feat: photoMaker in imageInference, IIpAdapter (guideImages, etc.), IEmbedding weight, dict coercion for nested options by Sirsho1997 · Pull Request #251 · Runware/sdk-python

Sirsho1997 · 2026-03-05T03:13:31Z

Added

Added photoMaker as nested object in imageInference using IPhotoMaker; IImageInference.photoMaker: Optional[Union[IPhotoMaker, Dict[str, Any]]]. IPhotoMaker has optional model, positivePrompt, height, width for nested use; images and inputImages (standalone uses inputImages, imageInference uses images). Existing taskType photoMaker unchanged; standalone validates required model, positivePrompt, height, width.
Added photoMaker payload building in imageInference flow (photo_maker_data) and _addImageSpecialFields; existing taskType photoMaker unchanged.
Added IIpAdapter.guideImages, combineMethod, weightType, embedScaling, weightComposition; guideImage optional; base supports guideImages with process_image.
Added IEmbedding.weight: Optional[float].
Added dict coercion in IImageInference.post_init for photoMaker, instantID, acePlusPlus, puLID, ultralytics, outpaint, refiner; list coercion for embeddings (dict → IEmbedding) and ipAdapters (dict → IIpAdapter).

Changed

IPhotoMaker: model, positivePrompt, height, width now Optional for use as nested photoMaker in imageInference; added images: Optional[List[Union[str, File]]]. Removed post_init validation (max 4 images, style).
IImageInference: refiner, outpaint, instantID, acePlusPlus, puLID, ultralytics, photoMaker typed as Union[Type, Dict[str, Any]]; embeddings as List[Union[IEmbedding, Dict[str, Any]]]; ipAdapters as List[Union[IIpAdapter, Dict[str, Any]]].

…Embedding weight, dict coercion for nested options

Copilot

Pull request overview

This PR extends the image inference request types and request-building flow to support PhotoMaker settings as a nested object, expands IP-Adapter options (including multi-image guidance), adds IEmbedding.weight, and adds dict/list coercion for several nested options inside IImageInference.

Changes:

Added IPhotoMakerSettings and IImageInference.photoMaker, plus payload building and request wiring in the image-inference flow.
Expanded IIpAdapter fields (including guideImages) and added dict/list coercion for ipAdapters and embeddings items.
Added IEmbedding.weight and consolidated PhotoMaker style validation into a shared _PHOTO_MAKER_VALID_STYLES constant.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 6 comments.

File	Description
`runware/types.py`	Adds new/expanded request dataclasses and `IImageInference.__post_init__` coercion to accept nested dicts/lists.
`runware/base.py`	Updates image-inference request building to process `guideImages` and include `photoMaker` payload in the request.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

runware/types.py

runware/base.py

runware/types.py

… in imageInference

Copilot

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

runware/base.py

runware/types.py

Copilot

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 8 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

runware/base.py

runware/types.py

runware/base.py

runware/types.py

Copilot

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

Comments suppressed due to low confidence (1)

runware/base.py:668

expected_results is set from requestPhotoMaker.numberResults, but IPhotoMaker.numberResults is now optional and may be None. That will break _handle_pending_operation_message (len(results) >= op["expected"] will raise TypeError). Default numberResults to an int (e.g., or 1) before registering the pending op (and ideally before putting it in the request payload).

        future, should_send = await self._register_pending_operation(
            task_uuid,
            expected_results=numberOfResults,
            complete_predicate=None,
            result_filter=lambda r: r.get("imageUUID") is not None

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-06T03:09:04Z

runware/base.py

+            if "inputImages" in instant_id_data:
+                instant_id_data["inputImages"] = await process_image(instant_id_data["inputImages"])
+            else:
+                instant_id_data["inputImages"] = await process_image([instant_id_data["inputImage"]])



instant_id_data["inputImage"] can still be present (and unprocessed) while you also set inputImages. Additionally, the fallback path assumes inputImage exists and will raise KeyError if neither inputImages nor inputImage was provided. Consider normalizing to a single field before sending (e.g., require at least one of inputImages/inputImage, build inputImages, process it, and remove inputImage from the payload dict).

Suggested change

if "inputImages" in instant_id_data:

instant_id_data["inputImages"] = await process_image(instant_id_data["inputImages"])

else:

instant_id_data["inputImages"] = await process_image([instant_id_data["inputImage"]])

# Normalize instant_id_data images to a single `inputImages` field

input_images = instant_id_data.get("inputImages")

single_input_image = instant_id_data.get("inputImage")

if input_images is None and single_input_image is not None:

# Build inputImages list from a single inputImage

input_images = [single_input_image]

if input_images is not None:

instant_id_data["inputImages"] = await process_image(input_images)

# Remove the legacy single-image field to avoid unprocessed duplicates

instant_id_data.pop("inputImage", None)

Copilot · 2026-03-06T03:09:04Z

runware/base.py

            response = results[0]
            self._handle_error_response(response)
-            return createImageToTextFromResponse(response)
+            return instantiateDataclass(IImageToText, response)


Switching to instantiateDataclass(IImageToText, response) drops the previous response coercions (e.g., taskType string -> ETaskType and cost -> float). Since instantiateDataclass currently doesn’t handle Enum or numeric coercion, callers will now get raw strings for these fields. Either restore a dedicated response parser for IImageToText or extend instantiateDataclass to coerce Enum/numeric types consistently (similar to createEnhancedPromptsFromResponse).

Suggested change

return instantiateDataclass(IImageToText, response)

# Coerce response fields to expected types before dataclass instantiation.

# This restores previous behaviour where taskType was an ETaskType

# and cost was a float, rather than raw strings.

if isinstance(response, dict):

response_dict = dict(response)

else:

try:

# Attempt a shallow conversion for mapping-like objects

response_dict = dict(response)

except TypeError:

# Fallback: leave as-is if it cannot be converted

response_dict = response

if isinstance(response_dict, dict):

task_type_value = response_dict.get("taskType")

if isinstance(task_type_value, str):

try:

response_dict["taskType"] = ETaskType(task_type_value)

except ValueError:

# If the value is not a valid ETaskType, leave it unchanged

pass

cost_value = response_dict.get("cost")

if isinstance(cost_value, str):

try:

response_dict["cost"] = float(cost_value)

except (ValueError, TypeError):

# If cost is not a valid float, leave it unchanged

pass

return instantiateDataclass(IImageToText, response_dict)

teith

need also to fix the problems copilot found

teith · 2026-03-10T00:58:05Z

runware/base.py


+        photo_maker_data = {}
+        if requestImage.photoMaker:
+            photo_maker_data = {}


lol. Will remove that double initialization

teith · 2026-03-10T01:00:09Z

runware/types.py

    inputImages: Optional[List[Union[File, str]]] = None  # Primary: array of images (UUIDs, URLs, base64, dataURI)
    inputImage: Optional[Union[File, str]] = None  # Convenience: single image, defaults to inputImages[0] if not provided
-    prompt: List[str] = field(default_factory=lambda: ["Describe this image in detail"])  # Array of prompts with default
+    prompt: Optional[List[str]] = None  


why? why is it optional?

Age detection (subset of imageCaption) does not need a prompt

feat: photoMaker in imageInference, IIpAdapter (guideImages, etc.), I…

6ea7401

…Embedding weight, dict coercion for nested options

Sirsho1997 requested a review from Copilot March 5, 2026 04:07

Copilot started reviewing on behalf of Sirsho1997 March 5, 2026 04:08 View session

Copilot AI reviewed Mar 5, 2026

View reviewed changes

runware/types.py Outdated Show resolved Hide resolved

runware/types.py Show resolved Hide resolved

runware/base.py Show resolved Hide resolved

runware/base.py Show resolved Hide resolved

runware/base.py Outdated Show resolved Hide resolved

runware/types.py Outdated Show resolved Hide resolved

Sirsho1997 added 2 commits March 5, 2026 15:12

Refactored to use IPhotomaker as independent photoMaker or as a param…

e4ed805

… in imageInference

Updated _addImageSpecialFields for embedding in imageInference

c0d9aa5

Sirsho1997 requested a review from Copilot March 5, 2026 20:17

Copilot AI reviewed Mar 5, 2026

View reviewed changes

runware/base.py Outdated Show resolved Hide resolved

runware/base.py Outdated Show resolved Hide resolved

runware/types.py Outdated Show resolved Hide resolved

Sirsho1997 added 2 commits March 5, 2026 15:55

Add inputImages to instantID

c554ad7

Add more fields to IPhotoMaker

ca6ed1d

Sirsho1997 requested a review from Copilot March 6, 2026 02:20

Copilot started reviewing on behalf of Sirsho1997 March 6, 2026 02:20 View session

Copilot AI reviewed Mar 6, 2026

View reviewed changes

Fixed imageCaption

beb4b52

Sirsho1997 requested a review from Copilot March 6, 2026 03:04

Copilot started reviewing on behalf of Sirsho1997 March 6, 2026 03:05 View session

Copilot AI reviewed Mar 6, 2026

View reviewed changes

teith requested changes Mar 10, 2026

View reviewed changes

-            if "inputImages" in instant_id_data:
-                instant_id_data["inputImages"] = await process_image(instant_id_data["inputImages"])
-            else:
-                instant_id_data["inputImages"] = await process_image([instant_id_data["inputImage"]])
+            # Normalize instant_id_data images to a single `inputImages` field
+            input_images = instant_id_data.get("inputImages")
+            single_input_image = instant_id_data.get("inputImage")
+            if input_images is None and single_input_image is not None:
+                # Build inputImages list from a single inputImage
+                input_images = [single_input_image]
+            if input_images is not None:
+                instant_id_data["inputImages"] = await process_image(input_images)
+            # Remove the legacy single-image field to avoid unprocessed duplicates
+            instant_id_data.pop("inputImage", None)

-            return instantiateDataclass(IImageToText, response)
+            # Coerce response fields to expected types before dataclass instantiation.
+            # This restores previous behaviour where taskType was an ETaskType
+            # and cost was a float, rather than raw strings.
+            if isinstance(response, dict):
+                response_dict = dict(response)
+            else:
+                try:
+                    # Attempt a shallow conversion for mapping-like objects
+                    response_dict = dict(response)
+                except TypeError:
+                    # Fallback: leave as-is if it cannot be converted
+                    response_dict = response
+            if isinstance(response_dict, dict):
+                task_type_value = response_dict.get("taskType")
+                if isinstance(task_type_value, str):
+                    try:
+                        response_dict["taskType"] = ETaskType(task_type_value)
+                    except ValueError:
+                        # If the value is not a valid ETaskType, leave it unchanged
+                        pass
+                cost_value = response_dict.get("cost")
+                if isinstance(cost_value, str):
+                    try:
+                        response_dict["cost"] = float(cost_value)
+                    except (ValueError, TypeError):
+                        # If cost is not a valid float, leave it unchanged
+                        pass
+            return instantiateDataclass(IImageToText, response_dict)

Conversation

Sirsho1997 commented Mar 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Added

Changed

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 6, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 6, 2026

Choose a reason for hiding this comment

Uh oh!

teith left a comment

Choose a reason for hiding this comment

Uh oh!

teith Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

Sirsho1997 Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

teith Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

Sirsho1997 Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Sirsho1997 commented Mar 5, 2026 •

edited

Loading