veadk-python 0.2.12__py3-none-any.whl → 0.2.14__py3-none-any.whl

{veadk_python-0.2.12.dist-info → veadk_python-0.2.14.dist-info}/RECORD

@@ -2,10 +2,10 @@ veadk/__init__.py,sha256=9l1lyb9ifhHQeetmIBWZnIdwUCVyMyz1EnKsKz8BBG8,1135
 veadk/agent.py,sha256=4ijmzPXvY9iAx4s2NqBzwJ0VEYSWiUezeP2OvQFO9ek,11108
 veadk/agent_builder.py,sha256=HI7mRrUZ72_7i-jVzNDx1anTZHy7UxX7jr2Drxxx6j8,3031
 veadk/config.py,sha256=Ezl9Lna9iriC_Uf7m1ZXTWzylLyd7YspUFAQqh94Ong,3203
-veadk/consts.py,sha256=LTl4NQYJf7C8EvconGa96NRschzZZmmOqxlEikJe2Nk,2831
+veadk/consts.py,sha256=8FcwnvKxQM50r8HDQxnhI-Ml_mjyfYdkGDxbyPBrW5Q,2679
 veadk/runner.py,sha256=_DGNwX-t3sHJFJvHs-rRHXbjCZza8I_zU8AN3Fw5nRY,14217
 veadk/types.py,sha256=zOOzG-QJy-MkzHeicWJzy2_L5U4ERrWziPubIUEbd8c,1656
-veadk/version.py,sha256=GswH4BBfIAlEQ2fTdwoeOJftSsPugnneZH-RnSuSYN4,654
+veadk/version.py,sha256=AKkIbQyneK3kC3HOfdbOb8RMQ58A-Ly6yj_PKz8jElQ,654
 veadk/a2a/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
 veadk/a2a/agent_card.py,sha256=lhtgW1acMpxYUdULHEZwVFXOi6Xh4lNkf4S7QIhbFFI,1525
 veadk/a2a/remote_ve_agent.py,sha256=L2nzT8PlDI-lLtcaTJqk-D2Uvw9beKl8OEUqp-8qCbA,3510
@@ -28,7 +28,8 @@ veadk/auth/veauth/prompt_pilot_veauth.py,sha256=cls1LK2Un4cOMfHdaAqRhDHIXuk7cTuA
 veadk/auth/veauth/utils.py,sha256=cVEKWQZeX5fzx3JLB1odv59D8lhOAF1Pb3rsgO6evmM,2152
 veadk/auth/veauth/vesearch_veauth.py,sha256=rgup3VBbRSLligrsDFOEwpneq1BEtFwf9xpgNFWHKqc,2008
 veadk/cli/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
-veadk/cli/cli.py,sha256=-IGJKG9bGZpZXdLirOHv0LlpN9Vo42qzeEyNnZKsKv8,1430
+veadk/cli/cli.py,sha256=Iti595PE0g_R04njivyl0_fgNlPEYhCP-9j9ZkN80vQ,1496
+veadk/cli/cli_create.py,sha256=1qsTCww60rotvMPi_ASjKEe9nAoqxtEQ6mo1d7jDJnE,3949
 veadk/cli/cli_deploy.py,sha256=-P4PmXLGByypXGksshBT7BQ0U42hIvlHibXd_k4YfhQ,5328
 veadk/cli/cli_eval.py,sha256=TVSraCTyTxo_pLu5fhtk3TiZUOZkN3G2BLam1ybFXBc,5446
 veadk/cli/cli_init.py,sha256=f2A3RwUj9pApmUTl6FHmMwTTwyKl83pkvZRorTgl-XM,3982
@@ -105,7 +106,7 @@ veadk/integrations/ve_prompt_pilot/ve_prompt_pilot.py,sha256=Wp16pejFObnfudzNv1A
 veadk/integrations/ve_tls/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
 veadk/integrations/ve_tls/utils.py,sha256=DmWS37gSuKsUgiqlLyk1obfAAZZnEpkemQuoTDxAEG4,3769
 veadk/integrations/ve_tls/ve_tls.py,sha256=zO8rSX0CqxcjKjosXI90ykNFO6dhpsIsPa9uZGrf50s,7253
-veadk/integrations/ve_tos/ve_tos.py,sha256=yjc3h6JH2MM_8InE42YqjZupu5pcfFwHAMqPhr3x544,25186
+veadk/integrations/ve_tos/ve_tos.py,sha256=8tWb-eO7IK140VwjI-Bl4LNCeQNI9CgvKsr9gLXe5aA,25365
 veadk/integrations/ve_viking_db_memory/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
 veadk/integrations/ve_viking_db_memory/ve_viking_db_memory.py,sha256=gX5asox3VDra6o1BIg7AickPec0-Vj8_lgZ1zz5ue5Y,10165
 veadk/knowledgebase/__init__.py,sha256=k-5WM0DAqNq94Dg64Rl12ozoS1BnE9i-MY_ZOR2s2CQ,705
@@ -116,10 +117,10 @@ veadk/knowledgebase/backends/base_backend.py,sha256=gqkJNIASc-ZrJ2e4enIeUvI9CWuc
 veadk/knowledgebase/backends/in_memory_backend.py,sha256=uC4_LCkjp9DFQzTgTKzEr6yw1hcNUpZ0h7WSR0fpifc,3332
 veadk/knowledgebase/backends/opensearch_backend.py,sha256=4oChkqd25RvL460p1-D7noxQIoZVlD03aB61pKREqNw,4953
 veadk/knowledgebase/backends/redis_backend.py,sha256=pc-T_QicyAmOp6KbMz0V8iIL1ZlKCKQpdNB9SkaNqm8,5226
-veadk/knowledgebase/backends/utils.py,sha256=GOIuOFIJqruIk5Imx4F7ou5qKTUmg_gWJoLYi1kkopQ,2661
-veadk/knowledgebase/backends/vikingdb_knowledge_backend.py,sha256=hVSfAc17wzEoEKlPa7x3ekCcApFgDzY7x6ceBDxd9mw,18126
+veadk/knowledgebase/backends/utils.py,sha256=383La0fsmPcbJEpGWk54EP8p8QO1SP_ThqJPmRWa5Rs,2705
+veadk/knowledgebase/backends/vikingdb_knowledge_backend.py,sha256=khk33XDDdeK1tb84lHtTAjtuZQLmubZFo3shyd70YLs,19245
 veadk/memory/__init__.py,sha256=GRIoRqapS2gUtRMB9JZvodU29PwMxFe2KDjPnVpqib0,1252
-veadk/memory/long_term_memory.py,sha256=YE5DH9qyJwlA7t0smwV4J90wnUBYVR6MRXHg1qsqKNU,7423
+veadk/memory/long_term_memory.py,sha256=122uacyNsKkUIfsc5Vru8lMGsJKE0A7JJqddLaxX-ls,7838
 veadk/memory/short_term_memory.py,sha256=MLR2Op_l_Zr-ZKCJ7pyNhe_OzPG06Zcb8pt8qA_zrFY,4922
 veadk/memory/short_term_memory_processor.py,sha256=vqmYSUNFGXEogJ8ZxIAhSSrv66LGT8eEHCzxF9h07LA,3276
 veadk/memory/long_term_memory_backends/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
@@ -128,7 +129,7 @@ veadk/memory/long_term_memory_backends/in_memory_backend.py,sha256=0IKRqstS3Rtm8
 veadk/memory/long_term_memory_backends/mem0_backend.py,sha256=sj6OxMHA4oG6kuzTMK5zsGEIIa0-ESPrRYv6dR3KI8E,4869
 veadk/memory/long_term_memory_backends/opensearch_backend.py,sha256=u5odPUv0fchH3IsZNTH0-nNpdW4KwkWojRQi3cn52Wo,4567
 veadk/memory/long_term_memory_backends/redis_backend.py,sha256=0XX21AJA4UwEqrteXdgOr5WAfuo31wK1fMLUtXMssfg,5262
-veadk/memory/long_term_memory_backends/vikingdb_memory_backend.py,sha256=EDhpMCH92lspykzRzQpySfD4G5LndANAf6B5eMYzYH8,5793
+veadk/memory/long_term_memory_backends/vikingdb_memory_backend.py,sha256=TfAtQXiJ-iZLRKPTBMakQTwRoHdDsN0l2R5udNFJatY,6683
 veadk/memory/short_term_memory_backends/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
 veadk/memory/short_term_memory_backends/base_backend.py,sha256=5AtnWuFE_TYpTAlAKnm2xglf4AT3iXmU9QSF7LkO1BA,1053
 veadk/memory/short_term_memory_backends/mysql_backend.py,sha256=_m-yuB7X-pPvs93-W8ZswDIkIpYpCFMRRdt9D5SZzn8,1513
@@ -143,13 +144,13 @@ veadk/tools/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
 veadk/tools/demo_tools.py,sha256=Gu3sxygcYVS2cv3WqUOl-Gq4JhMlDAktoCHOFT0gbFQ,2216
 veadk/tools/load_knowledgebase_tool.py,sha256=UUTv0Za9GkEXAkl1SXmyq0HGCKGvSlH_f8Ok6O6e52M,4704
 veadk/tools/builtin_tools/__init__.py,sha256=pkSabKw7_ai4NOo56pXKL40EcaxIDh6HYxPXOY7qWbo,634
-veadk/tools/builtin_tools/generate_image.py,sha256=aDMlR-IrxUMepIl9hPkywlH-4e7uIRiyFzLTtmezOnw,17495
 veadk/tools/builtin_tools/image_edit.py,sha256=KslsuabBchAYR3ZrWSO5viEe5ORUAe0GI1qQ6mxoIU0,11588
-veadk/tools/builtin_tools/image_generate.py,sha256=-L_3k3KRJ_arljLfKz946fbd9ppxTDNvJmkNGhgj6qQ,11357
+veadk/tools/builtin_tools/image_generate.py,sha256=frBUYEmizq6B4cRIUTfWKpLF2RywposJf-hZ8kkZvV8,18075
 veadk/tools/builtin_tools/lark.py,sha256=b2IWsN8fZFh9aweSGznaOqA30TCOLpVjNCDNa1LHZl4,2046
 veadk/tools/builtin_tools/las.py,sha256=rgKfnK5GsHVbmkp-rc7rtCvWg-yYNxMjeV0ayCyRpjM,913
 veadk/tools/builtin_tools/load_knowledgebase.py,sha256=Xqtq25DL720goRegCVmmkpH2Ye2VWLcrF5ncC37gK_Y,3427
 veadk/tools/builtin_tools/mcp_router.py,sha256=l3xcIHAHQ0AGCZG3mYyhwM0btqEMDe4TY2S-UYUM8M0,883
+veadk/tools/builtin_tools/run_code.py,sha256=C2CRMNZYgKoyVpe4JoGtUa-Xfb48IuyNIHzF5WKJfxw,2623
 veadk/tools/builtin_tools/vesearch.py,sha256=prPP0w6lYeIEPwuZdmV00RAzaW4MeH8lYtK-NluaXtU,1748
 veadk/tools/builtin_tools/video_generate.py,sha256=hlvwoLESUV8vOPiNFVNPF0ithWqH7N5c6ElMvyI-lBM,16101
 veadk/tools/builtin_tools/web_scraper.py,sha256=iVnxWVf2mVgOnEOeQ6Bg5ATYN-g1ZPCTK6VJm710be0,2408
@@ -180,9 +181,9 @@ veadk/utils/mcp_utils.py,sha256=aET7pX3LXmRe2-Jh7_xRvxrVyl1dN7uPAUk16luwMlQ,1525
 veadk/utils/misc.py,sha256=ghEqrqoDfKrW9ZD3IB0bwcfyyB0gRWN2yEP9eRxQ4nE,4953
 veadk/utils/patches.py,sha256=dcHdlJ8IciyMjDuMy6-_6McUqJYyLz0yHmJ0xH8lWOw,2752
 veadk/utils/volcengine_sign.py,sha256=3xn6ca2OAg_AFyP2dqFTSioqkeDel_BoKURUtCcO-EQ,6736
-veadk_python-0.2.12.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-veadk_python-0.2.12.dist-info/METADATA,sha256=yucua-pOLLODTgl43PDkHes4ZNkn2pRRRGyRw1J-x2Q,18428
-veadk_python-0.2.12.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-veadk_python-0.2.12.dist-info/entry_points.txt,sha256=-g28D6dNV-2UvAiRP9VF0oOVSDSJ5zlLUIZ34ArAqF8,46
-veadk_python-0.2.12.dist-info/top_level.txt,sha256=Qqi3ycJ4anKiZWBXtUBIy8zK9ZuXJsFa05oFq8O8qqY,6
-veadk_python-0.2.12.dist-info/RECORD,,
+veadk_python-0.2.14.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+veadk_python-0.2.14.dist-info/METADATA,sha256=Gk-dXheONF6hhaRA3yMsOON7aqet5oSPXV9bZe2ehY4,18428
+veadk_python-0.2.14.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+veadk_python-0.2.14.dist-info/entry_points.txt,sha256=-g28D6dNV-2UvAiRP9VF0oOVSDSJ5zlLUIZ34ArAqF8,46
+veadk_python-0.2.14.dist-info/top_level.txt,sha256=Qqi3ycJ4anKiZWBXtUBIy8zK9ZuXJsFa05oFq8O8qqY,6
+veadk_python-0.2.14.dist-info/RECORD,,

veadk/tools/builtin_tools/generate_image.py

@@ -1,430 +0,0 @@
-# Copyright (c) 2025 Beijing Volcano Engine Technology Co., Ltd. and/or its affiliates.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import base64
-import json
-import mimetypes
-import traceback
-from typing import Dict
-
-from google.adk.tools import ToolContext
-from google.genai.types import Blob, Part
-from opentelemetry import trace
-from opentelemetry.trace import Span
-from volcenginesdkarkruntime import Ark
-from volcenginesdkarkruntime.types.images.images import SequentialImageGenerationOptions
-
-from veadk.config import getenv, settings
-from veadk.consts import (
-    DEFAULT_IMAGE_GENERATE_MODEL_NAME,
-    DEFAULT_IMAGE_GENERATE_MODEL_API_BASE,
-)
-from veadk.utils.logger import get_logger
-from veadk.utils.misc import formatted_timestamp, read_file_to_bytes
-from veadk.version import VERSION
-import asyncio
-import concurrent.futures
-import contextvars
-
-
-logger = get_logger(__name__)
-
-client = Ark(
-    api_key=getenv(
-        "MODEL_IMAGE_API_KEY", getenv("MODEL_AGENT_API_KEY", settings.model.api_key)
-    ),
-    base_url=getenv("MODEL_IMAGE_API_BASE", DEFAULT_IMAGE_GENERATE_MODEL_API_BASE),
-)
-
-executor = concurrent.futures.ThreadPoolExecutor(max_workers=8)
-tracer = trace.get_tracer("gcp.vertex.agent")
-
-
-def _build_input_parts(item: dict, task_type: str, image_field):
-    input_part = {"role": "user"}
-    input_part["parts.0.type"] = "text"
-    input_part["parts.0.text"] = json.dumps(item, ensure_ascii=False)
-
-    if image_field:
-        if task_type.startswith("single"):
-            assert isinstance(image_field, str), (
-                f"single_* task_type image must be str, got {type(image_field)}"
-            )
-            input_part["parts.1.type"] = "image_url"
-            input_part["parts.1.image_url.name"] = "origin_image"
-            input_part["parts.1.image_url.url"] = image_field
-        elif task_type.startswith("multi"):
-            assert isinstance(image_field, list), (
-                f"multi_* task_type image must be list, got {type(image_field)}"
-            )
-            assert len(image_field) <= 10, (
-                f"multi_* task_type image list length must be <= 10, got {len(image_field)}"
-            )
-            for i, image_url in enumerate(image_field):
-                idx = i + 1
-                input_part[f"parts.{idx}.type"] = "image_url"
-                input_part[f"parts.{idx}.image_url.name"] = f"origin_image_{i}"
-                input_part[f"parts.{idx}.image_url.url"] = image_url
-
-    return input_part
-
-
-def handle_single_task_sync(
-    idx: int, item: dict, tool_context
-) -> tuple[list[dict], list[str]]:
-    logger.debug(f"handle_single_task_sync item {idx}: {item}")
-    success_list: list[dict] = []
-    error_list: list[str] = []
-    total_tokens = 0
-    output_tokens = 0
-    output_part = {"message.role": "model"}
-
-    task_type = item.get("task_type", "text_to_single")
-    prompt = item.get("prompt", "")
-    response_format = item.get("response_format", None)
-    size = item.get("size", None)
-    watermark = item.get("watermark", None)
-    image_field = item.get("image", None)
-    sequential_image_generation = item.get("sequential_image_generation", None)
-    max_images = item.get("max_images", None)
-
-    input_part = _build_input_parts(item, task_type, image_field)
-
-    inputs = {"prompt": prompt}
-    if size:
-        inputs["size"] = size
-    if response_format:
-        inputs["response_format"] = response_format
-    if watermark is not None:
-        inputs["watermark"] = watermark
-    if sequential_image_generation:
-        inputs["sequential_image_generation"] = sequential_image_generation
-
-    with tracer.start_as_current_span(f"call_llm_task_{idx}") as span:
-        try:
-            if (
-                sequential_image_generation
-                and sequential_image_generation == "auto"
-                and max_images
-            ):
-                response = client.images.generate(
-                    model=getenv("MODEL_IMAGE_NAME", DEFAULT_IMAGE_GENERATE_MODEL_NAME),
-                    **inputs,
-                    sequential_image_generation_options=SequentialImageGenerationOptions(
-                        max_images=max_images
-                    ),
-                )
-            else:
-                response = client.images.generate(
-                    model=getenv("MODEL_IMAGE_NAME", DEFAULT_IMAGE_GENERATE_MODEL_NAME),
-                    **inputs,
-                )
-
-            if not response.error:
-                logger.debug(f"task {idx} Image generate response: {response}")
-
-                total_tokens += getattr(response.usage, "total_tokens", 0) or 0
-                output_tokens += getattr(response.usage, "output_tokens", 0) or 0
-
-                for i, image_data in enumerate(response.data):
-                    image_name = f"task_{idx}_image_{i}"
-                    if "error" in image_data:
-                        logger.error(f"Image {image_name} error: {image_data.error}")
-                        error_list.append(image_name)
-                        continue
-
-                    if getattr(image_data, "url", None):
-                        image_url = image_data.url
-                    else:
-                        b64 = getattr(image_data, "b64_json", None)
-                        if not b64:
-                            logger.error(
-                                f"Image {image_name} missing data (no url/b64)"
-                            )
-                            error_list.append(image_name)
-                            continue
-                        image_bytes = base64.b64decode(b64)
-                        image_url = _upload_image_to_tos(
-                            image_bytes=image_bytes, object_key=f"{image_name}.png"
-                        )
-                        if not image_url:
-                            logger.error(f"Upload image to TOS failed: {image_name}")
-                            error_list.append(image_name)
-                            continue
-                        logger.debug(f"Image saved as ADK artifact: {image_name}")
-
-                    tool_context.state[f"{image_name}_url"] = image_url
-                    output_part[f"message.parts.{i}.type"] = "image_url"
-                    output_part[f"message.parts.{i}.image_url.name"] = image_name
-                    output_part[f"message.parts.{i}.image_url.url"] = image_url
-                    logger.debug(
-                        f"Image {image_name} generated successfully: {image_url}"
-                    )
-                    success_list.append({image_name: image_url})
-            else:
-                logger.error(
-                    f"Task {idx} No images returned by model: {response.error}"
-                )
-                error_list.append(f"task_{idx}")
-
-        except Exception as e:
-            logger.error(f"Error in task {idx}: {e}")
-            traceback.print_exc()
-            error_list.append(f"task_{idx}")
-
-        finally:
-            add_span_attributes(
-                span,
-                tool_context,
-                input_part=input_part,
-                output_part=output_part,
-                output_tokens=output_tokens,
-                total_tokens=total_tokens,
-                request_model=getenv(
-                    "MODEL_IMAGE_NAME", DEFAULT_IMAGE_GENERATE_MODEL_NAME
-                ),
-                response_model=getenv(
-                    "MODEL_IMAGE_NAME", DEFAULT_IMAGE_GENERATE_MODEL_NAME
-                ),
-            )
-    logger.debug(
-        f"task {idx} Image generate success_list: {success_list}\nerror_list: {error_list}"
-    )
-    return success_list, error_list
-
-
-async def image_generate(tasks: list[dict], tool_context) -> Dict:
-    """
-    Seedream 4.0: batch image generation via tasks.
-    Args:
-        tasks (list[dict]):
-            A list of image-generation tasks. Each task is a dict.
-    Per-task schema
-    ---------------
-    Required:
-        - task_type (str):
-            One of:
-              * "multi_image_to_group"   # 多图生组图
-              * "single_image_to_group"  # 单图生组图
-              * "text_to_group"          # 文生组图
-              * "multi_image_to_single"  # 多图生单图
-              * "single_image_to_single" # 单图生单图
-              * "text_to_single"         # 文生单图
-        - prompt (str)
-            Text description of the desired image(s). 中文/English 均可。
-            若要指定生成图片的数量，请在prompt中添加"生成N张图片"，其中N为具体的数字。
-    Optional:
-        - size (str)
-            指定生成图像的大小，有两种用法（二选一，不可混用）：
-            方式 1：分辨率级别
-                可选值: "1K", "2K", "4K"
-                模型会结合 prompt 中的语义推断合适的宽高比、长宽。
-            方式 2：具体宽高值
-                格式: "<宽度>x<高度>"，如 "2048x2048", "2384x1728"
-                约束:
-                    * 总像素数范围: [1024x1024, 4096x4096]
-                    * 宽高比范围: [1/16, 16]
-                推荐值:
-                    - 1:1   → 2048x2048
-                    - 4:3   → 2384x1728
-                    - 3:4   → 1728x2304
-                    - 16:9  → 2560x1440
-                    - 9:16  → 1440x2560
-                    - 3:2   → 2496x1664
-                    - 2:3   → 1664x2496
-                    - 21:9  → 3024x1296
-            默认值: "2048x2048"
-        - response_format (str)
-            Return format: "url" (default, URL 24h 过期) | "b64_json".
-        - watermark (bool)
-            Add watermark. Default: true.
-        - image (str | list[str])   # 仅“非文生图”需要。文生图请不要提供 image
-            Reference image(s) as URL or Base64.
-            * 生成“单图”的任务：传入 string（exactly 1 image）。
-            * 生成“组图”的任务：传入 array（2–10 images）。
-        - sequential_image_generation (str)
-            控制是否生成“组图”。Default: "disabled".
-            * 若要生成组图：必须设为 "auto"。
-        - max_images (int)
-            仅当生成组图时生效。控制模型能生成的最多张数，范围 [1, 15]， 不设置默认为15。
-            注意这个参数不等于生成的图片数量，而是模型最多能生成的图片数量。
-            在单图组图场景最多 14；多图组图场景需满足 (len(images)+max_images ≤ 15)。
-    Model 行为说明（如何由参数推断模式）
-    ---------------------------------
-    1) 文生单图: 不提供 image 且 (S 未设置或 S="disabled") → 1 张图。
-    2) 文生组图: 不提供 image 且 S="auto" → 组图，数量由 max_images 控制。
-    3) 单图生单图: image=string 且 (S 未设置或 S="disabled") → 1 张图。
-    4) 单图生组图: image=string 且 S="auto" → 组图，数量 ≤14。
-    5) 多图生单图: image=array (2–10) 且 (S 未设置或 S="disabled") → 1 张图。
-    6) 多图生组图: image=array (2–10) 且 S="auto" → 组图，需满足总数 ≤15。
-    返回结果
-    --------
-        Dict with generation summary.
-        Example:
-        {
-            "status": "success",
-            "success_list": [
-                {"image_name": "url"}
-            ],
-            "error_list": ["image_name"]
-        }
-    Notes:
-    - 组图任务必须 sequential_image_generation="auto"。
-    - 如果想要指定生成组图的数量，请在prompt里添加数量说明，例如："生成3张图片"。
-    - size 推荐使用 2048x2048 或表格里的标准比例，确保生成质量。
-    """
-    logger.debug(
-        f"Using model: {getenv('MODEL_IMAGE_NAME', DEFAULT_IMAGE_GENERATE_MODEL_NAME)}"
-    )
-    success_list: list[dict] = []
-    error_list: list[str] = []
-    logger.debug(f"image_generate tasks: {tasks}")
-    with tracer.start_as_current_span("image_generate"):
-        base_ctx = contextvars.copy_context()
-
-        def make_task(idx, item):
-            ctx = base_ctx.copy()
-            return lambda: ctx.run(handle_single_task_sync, idx, item, tool_context)
-
-        loop = asyncio.get_event_loop()
-        futures = [
-            loop.run_in_executor(executor, make_task(idx, item))
-            for idx, item in enumerate(tasks)
-        ]
-
-        results = await asyncio.gather(*futures, return_exceptions=True)
-
-        for res in results:
-            if isinstance(res, Exception):
-                logger.error(f"Task raised exception: {res}")
-                error_list.append("unknown_task_exception")
-                continue
-            s, e = res
-            success_list.extend(s)
-            error_list.extend(e)
-
-    if not success_list:
-        logger.debug(
-            f"image_generate success_list: {success_list}\nerror_list: {error_list}"
-        )
-        return {
-            "status": "error",
-            "success_list": success_list,
-            "error_list": error_list,
-        }
-    app_name = tool_context._invocation_context.app_name
-    user_id = tool_context._invocation_context.user_id
-    session_id = tool_context._invocation_context.session.id
-    artifact_service = tool_context._invocation_context.artifact_service
-
-    if artifact_service:
-        for image in success_list:
-            for _, image_tos_url in image.items():
-                filename = f"artifact_{formatted_timestamp()}"
-                await artifact_service.save_artifact(
-                    app_name=app_name,
-                    user_id=user_id,
-                    session_id=session_id,
-                    filename=filename,
-                    artifact=Part(
-                        inline_data=Blob(
-                            display_name=filename,
-                            data=read_file_to_bytes(image_tos_url),
-                            mime_type=mimetypes.guess_type(image_tos_url)[0],
-                        )
-                    ),
-                )
-
-    logger.debug(
-        f"image_generate success_list: {success_list}\nerror_list: {error_list}"
-    )
-    return {"status": "success", "success_list": success_list, "error_list": error_list}
-
-
-def add_span_attributes(
-    span: Span,
-    tool_context: ToolContext,
-    input_part: dict = None,
-    output_part: dict = None,
-    input_tokens: int = None,
-    output_tokens: int = None,
-    total_tokens: int = None,
-    request_model: str = None,
-    response_model: str = None,
-):
-    try:
-        # common attributes
-        app_name = tool_context._invocation_context.app_name
-        user_id = tool_context._invocation_context.user_id
-        agent_name = tool_context.agent_name
-        session_id = tool_context._invocation_context.session.id
-        span.set_attribute("gen_ai.agent.name", agent_name)
-        span.set_attribute("openinference.instrumentation.veadk", VERSION)
-        span.set_attribute("gen_ai.app.name", app_name)
-        span.set_attribute("gen_ai.user.id", user_id)
-        span.set_attribute("gen_ai.session.id", session_id)
-        span.set_attribute("agent_name", agent_name)
-        span.set_attribute("agent.name", agent_name)
-        span.set_attribute("app_name", app_name)
-        span.set_attribute("app.name", app_name)
-        span.set_attribute("user.id", user_id)
-        span.set_attribute("session.id", session_id)
-        span.set_attribute("cozeloop.report.source", "veadk")
-
-        # llm attributes
-        span.set_attribute("gen_ai.system", "openai")
-        span.set_attribute("gen_ai.operation.name", "chat")
-        if request_model:
-            span.set_attribute("gen_ai.request.model", request_model)
-        if response_model:
-            span.set_attribute("gen_ai.response.model", response_model)
-        if total_tokens:
-            span.set_attribute("gen_ai.usage.total_tokens", total_tokens)
-        if output_tokens:
-            span.set_attribute("gen_ai.usage.output_tokens", output_tokens)
-        if input_tokens:
-            span.set_attribute("gen_ai.usage.input_tokens", input_tokens)
-        if input_part:
-            span.add_event("gen_ai.user.message", input_part)
-        if output_part:
-            span.add_event("gen_ai.choice", output_part)
-
-    except Exception:
-        traceback.print_exc()
-
-
-def _upload_image_to_tos(image_bytes: bytes, object_key: str) -> None:
-    try:
-        import os
-        from datetime import datetime
-
-        from veadk.integrations.ve_tos.ve_tos import VeTOS
-
-        timestamp: str = datetime.now().strftime("%Y%m%d%H%M%S%f")[:-3]
-        object_key = f"{timestamp}-{object_key}"
-        bucket_name = os.getenv("DATABASE_TOS_BUCKET")
-        ve_tos = VeTOS()
-
-        tos_url = ve_tos.build_tos_signed_url(
-            object_key=object_key, bucket_name=bucket_name
-        )
-
-        ve_tos.upload_bytes(
-            data=image_bytes, object_key=object_key, bucket_name=bucket_name
-        )
-
-        return tos_url
-    except Exception as e:
-        logger.error(f"Upload to TOS failed: {e}")
-        return None