
- **NsNode**: Represents a single node in the neural network graph.
- **NsLink**: Used to define connections between nodes and data sources or sinks.
- **NsProvider**: Specifies the data source or sink type.
- **NsProviderType**: Enumerates different types of data providers.
- **Nstream**: Might be used for additional functionality directly related to stream operations.
**Usage**: These are used to define and link the functional components of a neural network pipeline, such as data inputs, transformations, and outputs.
### NsNeuron, NstreamLLM
**Import**: `from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM`
**Description**: This import involves classes related to the neural network computation units within the SDK:
- **NsNeuron**: Represents a neuron unit which can execute specific neural network computations or models.
- **NstreamLLM**: Pertains to specific large language model configurations that can be deployed within neurons.
**Usage**: These are used to specify and configure the large language models that perform the actual analytics and insights generation in the pipeline.
### NsGraph
**Import**: `from nstreamai_nstream_ai.core.nsgraph import NsGraph`
**Description**: Imports the `NsGraph` class, which manages the execution flow of neural network nodes defined with `NsNode`.
**Usage**: This class is crucial for defining the execution order and dependencies between nodes, as well as starting and stopping the data processing workflow.
### Configuration
Start by initializing the SDK and connecting to the Nstream service:
```python
try:
logger.info("Starting main execution")
conn = NsInit(
api_key="your_api_key",
username="your_username",
password="your_password").connect()
logger.info("Connected to NsInit")
except Exception as e:
logger.exception(
"Exception occurred while initializing NsInit"
)
print(e)
sys.exit()
```
### Building the Pipeline
#### Node 1: User Interaction Summarization and Insight Generation
Create the first node to summarize user interactions and generate insights:
```python
prompt_event = Nstream.event()
ns_node_1_prompt_text = f"Generate a general insight using a user's data - {prompt_event}"
ns_node_1_context_transform_prompt_text = "Transform unstructured user data into a detailed JSON format..."
ns_node_1 = NsNode(
node_name="GraphNode1",
prompt=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
prompt_text=ns_node_1_prompt_text
),
context=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
context_transform_prompt_text=ns_node_1_context_transform_prompt_text),
neuron=NsNeuron(NstreamLLM.mistral_7b()),
socket=conn)
logger.info("GraphNode1 configured")
```
#### Node 2: Personalized Recommendations
Set up the second node to use insights from Node 1 to generate personalized recommendations:
```python
ns_node_2_prompt_text = "Based on general insight, provide a list of recommended promotional offers"
ns_node_2_context_transform_prompt_text = "Transform unstructured promotional data into a structured JSON format..."
ns_node_2 = NsNode(
node_name="GraphNode2",
prompt=NsLink(
socket=conn,
provider=NsProvider(
type=NsProviderType().Source).mongodb(),prompt_text=ns_node_2_prompt_text),
context=ns_node_1.output(
context_transform_prompt_text=ns_node_2_context_transform_prompt_text),
neuron=NsNeuron(NstreamLLM.llama2_7b()),
socket=conn)
logger.info("GraphNode2 configured")
```
#### Running the Pipeline
Execute the configured pipeline and handle the output:
```python
ns_graph_sink = NsLink(socket=conn, provider=NsProvider(type=NsProviderType().Sink).terminal())
logger.info("Graph sink configured")
ns_graph = NsGraph(conn).start(ns_node_1).end(ns_node_2).submit(ns_graph_sink)
logger.info("Graph execution started")
ns_graph.terminate(run_time=6)
logger.info("Graph execution terminated")
print("Execution Completed")
logger.info("Main execution completed")
```
Final code would look like this:
```python
from nstreamai_nstream_ai.core.nsinit import NsInit
from nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream
from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM
from nstreamai_nstream_ai.core.nsgraph import NsGraph
from utils.logger import logger
import sys
if __name__ == "__main__":
try:
logger.info("Starting main execution")
conn = NsInit(
api_key="your_api_key",
username="your_username",
password="your_password").connect()
logger.info("Connected to NsInit")
except Exception as e:
logger.exception("Exception occurred while initializing NsInit")
print(e)
sys.exit()
# Node 1: User Interaction Summarization and Insight Generation
# Objective: Generate Insight using the raw cookie data and historical summarizes of the user’s interactions with promotions on a website.
prompt_event = Nstream.event()
# Prompt for insight generation
ns_node_1_prompt_text = f"Generate a general insight using a input user's data - {prompt_event}" # Uses stored structured summary from db
# Context Transformation Prompt:
ns_node_1_context_transform_prompt_text = "Transform unstructured user data into a detailed JSON format that captures comprehensive user interactions, particularly with promotional activities, and includes other vital user data points like geolocation data, data source, and timestamps."
ns_node_1 = NsNode(
node_name="GraphNode1",
prompt=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
prompt_text=ns_node_1_prompt_text
),
context=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
context_transform_prompt_text=ns_node_1_context_transform_prompt_text),
neuron=NsNeuron(NstreamLLM.mistral_7b()),
socket=conn)
logger.info("GraphNode1 configured")
# Node 2: Personalized Recommendations
# Objective: Use the insights from Node1 to generate personalized product recommendations based on the user's interactions with promotions.
# Prompt for Product Listing
ns_node_2_prompt_text = "Based on general insight, provide a list the recommended promotional offers"
# Context Transformation Prompt
ns_node_2_context_transform_prompt_text = "Transform unstructured promotional data into a structured JSON format for better recommendation accuracy. Include essential details like offer ID, title, discount, and audience criteria. Ensure all offers are consistently formatted to enhance analysis and comparison."
ns_node_2 = NsNode(
node_name="GraphNode2",
prompt=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
prompt_text=ns_node_2_prompt_text,
),
context=ns_node_1.output(
context_transform_prompt_text=ns_node_2_context_transform_prompt_text),
neuron=NsNeuron(NstreamLLM.llama2_7b()),
socket=conn)
logger.info("GraphNode2 configured")
ns_graph_sink = NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Sink).terminal(),
)
logger.info("Graph sink configured")
ns_graph = NsGraph(conn).start(ns_node_1).end(ns_node_2).submit(ns_graph_sink)
logger.info("Graph execution started")
ns_graph.terminate(run_time=6)
logger.info("Graph execution terminated")
logger.info("Main execution completed")
```
##
### Using the Nstream SDK for Genaration Usage Example
#### Building the Pipeline
##### Node 1: User Interaction Summarization
Create the first node to summarize user interactions:
```python
prompt_event = Nstream.event()
ns_node_1_prompt_text = f"Generate a general insight using input user's data - {prompt_event}"
# Configure the node with its necessary components
ns_node_1 = NsNode(
node_name="GraphNode1",
prompt=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
prompt_text=ns_node_1_prompt_text
),
neuron=NsNeuron(NstreamLLM.mistral_7b()),
socket=conn)
logger.info("GraphNode1 configured")
```
#### Running the Pipeline
Execute the configured pipeline and handle the output:
```python
ns_graph_sink = NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Sink).terminal(),
)
logger.info("Graph sink configured")
# Initialize the graph, start execution, and submit it for processing
ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)
logger.info("Graph execution started")
# Terminate the graph execution after a predefined runtime
ns_graph.terminate(run_time=6)
logger.info("Graph execution terminated")
# Signal completion of the main execution
logger.info("Main execution completed")
```
Final Code:
```python
# Import necessary modules from the Nstream SDK and utility libraries
from nstreamai_nstream_ai.core.nsinit import NsInit
from nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream
from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM
from nstreamai_nstream_ai.core.nsgraph import NsGraph
from utils.logger import logger
import sys
if __name__ == "__main__":
try:
# Initialize logging and connect to the Nstream service
logger.info("Starting main execution")
conn = NsInit(
api_key="your_api_key",
username="your_username",
password="your_password").connect()
logger.info("Connected to NsInit")
except Exception as e:
# Log any exceptions during initialization and exit
logger.exception("Exception occurred while initializing NsInit")
print(e)
sys.exit()
# Define Node 1: Summarize user interactions
# Create the first node to summarize user interactions:
prompt_event = Nstream.event()
ns_node_1_prompt_text = f"Generate a general insight using input user's data - {prompt_event}"
# Configure the node with its necessary components
ns_node_1 = NsNode(
node_name="GraphNode1",
prompt=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
prompt_text=ns_node_1_prompt_text
),
neuron=NsNeuron(NstreamLLM.mistral_7b()),
socket=conn)
logger.info("GraphNode1 configured")
# Configure the sink for the graph output, typically a terminal or logger
ns_graph_sink = NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Sink).terminal(),
)
logger.info("Graph sink configured")
# Initialize the graph, start execution, and submit it for processing
ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)
logger.info("Graph execution started")
# Terminate the graph execution after a predefined runtime
ns_graph.terminate(run_time=6)
logger.info("Graph execution terminated")
# Signal completion of the main execution
logger.info("Main execution completed")
```
### Using the Nstream SDK for Context transformation Usage Example
In this section, we demonstrate a use case of the Nstream SDK for transforming user data
into a structured format for analysis and insights.
```python
from nstreamai_nstream_ai.core.nsinit import NsInit
from nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream
from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM
from nstreamai_nstream_ai.core.nsgraph import NsGraph
from utils.logger import logger
import sys
if __name__ == "__main__":
try:
logger.info("Starting main execution")
conn = NsInit(
api_key="your_api_key",
username="your_username",
password="your_password").connect()
logger.info("Connected to NsInit")
except Exception as e:
logger.exception("Exception occurred while initializing NsInit")
print(e)
sys.exit()
# Node 1: User Data Transformation
# Context Transformation Prompt:
ns_node_1_context_transform_prompt_text = "Transform unstructured user data into a detailed JSON format that captures comprehensive user interactions, particularly with promotional activities, and includes other vital user data points like geolocation data, data source, and timestamps."
ns_node_1 = NsNode(
node_name="GraphNode1",
context=NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Source).mongodb(),
context_transform_prompt_text=ns_node_1_context_transform_prompt_text),
neuron=NsNeuron(NstreamLLM.mistral_7b()),
socket=conn)
logger.info("GraphNode1 configured")
ns_graph_sink = NsLink(
socket=conn,
provider=NsProvider(type=NsProviderType().Sink).terminal(),
)
logger.info("Graph sink configured")
ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)
logger.info("Graph execution started")
ns_graph.terminate(run_time=6)
logger.info("Graph execution terminated")
logger.info("Main execution completed")
```
Raw data
{
"_id": null,
"home_page": "https://github.com/nstream-ai/ms-public-sdk",
"name": "nstreamai",
"maintainer": null,
"docs_url": null,
"requires_python": "<4,>=3.8",
"maintainer_email": null,
"keywords": "nstreamai, streaming, rag, analytics, realtime",
"author": "Nstream AI",
"author_email": "hello@nstream.ai",
"download_url": "https://files.pythonhosted.org/packages/b6/91/5ff72b5ec637efff2ad9048b659355f0be4dfbe504affd0d7cf3d2e4e17b/nstreamai-0.0.14.tar.gz",
"platform": null,
"description": "# ms-public-sdk\n![Nstream SDK](src/images/mermaid-diagram.svg \"Nstream SDK\")\n\n## Quick Start Guide\n### Using the Nstream SDK for Recommending Promotions\n\n### Introduction\nThe Nstream SDK is designed for building and managing neural network pipelines that incorporate large language models (LLMs) and stream processing capabilities. This guide will walk you through creating a simple but powerful neural network pipeline for generating insights from streaming data using the Nstream SDK.\n\n### Installation\n\nTo begin using the Nstream SDK in your projects, install it via pip with the following command:\n\n```bash\npip install nstream\n```\n\nAfter installation, import the necessary modules and classes as follows:\n\n```python\nfrom nstreamai_nstream_ai.core.nsinit import NsInit\nfrom nstreamai_nstream_ai.core.nsnode import (\n NsNode, NsLink, NsProvider, NsProviderType, Nstream\n)\nfrom nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM\nfrom nstreamai_nstream_ai.core.nsgraph import NsGraph\nfrom utils.logger import logger\nimport sys\n```\n## Core Module Imports\n\n### NsInit\n**Import**: `from nstreamai_nstream_ai.core.nsinit import NsInit`\n\n**Description**: This import brings in the `NsInit` class responsible for initializing and configuring the connection settings for the Nstream SDK.\n\n**Usage**: It is used to set up the initial connection with the Nstream API using credentials like API key, username, and password.\n\n### NsNode, NsLink, NsProvider, NsProviderType, Nstream\n**Import**: `from nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream`\n\n**Description**: This import includes multiple classes that are fundamental to constructing nodes within an Nstream pipeline:\n- **NsNode**: Represents a single node in the neural network graph.\n- **NsLink**: Used to define connections between nodes and data sources or sinks.\n- **NsProvider**: Specifies the data source or sink type.\n- **NsProviderType**: Enumerates different types of data providers.\n- **Nstream**: Might be used for additional functionality directly related to stream operations.\n\n**Usage**: These are used to define and link the functional components of a neural network pipeline, such as data inputs, transformations, and outputs.\n\n### NsNeuron, NstreamLLM\n**Import**: `from nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM`\n\n**Description**: This import involves classes related to the neural network computation units within the SDK:\n- **NsNeuron**: Represents a neuron unit which can execute specific neural network computations or models.\n- **NstreamLLM**: Pertains to specific large language model configurations that can be deployed within neurons.\n\n**Usage**: These are used to specify and configure the large language models that perform the actual analytics and insights generation in the pipeline.\n\n### NsGraph\n**Import**: `from nstreamai_nstream_ai.core.nsgraph import NsGraph`\n\n**Description**: Imports the `NsGraph` class, which manages the execution flow of neural network nodes defined with `NsNode`.\n\n**Usage**: This class is crucial for defining the execution order and dependencies between nodes, as well as starting and stopping the data processing workflow.\n\n### Configuration\nStart by initializing the SDK and connecting to the Nstream service:\n\n```python\ntry:\n logger.info(\"Starting main execution\")\n conn = NsInit(\n api_key=\"your_api_key\", \n username=\"your_username\", \n password=\"your_password\").connect()\n logger.info(\"Connected to NsInit\")\nexcept Exception as e:\n logger.exception(\n \"Exception occurred while initializing NsInit\"\n )\n print(e)\n sys.exit()\n```\n\n### Building the Pipeline\n#### Node 1: User Interaction Summarization and Insight Generation\nCreate the first node to summarize user interactions and generate insights:\n\n```python\nprompt_event = Nstream.event()\nns_node_1_prompt_text = f\"Generate a general insight using a user's data - {prompt_event}\"\nns_node_1_context_transform_prompt_text = \"Transform unstructured user data into a detailed JSON format...\"\n\nns_node_1 = NsNode(\n node_name=\"GraphNode1\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n prompt_text=ns_node_1_prompt_text\n ),\n context=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n context_transform_prompt_text=ns_node_1_context_transform_prompt_text),\n neuron=NsNeuron(NstreamLLM.mistral_7b()),\n socket=conn)\nlogger.info(\"GraphNode1 configured\")\n```\n\n#### Node 2: Personalized Recommendations\nSet up the second node to use insights from Node 1 to generate personalized recommendations:\n\n```python\nns_node_2_prompt_text = \"Based on general insight, provide a list of recommended promotional offers\"\nns_node_2_context_transform_prompt_text = \"Transform unstructured promotional data into a structured JSON format...\"\n\nns_node_2 = NsNode(\n node_name=\"GraphNode2\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(\n type=NsProviderType().Source).mongodb(),prompt_text=ns_node_2_prompt_text),\n context=ns_node_1.output(\n context_transform_prompt_text=ns_node_2_context_transform_prompt_text),\n neuron=NsNeuron(NstreamLLM.llama2_7b()),\n socket=conn)\nlogger.info(\"GraphNode2 configured\")\n```\n#### Running the Pipeline\nExecute the configured pipeline and handle the output:\n```python\nns_graph_sink = NsLink(socket=conn, provider=NsProvider(type=NsProviderType().Sink).terminal())\nlogger.info(\"Graph sink configured\")\n\nns_graph = NsGraph(conn).start(ns_node_1).end(ns_node_2).submit(ns_graph_sink)\nlogger.info(\"Graph execution started\")\n\nns_graph.terminate(run_time=6)\nlogger.info(\"Graph execution terminated\")\n\nprint(\"Execution Completed\")\nlogger.info(\"Main execution completed\")\n```\n\nFinal code would look like this:\n```python\nfrom nstreamai_nstream_ai.core.nsinit import NsInit\nfrom nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream\nfrom nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM\nfrom nstreamai_nstream_ai.core.nsgraph import NsGraph\nfrom utils.logger import logger\nimport sys\n\nif __name__ == \"__main__\":\n try:\n logger.info(\"Starting main execution\")\n conn = NsInit(\n api_key=\"your_api_key\", \n username=\"your_username\", \n password=\"your_password\").connect()\n logger.info(\"Connected to NsInit\")\n except Exception as e:\n logger.exception(\"Exception occurred while initializing NsInit\")\n print(e)\n sys.exit()\n\n # Node 1: User Interaction Summarization and Insight Generation\n # Objective: Generate Insight using the raw cookie data and historical summarizes of the user\u2019s interactions with promotions on a website.\n prompt_event = Nstream.event()\n # Prompt for insight generation\n ns_node_1_prompt_text = f\"Generate a general insight using a input user's data - {prompt_event}\" # Uses stored structured summary from db\n \n # Context Transformation Prompt:\n ns_node_1_context_transform_prompt_text = \"Transform unstructured user data into a detailed JSON format that captures comprehensive user interactions, particularly with promotional activities, and includes other vital user data points like geolocation data, data source, and timestamps.\"\n\n ns_node_1 = NsNode(\n node_name=\"GraphNode1\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n prompt_text=ns_node_1_prompt_text\n ),\n context=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n context_transform_prompt_text=ns_node_1_context_transform_prompt_text),\n neuron=NsNeuron(NstreamLLM.mistral_7b()),\n socket=conn)\n logger.info(\"GraphNode1 configured\")\n\n # Node 2: Personalized Recommendations\n # Objective: Use the insights from Node1 to generate personalized product recommendations based on the user's interactions with promotions.\n \n # Prompt for Product Listing\n ns_node_2_prompt_text = \"Based on general insight, provide a list the recommended promotional offers\"\n \n # Context Transformation Prompt\n ns_node_2_context_transform_prompt_text = \"Transform unstructured promotional data into a structured JSON format for better recommendation accuracy. Include essential details like offer ID, title, discount, and audience criteria. Ensure all offers are consistently formatted to enhance analysis and comparison.\"\n\n ns_node_2 = NsNode(\n node_name=\"GraphNode2\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n prompt_text=ns_node_2_prompt_text,\n ),\n context=ns_node_1.output(\n context_transform_prompt_text=ns_node_2_context_transform_prompt_text),\n neuron=NsNeuron(NstreamLLM.llama2_7b()),\n socket=conn)\n logger.info(\"GraphNode2 configured\")\n\n ns_graph_sink = NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Sink).terminal(),\n )\n logger.info(\"Graph sink configured\")\n\n ns_graph = NsGraph(conn).start(ns_node_1).end(ns_node_2).submit(ns_graph_sink)\n logger.info(\"Graph execution started\")\n\n ns_graph.terminate(run_time=6)\n logger.info(\"Graph execution terminated\")\n\n logger.info(\"Main execution completed\")\n```\n\n##\n### Using the Nstream SDK for Genaration Usage Example\n\n#### Building the Pipeline\n##### Node 1: User Interaction Summarization\nCreate the first node to summarize user interactions:\n\n```python\n prompt_event = Nstream.event()\n ns_node_1_prompt_text = f\"Generate a general insight using input user's data - {prompt_event}\"\n\n # Configure the node with its necessary components\n ns_node_1 = NsNode(\n node_name=\"GraphNode1\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n prompt_text=ns_node_1_prompt_text\n ),\n neuron=NsNeuron(NstreamLLM.mistral_7b()),\n socket=conn)\n logger.info(\"GraphNode1 configured\")\n```\n\n#### Running the Pipeline\nExecute the configured pipeline and handle the output:\n```python\n ns_graph_sink = NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Sink).terminal(),\n )\n logger.info(\"Graph sink configured\")\n\n # Initialize the graph, start execution, and submit it for processing\n ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)\n logger.info(\"Graph execution started\")\n\n # Terminate the graph execution after a predefined runtime\n ns_graph.terminate(run_time=6)\n logger.info(\"Graph execution terminated\")\n\n # Signal completion of the main execution\n logger.info(\"Main execution completed\")\n```\n\nFinal Code:\n```python\n\n# Import necessary modules from the Nstream SDK and utility libraries\nfrom nstreamai_nstream_ai.core.nsinit import NsInit\nfrom nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream\nfrom nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM\nfrom nstreamai_nstream_ai.core.nsgraph import NsGraph\nfrom utils.logger import logger\nimport sys\n\nif __name__ == \"__main__\":\n try:\n # Initialize logging and connect to the Nstream service\n logger.info(\"Starting main execution\")\n conn = NsInit(\n api_key=\"your_api_key\", \n username=\"your_username\", \n password=\"your_password\").connect()\n logger.info(\"Connected to NsInit\")\n except Exception as e:\n # Log any exceptions during initialization and exit\n logger.exception(\"Exception occurred while initializing NsInit\")\n print(e)\n sys.exit()\n\n # Define Node 1: Summarize user interactions\n # Create the first node to summarize user interactions:\n prompt_event = Nstream.event()\n ns_node_1_prompt_text = f\"Generate a general insight using input user's data - {prompt_event}\"\n\n # Configure the node with its necessary components\n ns_node_1 = NsNode(\n node_name=\"GraphNode1\",\n prompt=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n prompt_text=ns_node_1_prompt_text\n ),\n neuron=NsNeuron(NstreamLLM.mistral_7b()),\n socket=conn)\n logger.info(\"GraphNode1 configured\")\n\n # Configure the sink for the graph output, typically a terminal or logger\n ns_graph_sink = NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Sink).terminal(),\n )\n logger.info(\"Graph sink configured\")\n\n # Initialize the graph, start execution, and submit it for processing\n ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)\n logger.info(\"Graph execution started\")\n\n # Terminate the graph execution after a predefined runtime\n ns_graph.terminate(run_time=6)\n logger.info(\"Graph execution terminated\")\n\n # Signal completion of the main execution\n logger.info(\"Main execution completed\")\n```\n\n### Using the Nstream SDK for Context transformation Usage Example\nIn this section, we demonstrate a use case of the Nstream SDK for transforming user data \ninto a structured format for analysis and insights.\n\n```python\nfrom nstreamai_nstream_ai.core.nsinit import NsInit\nfrom nstreamai_nstream_ai.core.nsnode import NsNode, NsLink, NsProvider, NsProviderType, Nstream\nfrom nstreamai_nstream_ai.core.nsneuron import NsNeuron, NstreamLLM\nfrom nstreamai_nstream_ai.core.nsgraph import NsGraph\nfrom utils.logger import logger\nimport sys\n\nif __name__ == \"__main__\":\n try:\n logger.info(\"Starting main execution\")\n conn = NsInit(\n api_key=\"your_api_key\", \n username=\"your_username\", \n password=\"your_password\").connect()\n logger.info(\"Connected to NsInit\")\n except Exception as e:\n logger.exception(\"Exception occurred while initializing NsInit\")\n print(e)\n sys.exit()\n\n # Node 1: User Data Transformation\n # Context Transformation Prompt:\n ns_node_1_context_transform_prompt_text = \"Transform unstructured user data into a detailed JSON format that captures comprehensive user interactions, particularly with promotional activities, and includes other vital user data points like geolocation data, data source, and timestamps.\"\n\n ns_node_1 = NsNode(\n node_name=\"GraphNode1\",\n context=NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Source).mongodb(),\n context_transform_prompt_text=ns_node_1_context_transform_prompt_text),\n neuron=NsNeuron(NstreamLLM.mistral_7b()),\n socket=conn)\n logger.info(\"GraphNode1 configured\")\n\n ns_graph_sink = NsLink(\n socket=conn,\n provider=NsProvider(type=NsProviderType().Sink).terminal(),\n )\n logger.info(\"Graph sink configured\")\n\n ns_graph = NsGraph(conn).start(ns_node_1).submit(ns_graph_sink)\n logger.info(\"Graph execution started\")\n\n ns_graph.terminate(run_time=6)\n logger.info(\"Graph execution terminated\")\n\n logger.info(\"Main execution completed\")\n```\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Official SDK for nstream ai stream processing powered by Gen AI",
"version": "0.0.14",
"project_urls": {
"Homepage": "https://github.com/nstream-ai/ms-public-sdk"
},
"split_keywords": [
"nstreamai",
" streaming",
" rag",
" analytics",
" realtime"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "13a35d770b4face0cd5d6e8447f132cb0c02087b5d4f84710eb542fea4d5d3ca",
"md5": "eff2bb9d4afb71779a4d8939fcd85291",
"sha256": "923738c0b092471c624fa3cb6c78376d8a5229fba725d3d0859baaefa4de4deb"
},
"downloads": -1,
"filename": "nstreamai-0.0.14-py3-none-any.whl",
"has_sig": false,
"md5_digest": "eff2bb9d4afb71779a4d8939fcd85291",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4,>=3.8",
"size": 27584,
"upload_time": "2024-04-18T06:19:55",
"upload_time_iso_8601": "2024-04-18T06:19:55.545894Z",
"url": "https://files.pythonhosted.org/packages/13/a3/5d770b4face0cd5d6e8447f132cb0c02087b5d4f84710eb542fea4d5d3ca/nstreamai-0.0.14-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "b6915ff72b5ec637efff2ad9048b659355f0be4dfbe504affd0d7cf3d2e4e17b",
"md5": "f54b5d527f4f3522c21557618e0b9765",
"sha256": "8e9a747b7f866cfc3c0145839c0be3c12a56366e5e581eed5b7c150989ca27b3"
},
"downloads": -1,
"filename": "nstreamai-0.0.14.tar.gz",
"has_sig": false,
"md5_digest": "f54b5d527f4f3522c21557618e0b9765",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4,>=3.8",
"size": 25196,
"upload_time": "2024-04-18T06:19:57",
"upload_time_iso_8601": "2024-04-18T06:19:57.183627Z",
"url": "https://files.pythonhosted.org/packages/b6/91/5ff72b5ec637efff2ad9048b659355f0be4dfbe504affd0d7cf3d2e4e17b/nstreamai-0.0.14.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-04-18 06:19:57",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "nstream-ai",
"github_project": "ms-public-sdk",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "nstreamai"
}
JSON
