dsh-plugin-image-workbench 1.1.5 → 1.1.7

package/vendor/chatgpt2api/source/services/protocol/openai_v1_image_edit.py

@@ -1,86 +1,86 @@
-from __future__ import annotations
-
-from io import BytesIO
-from typing import Any, Iterator
-
-from PIL import Image
-
-from services.protocol.conversation import (
-    ConversationRequest,
-    ImageGenerationError,
-    collect_image_outputs,
-    count_text_tokens,
-    encode_images,
-    stream_image_chunks,
-    stream_image_outputs_with_pool,
-)
-from utils.image_tokens import count_image_inputs_tokens, count_image_output_items_tokens, image_usage
-
-
-def _composite_mask(
-    images: list[tuple[bytes, str, str]],
-    masks: list[tuple[bytes, str, str]],
-) -> list[tuple[bytes, str, str]]:
-    """将 mask 的 alpha 通道合成到图片中，标识需要编辑的区域。
-    
-    mask 的透明区域（低 alpha）= 需要编辑的区域，
-    mask 的不透明区域（高 alpha）= 保留的区域。
-    如果无 mask 则返回原图。
-    """
-    if not masks:
-        return images
-    result: list[tuple[bytes, str, str]] = []
-    for i, (data, filename, mime_type) in enumerate(images):
-        mask_data = masks[i][0] if i < len(masks) else masks[-1][0]
-        img = Image.open(BytesIO(data)).convert("RGBA")
-        mask_img = Image.open(BytesIO(mask_data))
-        if mask_img.mode == "RGBA":
-            alpha = mask_img.split()[3]
-        elif mask_img.mode == "L":
-            alpha = mask_img
-        else:
-            alpha = mask_img.convert("L")
-        alpha = alpha.resize(img.size, Image.LANCZOS)
-        img.putalpha(alpha)
-        buf = BytesIO()
-        img.save(buf, format="PNG")
-        result.append((buf.getvalue(), filename, "image/png"))
-    return result
-
-
-def handle(body: dict[str, Any]) -> dict[str, Any] | Iterator[dict[str, Any]]:
-    prompt = str(body.get("prompt") or "")
-    images = body.get("images") or []
-    masks = body.get("mask") or []
-    images = _composite_mask(images, masks)
-    model = str(body.get("model") or "gpt-image-2")
-    n = int(body.get("n") or 1)
-    size = body.get("size")
-    quality = str(body.get("quality") or "auto")
-    response_format = str(body.get("response_format") or "b64_json")
-    base_url = str(body.get("base_url") or "") or None
-    progress_callback = body.get("progress_callback")
-    encoded_images = encode_images(images)
-    if not encoded_images:
-        raise ImageGenerationError("image is required")
-    outputs = stream_image_outputs_with_pool(ConversationRequest(
-        prompt=prompt,
-        model=model,
-        n=n,
-        size=size,
-        quality=quality,
-        response_format=response_format,
-        base_url=base_url,
-        images=encoded_images,
-        message_as_error=True,
-        progress_callback=progress_callback,
-    ))
-    if body.get("stream"):
-        return stream_image_chunks(outputs)
-    result = collect_image_outputs(outputs)
-    result["usage"] = image_usage(
-        input_text_tokens=count_text_tokens(prompt, model),
-        input_image_tokens=count_image_inputs_tokens(images, model),
-        output_tokens=count_image_output_items_tokens(result.get("data"), size, quality),
-    )
-    return result
+from __future__ import annotations
+
+from io import BytesIO
+from typing import Any, Iterator
+
+from PIL import Image
+
+from services.protocol.conversation import (
+    ConversationRequest,
+    ImageGenerationError,
+    collect_image_outputs,
+    count_text_tokens,
+    encode_images,
+    stream_image_chunks,
+    stream_image_outputs_with_pool,
+)
+from utils.image_tokens import count_image_inputs_tokens, count_image_output_items_tokens, image_usage
+
+
+def _composite_mask(
+    images: list[tuple[bytes, str, str]],
+    masks: list[tuple[bytes, str, str]],
+) -> list[tuple[bytes, str, str]]:
+    """将 mask 的 alpha 通道合成到图片中，标识需要编辑的区域。
+    
+    mask 的透明区域（低 alpha）= 需要编辑的区域，
+    mask 的不透明区域（高 alpha）= 保留的区域。
+    如果无 mask 则返回原图。
+    """
+    if not masks:
+        return images
+    result: list[tuple[bytes, str, str]] = []
+    for i, (data, filename, mime_type) in enumerate(images):
+        mask_data = masks[i][0] if i < len(masks) else masks[-1][0]
+        img = Image.open(BytesIO(data)).convert("RGBA")
+        mask_img = Image.open(BytesIO(mask_data))
+        if mask_img.mode == "RGBA":
+            alpha = mask_img.split()[3]
+        elif mask_img.mode == "L":
+            alpha = mask_img
+        else:
+            alpha = mask_img.convert("L")
+        alpha = alpha.resize(img.size, Image.LANCZOS)
+        img.putalpha(alpha)
+        buf = BytesIO()
+        img.save(buf, format="PNG")
+        result.append((buf.getvalue(), filename, "image/png"))
+    return result
+
+
+def handle(body: dict[str, Any]) -> dict[str, Any] | Iterator[dict[str, Any]]:
+    prompt = str(body.get("prompt") or "")
+    images = body.get("images") or []
+    masks = body.get("mask") or []
+    images = _composite_mask(images, masks)
+    model = str(body.get("model") or "gpt-image-2")
+    n = int(body.get("n") or 1)
+    size = body.get("size")
+    quality = str(body.get("quality") or "auto")
+    response_format = str(body.get("response_format") or "b64_json")
+    base_url = str(body.get("base_url") or "") or None
+    progress_callback = body.get("progress_callback")
+    encoded_images = encode_images(images)
+    if not encoded_images:
+        raise ImageGenerationError("image is required")
+    outputs = stream_image_outputs_with_pool(ConversationRequest(
+        prompt=prompt,
+        model=model,
+        n=n,
+        size=size,
+        quality=quality,
+        response_format=response_format,
+        base_url=base_url,
+        images=encoded_images,
+        message_as_error=True,
+        progress_callback=progress_callback,
+    ))
+    if body.get("stream"):
+        return stream_image_chunks(outputs)
+    result = collect_image_outputs(outputs)
+    result["usage"] = image_usage(
+        input_text_tokens=count_text_tokens(prompt, model),
+        input_image_tokens=count_image_inputs_tokens(images, model),
+        output_tokens=count_image_output_items_tokens(result.get("data"), size, quality),
+    )
+    return result

package/vendor/chatgpt2api/source/services/protocol/openai_v1_image_generations.py

@@ -1,42 +1,42 @@
-from __future__ import annotations
-
-from typing import Any, Iterator
-
-from services.protocol.conversation import (
-    ConversationRequest,
-    collect_image_outputs,
-    count_text_tokens,
-    stream_image_chunks,
-    stream_image_outputs_with_pool,
-)
-from utils.image_tokens import count_image_output_items_tokens, image_usage
-
-
-def handle(body: dict[str, Any]) -> dict[str, Any] | Iterator[dict[str, Any]]:
-    prompt = str(body.get("prompt") or "")
-    model = str(body.get("model") or "gpt-image-2")
-    n = int(body.get("n") or 1)
-    size = body.get("size")
-    quality = str(body.get("quality") or "auto")
-    response_format = str(body.get("response_format") or "b64_json")
-    base_url = str(body.get("base_url") or "") or None
-    progress_callback = body.get("progress_callback")
-    outputs = stream_image_outputs_with_pool(ConversationRequest(
-        prompt=prompt,
-        model=model,
-        n=n,
-        size=size,
-        quality=quality,
-        response_format=response_format,
-        base_url=base_url,
-        message_as_error=True,
-        progress_callback=progress_callback,
-    ))
-    if body.get("stream"):
-        return stream_image_chunks(outputs)
-    result = collect_image_outputs(outputs)
-    result["usage"] = image_usage(
-        input_text_tokens=count_text_tokens(prompt, model),
-        output_tokens=count_image_output_items_tokens(result.get("data"), size, quality),
-    )
-    return result
+from __future__ import annotations
+
+from typing import Any, Iterator
+
+from services.protocol.conversation import (
+    ConversationRequest,
+    collect_image_outputs,
+    count_text_tokens,
+    stream_image_chunks,
+    stream_image_outputs_with_pool,
+)
+from utils.image_tokens import count_image_output_items_tokens, image_usage
+
+
+def handle(body: dict[str, Any]) -> dict[str, Any] | Iterator[dict[str, Any]]:
+    prompt = str(body.get("prompt") or "")
+    model = str(body.get("model") or "gpt-image-2")
+    n = int(body.get("n") or 1)
+    size = body.get("size")
+    quality = str(body.get("quality") or "auto")
+    response_format = str(body.get("response_format") or "b64_json")
+    base_url = str(body.get("base_url") or "") or None
+    progress_callback = body.get("progress_callback")
+    outputs = stream_image_outputs_with_pool(ConversationRequest(
+        prompt=prompt,
+        model=model,
+        n=n,
+        size=size,
+        quality=quality,
+        response_format=response_format,
+        base_url=base_url,
+        message_as_error=True,
+        progress_callback=progress_callback,
+    ))
+    if body.get("stream"):
+        return stream_image_chunks(outputs)
+    result = collect_image_outputs(outputs)
+    result["usage"] = image_usage(
+        input_text_tokens=count_text_tokens(prompt, model),
+        output_tokens=count_image_output_items_tokens(result.get("data"), size, quality),
+    )
+    return result

package/vendor/chatgpt2api/source/services/protocol/openai_v1_models.py

@@ -1,53 +1,53 @@
-from __future__ import annotations
-
-from typing import Any
-
-from services.account_service import account_service
-from services.model_service import model_catalog_service
-from utils.helper import CODEX_IMAGE_MODEL
-
-
-def list_models() -> dict[str, Any]:
-    result = model_catalog_service.list_models()
-    data = result.get("data")
-    if not isinstance(data, list):
-        return result
-    seen = {str(item.get("id") or "").strip() for item in data if isinstance(item, dict)}
-    dynamic_models: set[str] = set()
-    accounts = account_service.list_accounts()
-    web_image_accounts = [
-        account
-        for account in accounts
-        if isinstance(account, dict)
-    ]
-    codex_types = {
-        normalized
-        for account in accounts
-        if isinstance(account, dict)
-           and account_service._normalize_source_type(account.get("source_type")) == "codex"
-           and (normalized := account_service._normalize_account_type(account.get("type")))
-    }
-
-    if web_image_accounts:
-        dynamic_models.add("gpt-image-2")
-    if codex_types & {"Plus", "Team", "Pro"}:
-        dynamic_models.add(CODEX_IMAGE_MODEL)
-    if "Plus" in codex_types:
-        dynamic_models.add(f"plus-{CODEX_IMAGE_MODEL}")
-    if "Team" in codex_types:
-        dynamic_models.add(f"team-{CODEX_IMAGE_MODEL}")
-    if "Pro" in codex_types:
-        dynamic_models.add(f"pro-{CODEX_IMAGE_MODEL}")
-
-    for model in sorted(dynamic_models):
-        if model not in seen:
-            data.append({
-                "id": model,
-                "object": "model",
-                "created": 0,
-                "owned_by": "chatgpt2api",
-                "permission": [],
-                "root": model,
-                "parent": None,
-            })
-    return result
+from __future__ import annotations
+
+from typing import Any
+
+from services.account_service import account_service
+from services.model_service import model_catalog_service
+from utils.helper import CODEX_IMAGE_MODEL
+
+
+def list_models() -> dict[str, Any]:
+    result = model_catalog_service.list_models()
+    data = result.get("data")
+    if not isinstance(data, list):
+        return result
+    seen = {str(item.get("id") or "").strip() for item in data if isinstance(item, dict)}
+    dynamic_models: set[str] = set()
+    accounts = account_service.list_accounts()
+    web_image_accounts = [
+        account
+        for account in accounts
+        if isinstance(account, dict)
+    ]
+    codex_types = {
+        normalized
+        for account in accounts
+        if isinstance(account, dict)
+           and account_service._normalize_source_type(account.get("source_type")) == "codex"
+           and (normalized := account_service._normalize_account_type(account.get("type")))
+    }
+
+    if web_image_accounts:
+        dynamic_models.add("gpt-image-2")
+    if codex_types & {"Plus", "Team", "Pro"}:
+        dynamic_models.add(CODEX_IMAGE_MODEL)
+    if "Plus" in codex_types:
+        dynamic_models.add(f"plus-{CODEX_IMAGE_MODEL}")
+    if "Team" in codex_types:
+        dynamic_models.add(f"team-{CODEX_IMAGE_MODEL}")
+    if "Pro" in codex_types:
+        dynamic_models.add(f"pro-{CODEX_IMAGE_MODEL}")
+
+    for model in sorted(dynamic_models):
+        if model not in seen:
+            data.append({
+                "id": model,
+                "object": "model",
+                "created": 0,
+                "owned_by": "chatgpt2api",
+                "permission": [],
+                "root": model,
+                "parent": None,
+            })
+    return result