batchwork-ai 0.2.2__tar.gz → 0.3.0__tar.gz

{batchwork_ai-0.2.2 → batchwork_ai-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: batchwork-ai
-Version: 0.2.2
+Version: 0.3.0
 Summary: Unified async batch API for AI providers
 Keywords: ai,anthropic,batch,gemini,llm,openai
 Author: Ajanraj
@@ -34,10 +34,10 @@ Description-Content-Type: text/markdown
 
 ![Batchwork — one typed API for provider-native AI batch jobs](docs/public/og-image.png)
 
-Unified async batch API for OpenAI, Anthropic, Google Gemini, Groq, Mistral, Together AI, and xAI. Providers price eligible batch requests up to 50% below standard synchronous calls, and Batchwork gives you that pricing through one typed Python interface and one CLI. It handles provider-specific serialization, submission, polling, and result parsing so you do not have to. Pricing and eligibility remain provider- and model-specific.
+Unified async batch API for OpenAI, Azure OpenAI, Anthropic, Google Gemini, Groq, Mistral, Together AI, and xAI. Providers price eligible batch requests up to 50% below standard synchronous calls, and Batchwork gives you that pricing through one typed Python interface and one CLI. It handles provider-specific serialization, submission, polling, and result parsing so you do not have to. Pricing and eligibility remain provider- and model-specific.
 
 - One typed API for provider-native batch jobs
-- Text, embedding, and image workloads
+- Text, embedding, image, image-edit, moderation, transcription, translation, and video workloads
 - Normalized jobs, results, usage, and errors
 - Messages, tools, structured content, and remote media
 - Polling, persistent stores, and signed webhooks
@@ -129,21 +129,38 @@ asyncio.run(main())
 
 Submitting returns a `BatchJob` immediately. Provider processing is asynchronous and may take minutes or, depending on the provider and workload, up to 24 hours.
 
+Beyond text, the same client covers seven more modalities — for example hosted-audio transcription:
+
+```python
+from batchwork import BatchTranscriptionRequest
+
+job = await client.batch_transcriptions(
+    model="groq/whisper-large-v3",
+    requests=[
+        BatchTranscriptionRequest(
+            custom_id="call-1",
+            audio_url="https://example.com/call-1.mp3",
+        )
+    ],
+)
+```
+
 Models use `provider/model` form. Results are correlated with requests through `custom_id`; provider output order is not guaranteed.
 
 ## Providers
 
 Output workloads supported by Batchwork:
 
-| Provider      | Text | Embeddings | Image generation |
-| ------------- | ---- | ---------- | ---------------- |
-| OpenAI        | Yes  | Yes        | Yes              |
-| Anthropic     | Yes  | No         | No               |
-| Google Gemini | Yes  | Yes        | Yes              |
-| Groq          | Yes  | No         | No               |
-| Mistral       | Yes  | Yes        | No               |
-| Together AI   | Yes  | No         | No               |
-| xAI           | Yes  | No         | Yes              |
+| Provider | Text | Embeddings | Image generation | Image edits | Moderation | Transcription | Translation | Video |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| OpenAI | Yes | Yes | Yes | Yes | Yes | No | No | No |
+| Azure OpenAI | Yes | No | No | No | No | No | No | No |
+| Anthropic | Yes | No | No | No | No | No | No | No |
+| Google Gemini | Yes | Yes | Yes | No | No | No | No | No |
+| Groq | Yes | No | No | No | No | Yes | Yes | No |
+| Mistral | Yes | Yes | No | No | Yes | Yes | No | No |
+| Together AI | Yes | No | No | No | No | Yes | Yes | No |
+| xAI | Yes | No | Yes | Yes | No | No | No | Yes |
 
 Image, PDF, text-file, and audio inputs for text requests vary separately from output modalities. Provider APIs may also impose model-specific limits. See the [provider overview](https://batchwork.ajanraj.com/docs/providers) for input support, submission transport, credentials, and restrictions.
 
@@ -165,7 +182,7 @@ The base package does not import or require `upstash-redis`.
 - [Configuration](https://batchwork.ajanraj.com/docs/configuration)
 - [Jobs](https://batchwork.ajanraj.com/docs/guides/jobs)
 - [Results](https://batchwork.ajanraj.com/docs/guides/results)
-- [Text, embeddings, and images](https://batchwork.ajanraj.com/docs/modalities/text)
+- [Text, embeddings, and other modalities](https://batchwork.ajanraj.com/docs/modalities/text)
 - [Provider overview](https://batchwork.ajanraj.com/docs/providers)
 - [Polling and webhooks](https://batchwork.ajanraj.com/docs/guides/server)
 - [Stores](https://batchwork.ajanraj.com/docs/guides/stores)
@@ -174,6 +191,10 @@ The base package does not import or require `upstash-redis`.
 - [Public API](https://batchwork.ajanraj.com/docs/api)
 - [FAQ](https://batchwork.ajanraj.com/docs/faq)
 
+## Built with Codex and GPT-5.6
+
+Batchwork was built entirely in the OpenAI Codex CLI with GPT-5.6 during [OpenAI Build Week](https://openai.devpost.com/) (July 2026) — the Python library, the `batchwork` CLI with its machine output contract and local job registry, the Agent Skill, the test suite, and the documentation site. Codex planned the design through a decision interview, implemented the seven provider adapters across three modalities, hardened security through repeated review loops, verified everything against live provider APIs, and drove the PyPI and documentation releases. The full collaboration story, plus a self-contained demo project for judges, lives in [`hackathon/`](hackathon/README.md).
+
 ## License
 
 [MIT](https://opensource.org/licenses/MIT) © [Ajanraj](https://github.com/ajanraj)

{batchwork_ai-0.2.2 → batchwork_ai-0.3.0}/README.md

@@ -2,10 +2,10 @@
 
 ![Batchwork — one typed API for provider-native AI batch jobs](docs/public/og-image.png)
 
-Unified async batch API for OpenAI, Anthropic, Google Gemini, Groq, Mistral, Together AI, and xAI. Providers price eligible batch requests up to 50% below standard synchronous calls, and Batchwork gives you that pricing through one typed Python interface and one CLI. It handles provider-specific serialization, submission, polling, and result parsing so you do not have to. Pricing and eligibility remain provider- and model-specific.
+Unified async batch API for OpenAI, Azure OpenAI, Anthropic, Google Gemini, Groq, Mistral, Together AI, and xAI. Providers price eligible batch requests up to 50% below standard synchronous calls, and Batchwork gives you that pricing through one typed Python interface and one CLI. It handles provider-specific serialization, submission, polling, and result parsing so you do not have to. Pricing and eligibility remain provider- and model-specific.
 
 - One typed API for provider-native batch jobs
-- Text, embedding, and image workloads
+- Text, embedding, image, image-edit, moderation, transcription, translation, and video workloads
 - Normalized jobs, results, usage, and errors
 - Messages, tools, structured content, and remote media
 - Polling, persistent stores, and signed webhooks
@@ -97,21 +97,38 @@ asyncio.run(main())
 
 Submitting returns a `BatchJob` immediately. Provider processing is asynchronous and may take minutes or, depending on the provider and workload, up to 24 hours.
 
+Beyond text, the same client covers seven more modalities — for example hosted-audio transcription:
+
+```python
+from batchwork import BatchTranscriptionRequest
+
+job = await client.batch_transcriptions(
+    model="groq/whisper-large-v3",
+    requests=[
+        BatchTranscriptionRequest(
+            custom_id="call-1",
+            audio_url="https://example.com/call-1.mp3",
+        )
+    ],
+)
+```
+
 Models use `provider/model` form. Results are correlated with requests through `custom_id`; provider output order is not guaranteed.
 
 ## Providers
 
 Output workloads supported by Batchwork:
 
-| Provider      | Text | Embeddings | Image generation |
-| ------------- | ---- | ---------- | ---------------- |
-| OpenAI        | Yes  | Yes        | Yes              |
-| Anthropic     | Yes  | No         | No               |
-| Google Gemini | Yes  | Yes        | Yes              |
-| Groq          | Yes  | No         | No               |
-| Mistral       | Yes  | Yes        | No               |
-| Together AI   | Yes  | No         | No               |
-| xAI           | Yes  | No         | Yes              |
+| Provider | Text | Embeddings | Image generation | Image edits | Moderation | Transcription | Translation | Video |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| OpenAI | Yes | Yes | Yes | Yes | Yes | No | No | No |
+| Azure OpenAI | Yes | No | No | No | No | No | No | No |
+| Anthropic | Yes | No | No | No | No | No | No | No |
+| Google Gemini | Yes | Yes | Yes | No | No | No | No | No |
+| Groq | Yes | No | No | No | No | Yes | Yes | No |
+| Mistral | Yes | Yes | No | No | Yes | Yes | No | No |
+| Together AI | Yes | No | No | No | No | Yes | Yes | No |
+| xAI | Yes | No | Yes | Yes | No | No | No | Yes |
 
 Image, PDF, text-file, and audio inputs for text requests vary separately from output modalities. Provider APIs may also impose model-specific limits. See the [provider overview](https://batchwork.ajanraj.com/docs/providers) for input support, submission transport, credentials, and restrictions.
 
@@ -133,7 +150,7 @@ The base package does not import or require `upstash-redis`.
 - [Configuration](https://batchwork.ajanraj.com/docs/configuration)
 - [Jobs](https://batchwork.ajanraj.com/docs/guides/jobs)
 - [Results](https://batchwork.ajanraj.com/docs/guides/results)
-- [Text, embeddings, and images](https://batchwork.ajanraj.com/docs/modalities/text)
+- [Text, embeddings, and other modalities](https://batchwork.ajanraj.com/docs/modalities/text)
 - [Provider overview](https://batchwork.ajanraj.com/docs/providers)
 - [Polling and webhooks](https://batchwork.ajanraj.com/docs/guides/server)
 - [Stores](https://batchwork.ajanraj.com/docs/guides/stores)
@@ -142,6 +159,10 @@ The base package does not import or require `upstash-redis`.
 - [Public API](https://batchwork.ajanraj.com/docs/api)
 - [FAQ](https://batchwork.ajanraj.com/docs/faq)
 
+## Built with Codex and GPT-5.6
+
+Batchwork was built entirely in the OpenAI Codex CLI with GPT-5.6 during [OpenAI Build Week](https://openai.devpost.com/) (July 2026) — the Python library, the `batchwork` CLI with its machine output contract and local job registry, the Agent Skill, the test suite, and the documentation site. Codex planned the design through a decision interview, implemented the seven provider adapters across three modalities, hardened security through repeated review loops, verified everything against live provider APIs, and drove the PyPI and documentation releases. The full collaboration story, plus a self-contained demo project for judges, lives in [`hackathon/`](hackathon/README.md).
+
 ## License
 
 [MIT](https://opensource.org/licenses/MIT) © [Ajanraj](https://github.com/ajanraj)

batchwork_ai-0.3.0/pyproject.toml

@@ -0,0 +1,98 @@
+[build-system]
+requires = ["uv_build>=0.11.29,<0.12"]
+build-backend = "uv_build"
+
+[project]
+name = "batchwork-ai"
+version = "0.3.0"
+description = "Unified async batch API for AI providers"
+readme = "README.md"
+requires-python = ">=3.11"
+license = "MIT"
+license-files = ["LICENSE"]
+keywords = [
+    "ai",
+    "anthropic",
+    "batch",
+    "gemini",
+    "llm",
+    "openai",
+]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Framework :: AsyncIO",
+    "Intended Audience :: Developers",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+    "Typing :: Typed",
+]
+dependencies = [
+    "click>=8.4,<9",
+    "httpcore>=1,<2",
+    "httpx>=0.28,<1",
+    "pydantic>=2.11,<3",
+]
+
+[[project.authors]]
+name = "Ajanraj"
+email = "hey@ajanraj.com"
+
+[[project.maintainers]]
+name = "Ajanraj"
+email = "hey@ajanraj.com"
+
+[project.scripts]
+batchwork = "batchwork.cli:main"
+
+[project.optional-dependencies]
+redis = ["upstash-redis>=1.4,<2"]
+
+[project.urls]
+Documentation = "https://batchwork.ajanraj.com"
+Issues = "https://github.com/ajanraj/batchwork-ai/issues"
+Repository = "https://github.com/ajanraj/batchwork-ai"
+
+[dependency-groups]
+dev = [
+    "jsonschema>=4.25,<5",
+    "pytest>=8.4,<9",
+    "pytest-asyncio>=1,<2",
+    "pytest-cov>=6.2,<7",
+    "ruff>=0.12,<1",
+    "twine>=6.1,<7",
+    "ty==0.0.59",
+]
+
+[tool.uv.build-backend]
+module-name = "batchwork"
+module-root = "src"
+
+[tool.pytest.ini_options]
+addopts = "-q"
+asyncio_mode = "auto"
+testpaths = ["tests"]
+markers = [
+    "live: requires real provider credentials",
+    "redis: requires Upstash Redis credentials",
+]
+
+[tool.ruff]
+line-length = 100
+target-version = "py311"
+
+[tool.ruff.lint]
+select = [
+    "E",
+    "F",
+    "I",
+    "UP",
+    "B",
+    "ASYNC",
+    "RUF",
+]
+
+[tool.ty.src]
+include = ["src"]

batchwork_ai-0.2.2/pyproject.toml → batchwork_ai-0.3.0/pyproject.toml.orig

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "batchwork-ai"
-version = "0.2.2"
+version = "0.3.0"
 description = "Unified async batch API for AI providers"
 readme = "README.md"
 requires-python = ">=3.11"

{batchwork_ai-0.2.2 → batchwork_ai-0.3.0}/src/batchwork/__init__.py

@@ -57,8 +57,15 @@ from .types import (
     BatchEmbeddingRequestSettings,
     BatchImage,
     BatchImageDefaults,
+    BatchImageEditDefaults,
+    BatchImageEditRequest,
+    BatchImageFileRef,
+    BatchImageRef,
     BatchImageRequest,
+    BatchImageUrlRef,
     BatchLimits,
+    BatchModeration,
+    BatchModerationRequest,
     BatchProvider,
     BatchRef,
     BatchRequest,
@@ -69,7 +76,15 @@ from .types import (
     BatchResultStatus,
     BatchSnapshot,
     BatchStatus,
+    BatchTranscriptionDefaults,
+    BatchTranscriptionRequest,
+    BatchTranscriptionSegment,
+    BatchTranslationDefaults,
+    BatchTranslationRequest,
     BatchUsage,
+    BatchVideo,
+    BatchVideoDefaults,
+    BatchVideoRequest,
     ContentToolOutput,
     CustomPart,
     DeprecatedToolOutputFileDataPart,
@@ -147,9 +162,16 @@ __all__ = [
     "BatchEmbeddingRequestSettings",
     "BatchImage",
     "BatchImageDefaults",
+    "BatchImageEditDefaults",
+    "BatchImageEditRequest",
+    "BatchImageFileRef",
+    "BatchImageRef",
     "BatchImageRequest",
+    "BatchImageUrlRef",
     "BatchJob",
     "BatchLimits",
+    "BatchModeration",
+    "BatchModerationRequest",
     "BatchPoller",
     "BatchProvider",
     "BatchRef",
@@ -164,7 +186,15 @@ __all__ = [
     "BatchStatus",
     "BatchStore",
     "BatchTimeoutError",
+    "BatchTranscriptionDefaults",
+    "BatchTranscriptionRequest",
+    "BatchTranscriptionSegment",
+    "BatchTranslationDefaults",
+    "BatchTranslationRequest",
     "BatchUsage",
+    "BatchVideo",
+    "BatchVideoDefaults",
+    "BatchVideoRequest",
     "BatchWebhookEvent",
     "Batchwork",
     "BatchworkError",

batchwork_ai-0.3.0/src/batchwork/_base_url.py

@@ -0,0 +1,56 @@
+"""Provider base URL validation and normalization."""
+
+from __future__ import annotations
+
+from ipaddress import ip_address
+from urllib.parse import urlsplit, urlunsplit
+
+
+class BaseUrlError(ValueError):
+    """A provider base URL violates the outbound credential policy."""
+
+
+def _policy_error(label: str) -> BaseUrlError:
+    return BaseUrlError(
+        f"{label} must be absolute HTTPS (HTTP allowed for loopback), without "
+        "userinfo, query, or fragment."
+    )
+
+
+def normalize_base_url(value: str, label: str = "base URL") -> str:
+    """Validate a provider endpoint and remove trailing path slashes."""
+    raw_scheme, separator, remainder = value.partition("://")
+    authority = remainder
+    for delimiter in "/?#":
+        authority = authority.partition(delimiter)[0]
+    try:
+        parsed = urlsplit(value)
+        hostname = parsed.hostname
+    except ValueError as error:
+        raise _policy_error(label) from error
+    try:
+        _ = parsed.port
+    except ValueError as error:
+        raise BaseUrlError(f"{label} contains an invalid port.") from error
+
+    loopback = hostname == "localhost"
+    if hostname is not None:
+        try:
+            loopback = loopback or ip_address(hostname).is_loopback
+        except ValueError:
+            pass
+
+    if (
+        not separator
+        or raw_scheme.lower() != parsed.scheme
+        or not hostname
+        or any(character.isspace() for character in authority)
+        or parsed.username is not None
+        or parsed.password is not None
+        or "?" in value
+        or "#" in value
+        or (parsed.scheme != "https" and not (parsed.scheme == "http" and loopback))
+    ):
+        raise _policy_error(label)
+
+    return urlunsplit((parsed.scheme, parsed.netloc, parsed.path.rstrip("/"), "", ""))

{batchwork_ai-0.2.2 → batchwork_ai-0.3.0}/src/batchwork/_limits.py

@@ -16,3 +16,4 @@ MAX_AGGREGATE_RESULTS_BYTES = 1024 * 1024 * 1024
 LARGE_BATCH_REQUESTS = 10_000
 LARGE_BATCH_UPLOAD_BYTES = 50 * 1024 * 1024
 LARGE_BATCH_IMAGES = 100
+LARGE_BATCH_VIDEOS = 20

batchwork_ai-0.3.0/src/batchwork/_read_retry.py

@@ -0,0 +1,89 @@
+"""Provider-neutral retry policy for safe read operations."""
+
+from __future__ import annotations
+
+import asyncio
+import random
+import ssl
+import time
+from collections.abc import Awaitable, Callable
+from typing import TypeVar
+
+import httpx
+
+from ._provider_failure import ProviderFailureError, ProviderFailureKind
+
+_READ_ATTEMPTS = 3
+_RETRYABLE_HTTP_STATUSES = frozenset({408, 429, 500, 502, 503, 504})
+_MAX_RETRY_DELAY_SECONDS = 60.0
+_ReadResult = TypeVar("_ReadResult")
+
+
+class _ReadRetryDeadlineExceeded(TimeoutError):
+    """The local read-retry deadline expired before another attempt."""
+
+
+def _cause_chain(error: BaseException) -> list[BaseException]:
+    causes: list[BaseException] = []
+    current: BaseException | None = error
+    while current is not None and current not in causes:
+        causes.append(current)
+        current = current.__cause__ or current.__context__
+    return causes
+
+
+def _retryable_read_failure(error: BaseException) -> bool:
+    if not isinstance(error, ProviderFailureError):
+        return False
+    failure = error.failure
+    if failure.kind is ProviderFailureKind.UNAVAILABLE:
+        return failure.status_code in _RETRYABLE_HTTP_STATUSES
+    if failure.kind is not ProviderFailureKind.TRANSPORT:
+        return False
+    causes = _cause_chain(error)
+    if any(isinstance(cause, ssl.SSLError) for cause in causes):
+        return False
+    return any(
+        isinstance(
+            cause,
+            (
+                httpx.ConnectError,
+                httpx.ConnectTimeout,
+                httpx.ReadError,
+                httpx.ReadTimeout,
+            ),
+        )
+        for cause in causes
+    )
+
+
+def _retry_delay(error: BaseException, retry_index: int) -> float:
+    if isinstance(error, ProviderFailureError):
+        retry_after = error.failure.retry_after_seconds
+        if retry_after is not None:
+            return min(float(retry_after), _MAX_RETRY_DELAY_SECONDS)
+    ceiling = min(0.25 * (2**retry_index), _MAX_RETRY_DELAY_SECONDS)
+    return random.uniform(ceiling / 2, ceiling)
+
+
+async def _retry_read(
+    read: Callable[[], Awaitable[_ReadResult]],
+    *,
+    deadline: float | None = None,
+) -> _ReadResult:
+    for attempt in range(_READ_ATTEMPTS):
+        try:
+            return await read()
+        except Exception as error:
+            if attempt + 1 == _READ_ATTEMPTS or not _retryable_read_failure(error):
+                raise
+            delay = _retry_delay(error, attempt)
+            if deadline is not None:
+                remaining = deadline - time.monotonic()
+                if remaining <= 0:
+                    raise _ReadRetryDeadlineExceeded from None
+                if delay >= remaining:
+                    await asyncio.sleep(remaining)
+                    raise _ReadRetryDeadlineExceeded from None
+            await asyncio.sleep(delay)
+    raise RuntimeError("batchwork: exhausted retry loop")