Extended Data

The Problem

Every Python project reinvents the same infrastructure: YAML parsing, environment variable handling, structured logging, API client management. You copy-paste from old projects, fix edge cases again, and hope the types are right.

Extended Data is the answer. Five packages that handle the infrastructure layer of your Python applications — each independently useful, all designed to work together.

---

Core Packages

Extended Data Types

Safe, typed serialization for YAML, JSON, TOML, and Base64. String transformations, deep merge, flatten, and state utilities like is_nothing() and first_non_empty().

pip install extended-data-types

Directed Inputs

Decorator-based input handling. Load from environment variables, stdin, or config — with automatic type coercion. No inheritance required.

pip install directed-inputs-class

Lifecycle Logging

Structured logging with verbosity control, message storage, context markers, and Gunicorn integration. Rich console output included.

pip install lifecyclelogging

Vendor Connectors

Unified connectors for AWS, GCP, GitHub, Slack, Vault, Zoom, Meshy AI, Anthropic, and Cursor. Each with Direct API + LangChain Tools + MCP Server interfaces.

pip install vendor-connectors

---

See It in Action

Serialization

from extended_data_types import encode_yaml, decode_json, deep_merge

config = {"name": "myapp", "features": ["yaml", "json"]}
yaml_str = encode_yaml(config)

# Deep merge with overrides
base = {"db": {"host": "localhost", "port": 5432}}
override = {"db": {"port": 5433}, "debug": True}
merged = deep_merge(base, override)
# {"db": {"host": "localhost", "port": 5433}, "debug": True}

Input Handling

from directed_inputs_class import directed_inputs

@directed_inputs(from_env=True, env_prefix="APP_")
class Config:
    def get_settings(self, debug: bool = False, port: int = 8080):
        return {"debug": debug, "port": port}

# APP_DEBUG=true APP_PORT=3000 python app.py
config = Config()
config.get_settings()  # {"debug": True, "port": 3000}

Logging

from lifecyclelogging import Logging

logger = Logging(enable_console=True, enable_verbose_output=True)
logger.logged_statement("Server starting", log_level="info")
logger.logged_statement(
    "Request",
    json_data={"method": "POST", "path": "/api"},
    storage_marker="REQUESTS",
    log_level="debug",
)

Vendor APIs

from vendor_connectors import VendorConnectors

vc = VendorConnectors()

# Credentials loaded automatically from environment
github = vc.get_github_client(github_owner="myorg")
s3 = vc.get_aws_client("s3")
anthropic = vc.get_anthropic_client()

# Each connector also supports LangChain and MCP
from vendor_connectors.meshy.tools import get_tools
tools = get_tools()  # StructuredTools for LangChain/CrewAI

---

Architecture

All packages share a common pattern via directed-inputs-class:

                     vendor-connectors
           AWS | GCP | GitHub | Slack | Vault | AI
          ─────────────────────────────────────────
                  directed-inputs-class
             Transparent credential & config loading
          ─────────────────────────────────────────
     extended-data-types     |     lifecyclelogging
      Serialization & Utils  |    Structured Logging

Credentials are loaded automatically from environment variables, stdin, or config files. No boilerplate. No manual plumbing.

---

Why Extended Data?

Type-Safe

Full type annotations, py.typed markers, and strict mypy compliance. Your IDE autocompletes everything.

Zero Config

Sensible defaults. Install, import, use. Override when you need to.

Composable

Use one package or all five. They share patterns but don’t force coupling.

Production-Proven

CI/CD on every commit. 90%+ test coverage. Used in real production systems.

---

Install Everything

pip install extended-data-types directed-inputs-class lifecyclelogging vendor-connectors

Or install individually — each package works on its own.

---

Packages at a Glance

Package	PyPI	Python
extended-data-types		3.10+
directed-inputs-class		3.10+
lifecyclelogging		3.10+
vendor-connectors		3.10+
secretssync	Go binary	Go 1.25+

---

Getting Started Installation, setup, and your first code with Extended Data

Extended Data Types Serialization, transformations, data structures, and state utilities

Directed Inputs Decorator-based input handling with automatic type coercion

Vendor Connectors AWS, GCP, GitHub, Slack, AI APIs — with LangChain and MCP