prdb-sdk (Python)

Python client for the prdb Public API.

Install

pip install prdb-sdk

Requires Python 3.10 or newer. The client is async and uses httpx under the hood, by way of the Kiota bundle.

Usage

import asyncio
from uuid import UUID

from prdb_sdk import create_client


async def main() -> None:
    client = create_client(api_key="...")

    # GET /videos
    page = await client.videos.get()
    for video in page.items or []:
        print(video.title)

    # GET /videos/{id}
    video = await client.videos.by_id(UUID("...")).get()

    # Query parameters are typed, including the closed-set ones.
    from prdb_sdk.generated.videos.videos_request_builder import VideosRequestBuilder

    config = VideosRequestBuilder.VideosRequestBuilderGetRequestConfiguration(
        query_parameters=VideosRequestBuilder.VideosRequestBuilderGetQueryParameters(
            page=2,
            page_size=50,
            search="...",
        )
    )
    page_two = await client.videos.get(request_configuration=config)


asyncio.run(main())

The request builders mirror the API's URL structure, so GET /videos/{id}/filehashes is client.videos.by_id(video_id).filehashes.get().

Authentication

create_client sends the key in the X-Api-Key header, and keeps it on the API host: a redirect to a different origin raises CrossOriginRedirectError rather than handing your credential to whoever answers there. Redirects that stay on the same origin are followed normally.

base_url must use https, so the key is never sent in cleartext. A loopback address (localhost, 127.0.0.1 or ::1) is the exception: the request never leaves the machine, so plain http is accepted there and a local stand-in for the API needs no certificate.

GET /health is the only endpoint that works without a key; use create_anonymous_client() for health probes. That one has no credential to protect, so it accepts a plain http base URL.

Options

from prdb_sdk import RetryOptions, create_client

client = create_client(
    api_key="...",
    base_url="https://api.prdb.net",   # override for a staging deployment
    http_client=my_httpx_async_client, # control timeouts, proxies, connection limits
    retry=RetryOptions.disabled(),     # see below
)

A client you pass in is left as it is. The SDK copies it and installs its middleware on the copy, so the client it sends through behaves like the one built for you — same redirect rule, same retry handling — while yours keeps behaving the way you configured it.

The copy shares your transport, so your connection pool, TLS settings and proxies are the ones actually used. That also ties the lifetimes together: closing your client closes the connections the SDK sends through.

Retrying

By default the SDK retries a 429, 503 or 504 up to three times, honouring Retry-After.

Turn that off if your application already retries prdb calls:

client = create_client("...", retry=RetryOptions.disabled())

Otherwise the two policies multiply — one logical call becomes up to n×m requests against an API that rate limits, and an outer circuit breaker never sees a stable failure to open on. The built-in policy also retries writes, so an application that must not repeat one should own the retry itself.

To keep it but change it:

client = create_client("...", retry=RetryOptions(max_retries=5, delay=1.0))

Reading the response status

A typed call returns the deserialised body but not the response status. Pass a ResponseStatusOption when the status itself matters; the conditional-request example below uses it to distinguish a 304 Not Modified response from other responses with no body.

Pass a ResponseStatusOption to read it:

from kiota_abstractions.base_request_configuration import RequestConfiguration

from prdb_sdk import ResponseStatusOption

status = ResponseStatusOption()

health = await client.health.get(
    request_configuration=RequestConfiguration(options=[status])
)

assert health.status == "healthy"
assert status.status_code == 200

Kiota's own NativeResponseHandler surfaces the raw response but suppresses deserialisation while doing so. This option keeps the typed result and records the status alongside it.

Reading the rate limit

Every metered response carries the rate limit it was counted against, so you can pace off the answers you are already getting instead of spending a request on GET /rate-limit to ask.

from kiota_abstractions.base_request_configuration import RequestConfiguration

from prdb_sdk import RateLimitOption

limits = RateLimitOption()

sites = await client.sites.get(
    request_configuration=RequestConfiguration(options=[limits])
)

if limits.hour and limits.hour.remaining < 50:
    ...  # slow down; limits.hour.reset_in_seconds until a slot frees up

hour and month are each a RateLimitWindow with limit, remaining and reset_in_seconds, or None.

reset_in_seconds is the wait until the oldest request leaves the sliding window and frees one slot — not a timestamp, and not the time until the whole window resets. It is the same quantity resetsInSeconds carries on GET /rate-limit.

None is an answer rather than a gap. A response the API did not meter — 401, 403, 503, and GET /rate-limit itself — carries no headers at all, and a 429 carries only the window that refused the request, so exactly one of the two being set is normal. A refusal records too, so the reading is there for a caller that catches the error.

Kiota can also surface response headers itself, through HeadersInspectionHandlerOption, as raw multi-valued strings. This option is the typed reading of the six that matter.

Conditional requests

GET /sites returns a weak ETag covering the matched rows and the paging, sorting and search parameters. Send it back as If-None-Match and the endpoint answers 304 Not Modified with no body while nothing has changed — the whole site list fits in one request at page_size=1000, so this is worth doing.

from kiota_abstractions.base_request_configuration import RequestConfiguration
from kiota_http.middleware.options import HeadersInspectionHandlerOption

from prdb_sdk import ResponseStatusOption

# First call: read the validator off the response.
inspect = HeadersInspectionHandlerOption(inspect_response_headers=True)
sites = await client.sites.get(
    request_configuration=RequestConfiguration(options=[inspect])
)
etag = next(iter(inspect.response_headers.try_get("etag")))

# Later: ask only for what changed.
status = ResponseStatusOption()
configuration = RequestConfiguration(options=[status])
configuration.headers.add("If-None-Match", etag)

sites = await client.sites.get(request_configuration=configuration)

if status.status_code == 304:
    ...  # nothing changed; keep the copy you already have

A 304 returns None from the typed call rather than raising. None alone does not distinguish "not modified" from "no rows", so pass a ResponseStatusOption when you need to tell them apart.

One wrinkle from the API side: the shared read-only cache does not vary by If-None-Match, so a request that hits it is answered 200 with a body even when your validator still matches. That is expected rather than an error.

Use one instance per call. It is written when the response arrives, so sharing one across concurrent calls means whichever finishes last wins.

The status recorded is the one the result was built from: after a redirect the SDK followed, and after the last retry. A call that raises records too, so an APIError caught from a 403 still has its status alongside. It stays None when no response was reached at all — a failed connection, a timeout, or a refused cross-origin redirect.

Generated code

Everything under prdb_sdk/generated/ is produced by Kiota from spec/openapi.json in the repository root and is overwritten on every regeneration. Do not edit it — see the root README.