incorporator 1.0.3

The Dynamic Class Building and Zero-Boilerplate Universal Data Gateway.

🌌 Incorporator (v1.0.3)

The Dynamic Class Building and Zero-Boilerplate Universal Data Gateway.

Stop writing boilerplate models, manual HTTP connection loops, pagination state-trackers, and fragile data-cleaning lambda functions.

Incorporator is an elite Python framework that transforms raw JSON, CSV, and XML APIs into fully typed, relational Python Object Graphs in a single line of code. Trade away pages of unrelated code for an easy prebuilt engine.

This is a framework that handles dynamic Pydantic metaprogramming, graph relational mapping, asynchronous connection pooling, and declarative ETL in 27KB.

🚀 Installation

pip install incorporator

⚡ The "Zero-Boilerplate" Philosophy

The Old Way: Define a rigid BaseModel, write an httpx loop, handle 429 retries, write a custom paginator, manually link foreign keys, catch KeyErrors, and hope the API schema doesn't change.

The Incorporator Way:

class Crypto(Incorporator): pass

async def crypto_coins():
    coins = await Crypto.incorp(
        inc_url="https://api.coingecko.com/api/v3/coins/markets?vs_currency=usd",
        inc_code="id",
        inc_name="name",
    )

    # Returns Dynamically created class and dictionary
    bitcoin = coins.inc_dict['bitcoin']
    print(bitcoin.circulating_supply)       # 20021206.0
    print(bitcoin.current_price)            # 78321

asyncio.run(crypto_coins())

---

🛠️ The Core Architectural Pillars

1. The Holy Trinity API & Dynamic Registries

• incorp(): Extracts raw data, compiles dynamic Pydantic schemas natively, and loads data into intelligent IncorporatorList wrappers.
• refresh(): Hydrates existing instances seamlessly with new data (perfect for live feeds).
• export(): Dumps stateful object graphs back into sanitized JSON, XML, or CSV files.
• The inc_dict: Every object automatically registers itself into a memory-safe WeakValueDictionary. Look up any object instantly: coins.inc_dict.get('bitcoin').

2. Declarative ETL & Null-Safe Converters

Data is messy. Incorporator's built-in conv_dict tools intercept bad data before Pydantic validation, shielding you from crashes with beautiful, readable syntax.

• inc(type): Automatically ranks fallbacks. inc(datetime) will parse ISO-8601 or 10+ standard string formats natively.
• calc(func, *keys): Multi-column row calculations. calc(len, 'residents', default=0).
• link_to & link_to_list: Zero-boilerplate Graph Relational Mapping.

3. Native Concurrency & Invisible Resilience

Pass a list of 500 URLs or trigger a deep-drill. Incorporator automatically spins up an asyncio.Semaphore, shares a single httpx.AsyncClient pool, and batches requests. Hit a 429 Too Many Requests? It automatically jitter-retries via tenacity. Still 429? It gracefully skips the failed row, logs it to results.failed_sources, and returns the remaining objects without crashing your pipeline.

4. Advanced Asynchronous Pagination

Isolated OOP strategies to gracefully handle pagination without infinite loops. Includes NextUrlPaginator, CursorPaginator, OffsetPaginator, PageNumberPaginator, and LinkHeaderPaginator.

---

📖 Real-World Showcases

Showcase 1: HATEOAS & Relational Mapping (Star Wars API)

Turn disconnected flat APIs into deeply nested, traversable object graphs using link_to and link_to_list.

class Planet(Incorporator): pass

class Film(Incorporator): pass

class Person(Incorporator): pass

async def far_far_away():
    BASE_URL = "https://swapi.dev/api"
    # 1. Build the foundational Graph Nodes
    planets = await Planet.incorp(
        inc_url=f"{BASE_URL}/planets/", rec_path="results",
        inc_code="id", inc_name="name",
        inc_page=NextUrlPaginator("next"), ignore_ssl=True,
        conv_dict={"url": extract_url_id(int)},
        name_chg=[("url", "id")]
    )

    films = await Film.incorp(
        inc_url=f"{BASE_URL}/films/", rec_path="results",
        inc_code="id", inc_name="title",
        inc_page=NextUrlPaginator("next"), ignore_ssl=True,
        conv_dict={"url": extract_url_id(int)},
        name_chg=[("url", "id")]
    )

    # 2. Fetch People and map relations natively
    people = await Person.incorp(
        inc_url=f"{BASE_URL}/people/", rec_path="results",
        inc_code="id", inc_name="name",
        inc_page=NextUrlPaginator("next"), ignore_ssl=True,
        conv_dict={
            "url": extract_url_id(int),
            "homeworld": calc(link_to(planets, extractor=extract_url_id(int)), default=None),
            "films": calc(link_to_list(films, extractor=extract_url_id(int)), default=[])
        },
        name_chg=[("url", "id")]
    )

    # Yoda, you seek yoda with a unique key, readable name, and last known whereabouts.
    for person in people[17:22]:
        person.display()            #<class, inc_code (key), inc_name, lact_rcd>
    print('\n')

    # Find Boba, I'd say you have with at O(1) speed with graph mapping already built.
    boba_fett = people.inc_dict[22]
    print(boba_fett.homeworld.inc_name)  # "Kamino"
    print(boba_fett.films[0].inc_name)  # "The Empire Strikes Back"

asyncio.run(far_far_away())

Showcase 2: Parent-Based Enrichment (PokéAPI)

Pass shallow objects into inc_parent to trigger automatic concurrent bulk detail scraping.

class Nav(Incorporator): pass

class Pokemon(Incorporator): pass

async def inc_pokedex():
    BASE_URL = "https://pokeapi.co/api/v2"

    # 1. SHALLOW DISCOVERY: Fetch 150 navigation URLs
    pokemon_nav = await Nav.incorp(
        inc_url=f"{BASE_URL}/pokemon/?limit=50&offset=0",
        rec_path="results",
        inc_name="name",
        name_chg=[('url', 'detail_url')],
        inc_page=NextUrlPaginator("next"),
        call_lim=3  # 3 pages * 50 = 150 Pokemon
    )

    def calculate_bst(stats: list) -> int:
        return sum(s.get("base_stat", 0) for s in stats if isinstance(s, dict))

    # 2. DEEP ENRICHMENT: Pass the parent objects. The framework tears out 'detail_url',
    # fires 150 concurrent requests, and builds deep objects automatically.
    enriched_pokemon = await Pokemon.incorp(
        inc_parent=pokemon_nav,
        inc_code="id",
        inc_name="name",
        excl_lst=["sprites", "moves", "game_indices", "held_items"]
    )

    # Query a parent-child pattern while building it for deep analysis.
    for pokemon in enriched_pokemon[:3]:
        print(pokemon.inc_name, pokemon.abilities[0].ability.name)

asyncio.run(inc_pokedex())

Showcase 3: Local XML to Live JSON Bulk POST (NHTSA API)

Seamlessly bridge deep local XML data with live JSON REST APIs.

class JimmyInvoice(Incorporator): pass

class NHTSARecord(Incorporator): pass

async def audit_jimmys():
    # 1. Extract nested data from a local XML file into the same Pydantic engine.
    invoices = await JimmyInvoice.incorp(
        inc_file="shady_jimmy.xml",
        rec_path="Dealership.AuditFile.Invoices.Invoice"
    )

    # Easy dot notation for one-line on your way to your next one line API call
    vin_batch_string = ";".join([getattr(inv.Vehicle, "VIN", "") for inv in invoices])

    # 2. Hit a live JSON Bulk Endpoint using a POST payload using the same syntax.
    live_records = await NHTSARecord.incorp(
        inc_url="https://vpic.nhtsa.dot.gov/api/vehicles/DecodeVINValuesBatch/",
        method="POST",
        form_payload={"format": "json", "DATA": vin_batch_string},
        rec_path="Results",
        inc_code="VIN",
        conv_dict={"ModelYear": inc(int)}  # Force string years to integers
    )

    # 3. Audit instantly via the memory-safe registry
    for inv in invoices:
        vin = inv.Vehicle.VIN
        actual_car = live_records.inc_dict.get(vin)
        if actual_car.ModelYear != int(inv.Vehicle.Year):
            print("Fraud Detected!", inv.inc_code, inv.Vehicle.Model)

asyncio.run(audit_jimmys())

---

🕵️ Non-Blocking Observability

Need production logs without starving your async event loop?

from incorporator import LoggedIncorporator

class WebAPI(LoggedIncorporator): pass

# Configures background multithreaded queue logging automatically
instance = await WebAPI.incorp(
    inc_url="https://api.example.com/data",
    enable_logging=True
)

instance.log_info("Standard trace")
instance.log_error("API Offline", exc_info=True)
instance.log_api("Web traffic trace") # Routes to isolated api.log

🤝 Contributing

1. Let's go!

Built for data engineers who want to sleep at night.