apify-client 3.2.1b4__tar.gz → 3.2.2b1__tar.gz

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/CHANGELOG.md

@@ -3,7 +3,15 @@
 All notable changes to this project will be documented in this file.
 
 <!-- git-cliff-unreleased-start -->
-## 3.2.1 - **not yet released**
+## 3.2.2 - **not yet released**
+
+### 🚀 Features
+
+- Stream request bodies from files, iterables, and responses ([#1060](https://github.com/apify/apify-client-python/pull/1060)) ([9be215b](https://github.com/apify/apify-client-python/commit/9be215b23a13de0abc4a255a18450cc764e7a6c7)) by [@vdusek](https://github.com/vdusek), closes [#972](https://github.com/apify/apify-client-python/issues/972)
+
+
+<!-- git-cliff-unreleased-end -->
+## [3.2.1](https://github.com/apify/apify-client-python/releases/tag/v3.2.1) (2026-09-25)
 
 ### 🐛 Bug Fixes
 
@@ -12,7 +20,6 @@ All notable changes to this project will be documented in this file.
 - Make `JsonSerializable` type alias read-only ([#1072](https://github.com/apify/apify-client-python/pull/1072)) ([bd82a26](https://github.com/apify/apify-client-python/commit/bd82a2645d3c64f0e4a30de3a016c16d92d48340)) by [@Pijukatel](https://github.com/Pijukatel)
 
 
-<!-- git-cliff-unreleased-end -->
 ## [3.2.0](https://github.com/apify/apify-client-python/releases/tag/v3.2.0) (2026-09-03)
 
 ### 🚀 Features

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: apify_client
-Version: 3.2.1b4
+Version: 3.2.2b1
 Summary: Apify API client for Python
 Keywords: apify,api,client,automation,crawling,scraping
 Author: Apify Technologies s.r.o.
@@ -170,7 +170,7 @@ For a guided walkthrough — authenticating, running an Actor, and reading its r
 - **Fully typed responses** — every method returns a [Pydantic](https://docs.pydantic.dev/) model generated from the Apify OpenAPI spec, with IDE autocomplete and runtime validation ([Typed models](https://docs.apify.com/api/client/python/docs/concepts/typed-models)).
 - **Automatic retries** — exponential backoff for network errors, HTTP 429, and 5xx responses, configurable per client ([Retries](https://docs.apify.com/api/client/python/docs/concepts/retries)).
 - **Tiered timeouts** — short / medium / long tiers picked per endpoint, overridable per call ([Timeouts](https://docs.apify.com/api/client/python/docs/concepts/timeouts)).
-- **Pagination and streaming** — iterate datasets, key-value store keys, or live logs without manual paging or buffering ([Pagination](https://docs.apify.com/api/client/python/docs/concepts/pagination), [Streaming](https://docs.apify.com/api/client/python/docs/concepts/streaming-resources)).
+- **Pagination and streaming** — iterate datasets, key-value store keys, or live logs without manual paging or buffering, and upload large records straight from a file or another stream ([Pagination](https://docs.apify.com/api/client/python/docs/concepts/pagination), [Streaming](https://docs.apify.com/api/client/python/docs/concepts/streaming-resources)).
 - **Convenience methods** — `call()`, `wait_for_finish()`, nested resource access, and other shortcuts that hide platform quirks ([Convenience methods](https://docs.apify.com/api/client/python/docs/concepts/convenience-methods)).
 - **Pluggable HTTP layer** — use the default [Impit](https://github.com/apify/impit)-based client, opt in to the built-in [HTTPX2](https://github.com/pydantic/httpx2) client, or plug in any custom implementation ([HTTP clients](https://docs.apify.com/api/client/python/docs/concepts/custom-http-clients)).
 - **Structured errors** — every API error surfaces as an [`ApifyApiError`](https://docs.apify.com/api/client/python/reference/class/ApifyApiError) with HTTP-specific subclasses for precise handling ([Error handling](https://docs.apify.com/api/client/python/docs/concepts/error-handling)).

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/README.md

@@ -133,7 +133,7 @@ For a guided walkthrough — authenticating, running an Actor, and reading its r
 - **Fully typed responses** — every method returns a [Pydantic](https://docs.pydantic.dev/) model generated from the Apify OpenAPI spec, with IDE autocomplete and runtime validation ([Typed models](https://docs.apify.com/api/client/python/docs/concepts/typed-models)).
 - **Automatic retries** — exponential backoff for network errors, HTTP 429, and 5xx responses, configurable per client ([Retries](https://docs.apify.com/api/client/python/docs/concepts/retries)).
 - **Tiered timeouts** — short / medium / long tiers picked per endpoint, overridable per call ([Timeouts](https://docs.apify.com/api/client/python/docs/concepts/timeouts)).
-- **Pagination and streaming** — iterate datasets, key-value store keys, or live logs without manual paging or buffering ([Pagination](https://docs.apify.com/api/client/python/docs/concepts/pagination), [Streaming](https://docs.apify.com/api/client/python/docs/concepts/streaming-resources)).
+- **Pagination and streaming** — iterate datasets, key-value store keys, or live logs without manual paging or buffering, and upload large records straight from a file or another stream ([Pagination](https://docs.apify.com/api/client/python/docs/concepts/pagination), [Streaming](https://docs.apify.com/api/client/python/docs/concepts/streaming-resources)).
 - **Convenience methods** — `call()`, `wait_for_finish()`, nested resource access, and other shortcuts that hide platform quirks ([Convenience methods](https://docs.apify.com/api/client/python/docs/concepts/convenience-methods)).
 - **Pluggable HTTP layer** — use the default [Impit](https://github.com/apify/impit)-based client, opt in to the built-in [HTTPX2](https://github.com/pydantic/httpx2) client, or plug in any custom implementation ([HTTP clients](https://docs.apify.com/api/client/python/docs/concepts/custom-http-clients)).
 - **Structured errors** — every API error surfaces as an [`ApifyApiError`](https://docs.apify.com/api/client/python/reference/class/ApifyApiError) with HTTP-specific subclasses for precise handling ([Error handling](https://docs.apify.com/api/client/python/docs/concepts/error-handling)).

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "apify_client"
-version = "3.2.1b4"
+version = "3.2.2b1"
 description = "Apify API client for Python"
 license = "Apache-2.0"
 license-files = ["LICENSE"]
@@ -178,6 +178,7 @@ indent-style = "space"
 addopts = "-r a --verbose --dist worksteal"
 asyncio_default_fixture_loop_scope = "function"
 asyncio_mode = "auto"
+filterwarnings = ["ignore:Streaming a request body is experimental:UserWarning"]
 pythonpath = ["."]
 timeout = 1800
 

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/pyproject.toml.orig

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "apify_client"
-version = "3.2.1b4"
+version = "3.2.2b1"
 description = "Apify API client for Python"
 authors = [{ name = "Apify Technologies s.r.o.", email = "support@apify.com" }]
 license = "Apache-2.0"
@@ -187,6 +187,10 @@ max-branches = 18
 addopts = "-r a --verbose --dist worksteal"
 asyncio_default_fixture_loop_scope = "function"
 asyncio_mode = "auto"
+filterwarnings = [
+    # The tests stream request bodies throughout, and the warning has a unit test of its own.
+    "ignore:Streaming a request body is experimental:UserWarning",
+]
 pythonpath = ["."]
 timeout = 1800
 

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/_consts.py

@@ -42,6 +42,15 @@ A smaller body already fits in a single network packet, so compressing it costs
 saving a round trip.
 """
 
+STREAMED_BODY_CHUNK_SIZE = 64 * 1024
+"""Size of the chunks a streamed request body reads from an `io.IOBase` source, in bytes from a binary stream and in
+characters from a text-mode one.
+
+A chunk is the most of a streamed body that is in memory at once, and in the asynchronous client every chunk costs
+one worker-thread hop, so the size balances memory against per-chunk overhead. It matches the buffer size the
+standard library and common HTTP libraries use to copy files.
+"""
+
 ALREADY_COMPRESSED_MEDIA_TYPE_PREFIXES = ('audio/', 'image/', 'video/')
 """Media type prefixes whose payloads carry their own compression, so compressing the request body is wasted work."""
 

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/_resource_clients/actor.py

@@ -236,7 +236,10 @@ class ActorClient(ResourceClient):
         https://docs.apify.com/api/v2#/reference/actors/run-collection/run-actor
 
         Args:
-            run_input: The input to pass to the Actor run.
+            run_input: The input to pass to the Actor run. Accepts the same values as
+                `KeyValueStoreClient.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             build: Specifies the Actor build to run. It can be either a build tag or build number. By default,
                 the run uses the build specified in the default run configuration for the Actor (typically latest).
@@ -315,7 +318,10 @@ class ActorClient(ResourceClient):
         https://docs.apify.com/api/v2#/reference/actors/run-collection/run-actor
 
         Args:
-            run_input: The input to pass to the Actor run.
+            run_input: The input to pass to the Actor run. Accepts the same values as
+                `KeyValueStoreClient.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             build: Specifies the Actor build to run. It can be either a build tag or build number. By default,
                 the run uses the build specified in the default run configuration for the Actor (typically latest).
@@ -540,7 +546,10 @@ class ActorClient(ResourceClient):
         """Validate an input for the Actor that defines an input schema.
 
         Args:
-            run_input: The input to validate.
+            run_input: The input to validate. Accepts the same values as `KeyValueStoreClient.set_record`,
+                including an `io.IOBase` stream such as an open file, an iterable of byte chunks, or a streamed
+                `HttpResponse`, which are uploaded in chunks without being held in memory. Streaming is experimental,
+                and its behavior may change in future versions.
             build_tag: The Actor's build tag.
             content_type: The content type of the input.
             timeout: Timeout for the API HTTP request.
@@ -738,7 +747,10 @@ class ActorClientAsync(ResourceClientAsync):
         https://docs.apify.com/api/v2#/reference/actors/run-collection/run-actor
 
         Args:
-            run_input: The input to pass to the Actor run.
+            run_input: The input to pass to the Actor run. Accepts the same values as
+                `KeyValueStoreClientAsync.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             build: Specifies the Actor build to run. It can be either a build tag or build number. By default,
                 the run uses the build specified in the default run configuration for the Actor (typically latest).
@@ -817,7 +829,10 @@ class ActorClientAsync(ResourceClientAsync):
         https://docs.apify.com/api/v2#/reference/actors/run-collection/run-actor
 
         Args:
-            run_input: The input to pass to the Actor run.
+            run_input: The input to pass to the Actor run. Accepts the same values as
+                `KeyValueStoreClientAsync.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             build: Specifies the Actor build to run. It can be either a build tag or build number. By default,
                 the run uses the build specified in the default run configuration for the Actor (typically latest).
@@ -1043,7 +1058,10 @@ class ActorClientAsync(ResourceClientAsync):
         """Validate an input for the Actor that defines an input schema.
 
         Args:
-            run_input: The input to validate.
+            run_input: The input to validate. Accepts the same values as `KeyValueStoreClientAsync.set_record`,
+                including an `io.IOBase` stream such as an open file, an iterable of byte chunks, or a streamed
+                `HttpResponse`, which are uploaded in chunks without being held in memory. Streaming is experimental,
+                and its behavior may change in future versions.
             build_tag: The Actor's build tag.
             content_type: The content type of the input.
             timeout: Timeout for the API HTTP request.

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/_resource_clients/key_value_store.py

@@ -369,14 +369,19 @@ class KeyValueStoreClient(ResourceClient):
 
         Args:
             key: The key of the record to save the value to.
-            value: The value to save into the record.
+            value: The value to save into the record. An `io.IOBase` stream such as an open file, an iterable of byte
+                chunks, or a streamed `HttpResponse` is uploaded in chunks as it is read, without being held in memory
+                whole or compressed. Any other object with a callable `read` is read whole and sent uncompressed. Only
+                a seekable `io.IOBase` value can be retried; any other streamed value gets a single attempt. See
+                `StreamedRequestBody` for details. Streaming is experimental, and its behavior may change in future
+                versions.
             content_type: The content type of the saved value.
             content_encoding: The encoding already applied to `value`, sent as the `Content-Encoding` header. Pass it
                 to upload a pre-compressed value - the client then forwards the bytes as they are instead of
                 compressing them itself. The API accepts `gzip`, `br`, `deflate`, and `identity`, and stores the
-                record exactly as uploaded, so this also becomes the encoding the record is served with. Only a
-                bytes-like `value`, or a file-like one that reads into bytes, can carry a compression - anything
-                else raises `TypeError` instead of being stored under a header that misdescribes it.
+                record exactly as uploaded, so this also becomes the encoding the record is served with. A `str`, a
+                JSON-serializable object, or a text-mode file cannot be carrying a compression and raises
+                `TypeError`; a streamed value is taken at its word, since its bytes are only seen as they are sent.
             timeout: Timeout for the API HTTP request.
         """
         value, content_type = encode_key_value_store_record_value(
@@ -802,14 +807,19 @@ class KeyValueStoreClientAsync(ResourceClientAsync):
 
         Args:
             key: The key of the record to save the value to.
-            value: The value to save into the record.
+            value: The value to save into the record. An `io.IOBase` stream such as an open file, an iterable of byte
+                chunks, or a streamed `HttpResponse` is uploaded in chunks as it is read, without being held in memory
+                whole or compressed. Any other object with a callable `read` is read whole and sent uncompressed. Only
+                a seekable `io.IOBase` value can be retried; any other streamed value gets a single attempt. See
+                `StreamedRequestBody` for details. Streaming is experimental, and its behavior may change in future
+                versions.
             content_type: The content type of the saved value.
             content_encoding: The encoding already applied to `value`, sent as the `Content-Encoding` header. Pass it
                 to upload a pre-compressed value - the client then forwards the bytes as they are instead of
                 compressing them itself. The API accepts `gzip`, `br`, `deflate`, and `identity`, and stores the
-                record exactly as uploaded, so this also becomes the encoding the record is served with. Only a
-                bytes-like `value`, or a file-like one that reads into bytes, can carry a compression - anything
-                else raises `TypeError` instead of being stored under a header that misdescribes it.
+                record exactly as uploaded, so this also becomes the encoding the record is served with. A `str`, a
+                JSON-serializable object, or a text-mode file cannot be carrying a compression and raises
+                `TypeError`; a streamed value is taken at its word, since its bytes are only seen as they are sent.
             timeout: Timeout for the API HTTP request.
         """
         value, content_type = encode_key_value_store_record_value(

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/_resource_clients/run.py

@@ -180,7 +180,10 @@ class RunClient(ResourceClient):
             target_actor_build: The build of the target Actor. It can be either a build tag or build number.
                 By default, the run uses the build specified in the default run configuration for the target Actor
                 (typically the latest build).
-            run_input: The input to pass to the new run.
+            run_input: The input to pass to the new run. Accepts the same values as
+                `KeyValueStoreClient.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             timeout: Timeout for the API HTTP request.
 
@@ -608,7 +611,10 @@ class RunClientAsync(ResourceClientAsync):
             target_actor_build: The build of the target Actor. It can be either a build tag or build number.
                 By default, the run uses the build specified in the default run configuration for the target Actor
                 (typically the latest build).
-            run_input: The input to pass to the new run.
+            run_input: The input to pass to the new run. Accepts the same values as
+                `KeyValueStoreClientAsync.set_record`, including an `io.IOBase` stream such as an open file, an iterable
+                of byte chunks, or a streamed `HttpResponse`, which are uploaded in chunks without being held in
+                memory. Streaming is experimental, and its behavior may change in future versions.
             content_type: The content type of the input.
             timeout: Timeout for the API HTTP request.
 

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/_utils/encoding.py

@@ -1,32 +1,35 @@
 from __future__ import annotations
 
+import io
 import json
 from base64 import b64encode
 from functools import cache
-from inspect import isawaitable, iscoroutine
 from typing import TYPE_CHECKING, Any
 
 from apify_client._models import WebhookCreate, WebhookRepresentation
+from apify_client.http_clients._streamed_body import StreamedRequestBody
 
 if TYPE_CHECKING:
-    from apify_client.types import WebhooksList
+    from apify_client.types import StreamedBodySource, WebhooksList
 
 
 def encode_key_value_store_record_value(
     value: Any, *, content_type: str | None = None, content_encoding: str | None = None
-) -> tuple[bytes | bytearray | str, str]:
+) -> tuple[bytes | bytearray | str | StreamedBodySource, str]:
     """Encode a value for storage in a key-value store record.
 
     Args:
-        value: The value to encode. Anything exposing a callable `read` is treated as a file-like object: `read`
-            is called with no arguments, so the value is consumed from its current position and buffered in
-            memory whole - the object is neither rewound nor closed, and async file-like objects are rejected.
-            Any other value is JSON-serialized unless it is already bytes or a string.
-        content_type: The content type; if None, it's inferred from the value type.
+        value: The value to encode. A file-like object (anything with a callable `read`), an iterable of byte chunks,
+            or a streamed `HttpResponse` is returned as it is, to be streamed to the API from its current position -
+            the object is neither rewound nor closed. See `StreamedRequestBody.is_streamable` for the exact rules. Any
+            other value is JSON-serialized unless it is already bytes or a string.
+        content_type: The content type; if None, it's inferred from the value type. An `io.TextIOBase`, which is what
+            the standard library returns for a file opened in text mode, is `text/plain; charset=utf-8`; any other
+            streamed value is `application/octet-stream`.
         content_encoding: The encoding the caller declares the value already carries, if any. Anything other than
-            `identity` means the value is compressed, which only a bytes-like payload can be, so any other value
-            is rejected. The check belongs here because a file-like value has to be read before its payload type
-            is known, and reading it a second time in the caller is not possible.
+            `identity` means the value is compressed, which only a bytes-like payload can be, so a string, a
+            JSON-serialized object, or a text-mode file is rejected. Any other streamed value is taken at its word,
+            since its bytes are only seen as they are sent.
 
     Returns:
         A tuple of (encoded_value, content_type).
@@ -35,29 +38,23 @@ def encode_key_value_store_record_value(
         TypeError: If the value cannot be encoded into a body the transport accepts, or if it cannot be carrying
             the declared `content_encoding`.
     """
-    # Read file-like values into memory; the transport only accepts bytes-like bodies. Detect them by a
-    # callable `read` (not `io.IOBase`) so duck-typed file-likes are read, not JSON-serialized. Impit exposes
-    # no streaming `content=` API, so the value has to be buffered whole.
-    read = getattr(value, 'read', None)
-    if callable(read):
-        value = read()
-
-        if isawaitable(value):
-            if iscoroutine(value):
-                value.close()  # Prevent a "coroutine was never awaited" warning.
+    declared_encoding = (content_encoding or '').strip().lower()
+    declares_compression = declared_encoding not in ('', 'identity')
+
+    if StreamedRequestBody.is_streamable(value):
+        is_text = isinstance(value, io.TextIOBase)
+        if declares_compression and is_text:
             raise TypeError(
-                'Async file-like objects are not supported. Await the read yourself and pass the resulting '
-                'bytes or string.'
+                f'Cannot upload a file-like value opened in text mode with `Content-Encoding: {content_encoding}`. '
+                'An encoding other than `identity` declares the value is already compressed, so pass the compressed '
+                'bytes, or a binary file-like object that reads them.'
             )
+        return (value, content_type or ('text/plain; charset=utf-8' if is_text else 'application/octet-stream'))
 
-        if not isinstance(value, (bytes, bytearray, str)):
-            raise TypeError(f'Reading the file-like value returned {type(value).__name__}, expected bytes or str.')
-
-    # A declared compression describes bytes the caller compressed. A string, a JSON-serializable object, or a
-    # text-mode file cannot be carrying one, and would otherwise be stored under a header that misdescribes it -
-    # the client forwards the header untouched and never inspects the body.
-    declared_encoding = (content_encoding or '').strip().lower()
-    if declared_encoding not in ('', 'identity') and not isinstance(value, (bytes, bytearray)):
+    # A declared compression describes bytes the caller compressed. A string or a JSON-serializable object cannot
+    # be carrying one, and would otherwise be stored under a header that misdescribes it - the client forwards the
+    # header untouched and never inspects the body.
+    if declares_compression and not isinstance(value, (bytes, bytearray)):
         raise TypeError(
             f'Cannot upload a {type(value).__name__} value with `Content-Encoding: {content_encoding}`. An encoding '
             'other than `identity` declares the value is already compressed, so pass the compressed bytes, or a '

{apify_client-3.2.1b4 → apify_client-3.2.2b1}/src/apify_client/http_clients/__init__.py

@@ -2,6 +2,7 @@ from apify_client._utils.try_import import install_import_hook as _install_impor
 from apify_client._utils.try_import import try_import as _try_import
 from apify_client.http_clients._base import HttpClient, HttpClientAsync, HttpResponse
 from apify_client.http_clients._impit import ImpitHttpClient, ImpitHttpClientAsync
+from apify_client.http_clients._streamed_body import StreamedRequestBody
 
 _install_import_hook(__name__)
 
@@ -22,6 +23,7 @@ __all__ = [
     'HttpResponse',
     'ImpitHttpClient',
     'ImpitHttpClientAsync',
+    'StreamedRequestBody',
 ]
 
 if _httpx2_import.available: