masa-ai 0.2.4

MASA SDK - Masa's AI Software Architecture

Masa AI Software Architecture

MASA is a project for data retrieval, quality control, and orchestration. It currently provides tools to retrieve data from Twitter using the Masa Protocol Node API, with plans to expand to other data sources and functionalities in the future.

Note: This SDK requires a Masa Protocol Node to be running on the system. Instructions on how to install and run a node can be found here.

Quick Start

1. Install the MASA package:

   pip install masa-ai
   

   If you encounter issues running or installing masa-ai, please refer to the System Requirements section to ensure you have the necessary system dependencies installed.

2. Create a request_list.json file with the queries you'd like to process. This file can be placed anywhere on your system. Here is an example of what the request_list.json might look like:

   [
       {
           "scraper": "XTwitterScraper",
           "endpoint": "data/twitter/tweets/recent",
           "priority": 1,
           "params": {
               "query": "#example",
               "max_results": 100
           }
       },
       {
           "scraper": "XTwitterScraper",
           "endpoint": "data/twitter/tweets/recent",
           "priority": 2,
           "params": {
               "query": "from:example_user",
               "max_results": 50
           }
       }
   ]
   

   Note: max_results can be no greater than 450. To be safe, set it slightly lower than this limit to avoid exceeding the rate limit.

   An example request_list.json file is included in the package. You can find it in the examples folder at the following path:

   EXAMPLE_PATH=$(pip show masa-ai | grep Location | awk '{print $2"/masa_ai/examples/request_list.json"}')
   echo "Example request_list.json path: $EXAMPLE_PATH"
   

3. Use the MASA CLI:

   masa-ai-cli <command> [options]
   

   Available commands:

   • process [path_to_requests_json]: Process all requests (both resumed and new).
   • docs [page_name]: Rebuild and view the documentation for the specified page (page_name is optional).
   • data: List the scraped data files.
   • list-requests [--statuses STATUS_LIST]: List requests filtered by statuses.
   • clear-requests [REQUEST_IDS]: Clear queued or in-progress requests by IDs.

   Examples:

   # Process requests from a JSON file
   masa-ai-cli process /path/to/request_list.json
   
   # View the usage documentation
   masa-ai-cli docs usage
   
   # List scraped data files
   masa-ai-cli data
   
   # List queued and in-progress requests
   masa-ai-cli list-requests
   
   # List requests with specific statuses
   masa-ai-cli list-requests --statuses completed,failed
   
   # Clear all queued and in-progress requests
   masa-ai-cli clear-requests
   
   # Clear specific requests by IDs
   masa-ai-cli clear-requests req1,req2,req3
   

4. Accessing Scraped Data:

   By default, the data that is scraped is saved to the current working directory under the data folder. You can designate a different directory by setting the DATA_DIRECTORY in the configuration. To list all scraped data files, use the following command:

   masa-ai-cli data
   

   This will display the structure of the data folder and list all the files contained within it.

5. Recommendations for Accessing and Using Scraped Data:

   • Command Line: You can navigate to the data folder using the command line to view and manipulate the files directly.

     IMPORTANT: The data folder is created when you run the masa-ai-cli process [path_to_requests_json] command.

     # Navigate to the data directory
     cd /path/to/your/data_directory
     

     If you have set a custom DATA_DIRECTORY in your configuration, replace /path/to/your/data_directory with the path you have designated. You can use this path to access data for further processing, analysis, and utilization with agents.

6. For detailed usage instructions, please refer to the Usage Guide.

Managing Requests

The MASA CLI now provides commands to manage your data retrieval requests more effectively.

Listing Requests

You can list the current requests that are queued or in progress:

masa-ai-cli list-requests

By default, this command lists requests with statuses queued and in_progress. You can specify other statuses using the --statuses option:

masa-ai-cli list-requests --statuses completed,failed

To list all requests regardless of their status:

masa-ai-cli list-requests --statuses all

Clearing Requests

To clear all requests that are queued or in progress:

masa-ai-cli clear-requests

To clear specific requests by their IDs:

masa-ai-cli clear-requests req1,req2,req3

Requests that are cleared will have their status changed to cancelled and will not be processed.

Configuration

The project uses YAML files for configuration:

• configs/settings.yaml: Main configuration file containing settings for Twitter API, request management, and logging.
• configs/.secrets.yaml: (Optional) File for storing sensitive information like API keys.

The settings.yaml file is loaded using Dynaconf, which allows for easy environment-based configuration management.

Advanced Twitter Search

The Masa Protocol Node API provides advanced search capabilities for retrieving Twitter data. Some of the available search options include:

• Hashtag Search: #hashtag
• Mention Search: @username
• From User Search: from:username
• Keyword Exclusion: -keyword
• OR Operator: term1 OR term2
• Geo-location Based Search: geocode:latitude,longitude,radius
• Language-Specific Search: lang:language_code

For more details, refer to the Masa Protocol Twitter Docs.

Project Structure

• masa_ai/: Main package directory
  • configs/: Configuration files
  • connections/: API connection handlers
  • tools/: Core functionality modules
    • qc/: Quality control tools
    • retrieve/: Data retrieval tools
    • utils/: Utility functions
  • orchestration/: Request management and processing
  • logs/: Log files
  • data/: Scraped data
  • examples/: Example files

System Requirements {#system-requirements}

If you run into issues running or installing masa-ai, ensure you have the necessary system dependencies installed.

On Debian-based systems (e.g., Ubuntu)

Install build-essential:

sudo apt-get update
sudo apt-get install -y build-essential

On Red Hat-based systems (e.g., CentOS)

Install Development Tools:

sudo yum groupinstall 'Development Tools'

On macOS

Install Xcode Command Line Tools:

xcode-select --install

On Windows

1. Download and install the Microsoft Visual C++ Build Tools.
2. Ensure that the installation includes the "Desktop development with C++" workload.
3. Install make using Chocolatey:

choco install make

Dependencies

Key dependencies include:

• Data Processing: numpy, pandas
• API Interaction: requests
• Configuration: dynaconf, pyyaml, python-dotenv
• Quality Control: colorlog
• Progress Display: tqdm
• Documentation: sphinx, sphinx_rtd_theme, recommonmark, myst-parser
• Jupyter Notebooks: jupyter, notebook, ipykernel
• Database Interaction: psycopg2-binary
• Data Parsing: feedparser

For a full list of dependencies, refer to pyproject.toml.

Documentation

The MASA project uses Sphinx to generate its documentation. The documentation is automatically rebuilt and viewed when using the docs command with the masa-ai-cli command.

To view the documentation:

masa-ai-cli docs [page_name]

This command will rebuild and view the documentation for the specified page. Note that the [page_name] is optional. If no page name is provided, the documentation for the entire project will be displayed.

License

This project is licensed under the MIT License. See the LICENSE file for details.