Merge main into feat/plugin

.github/workflows/web-tests.yml

@@ -0,0 +1,46 @@
+name: Web Tests
+
+on:
+  pull_request:
+    branches:
+      - main
+    paths:
+      - web/**
+
+concurrency:
+  group: web-tests-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  test:
+    name: Web Tests
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: ./web
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Check changed files
+        id: changed-files
+        uses: tj-actions/changed-files@v45
+        with:
+          files: web/**
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        if: steps.changed-files.outputs.any_changed == 'true'
+        with:
+          node-version: 20
+          cache: yarn
+          cache-dependency-path: ./web/package.json
+
+      - name: Install dependencies
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: yarn install --frozen-lockfile
+
+      - name: Run tests
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: yarn test

api/app.py

@@ -53,11 +53,9 @@ from services.account_service import AccountService
 
 warnings.simplefilter("ignore", ResourceWarning)
 
-# fix windows platform
-if os.name == "nt":
-    os.system('tzutil /s "UTC"')
-else:
-    os.environ["TZ"] = "UTC"
+os.environ["TZ"] = "UTC"
+# windows platform not support tzset
+if hasattr(time, "tzset"):
     time.tzset()
 
 

api/core/app/apps/base_app_runner.py

@@ -309,7 +309,7 @@ class AppRunner:
             if not prompt_messages:
                 prompt_messages = result.prompt_messages
 
-            if not usage and result.delta.usage:
+            if result.delta.usage:
                 usage = result.delta.usage
 
         if not usage:

api/core/embedding/cached_embedding.py

@@ -5,6 +5,7 @@ from typing import Optional, cast
 import numpy as np
 from sqlalchemy.exc import IntegrityError
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_manager import ModelInstance
 from core.model_runtime.entities.model_entities import ModelPropertyKey
 from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel
@@ -56,7 +57,9 @@ class CacheEmbedding(Embeddings):
                 for i in range(0, len(embedding_queue_texts), max_chunks):
                     batch_texts = embedding_queue_texts[i : i + max_chunks]
 
-                    embedding_result = self._model_instance.invoke_text_embedding(texts=batch_texts, user=self._user)
+                    embedding_result = self._model_instance.invoke_text_embedding(
+                        texts=batch_texts, user=self._user, input_type=EmbeddingInputType.DOCUMENT
+                    )
 
                     for vector in embedding_result.embeddings:
                         try:
@@ -100,7 +103,9 @@ class CacheEmbedding(Embeddings):
             redis_client.expire(embedding_cache_key, 600)
             return list(np.frombuffer(base64.b64decode(embedding), dtype="float"))
         try:
-            embedding_result = self._model_instance.invoke_text_embedding(texts=[text], user=self._user)
+            embedding_result = self._model_instance.invoke_text_embedding(
+                texts=[text], user=self._user, input_type=EmbeddingInputType.QUERY
+            )
 
             embedding_results = embedding_result.embeddings[0]
             embedding_results = (embedding_results / np.linalg.norm(embedding_results)).tolist()

api/core/embedding/embedding_constant.py

@@ -0,0 +1,10 @@
+from enum import Enum
+
+
+class EmbeddingInputType(Enum):
+    """
+    Enum for embedding input type.
+    """
+
+    DOCUMENT = "document"
+    QUERY = "query"

api/core/model_manager.py

@@ -3,6 +3,7 @@ import os
 from collections.abc import Callable, Generator, Sequence
 from typing import IO, Literal, Optional, Union, cast, overload
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.entities.provider_configuration import ProviderConfiguration, ProviderModelBundle
 from core.entities.provider_entities import ModelLoadBalancingConfiguration
 from core.errors.error import ProviderTokenNotInitError
@@ -194,12 +195,15 @@ class ModelInstance:
             tools=tools,
         )
 
-    def invoke_text_embedding(self, texts: list[str], user: Optional[str] = None) -> TextEmbeddingResult:
+    def invoke_text_embedding(
+        self, texts: list[str], user: Optional[str] = None, input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT
+    ) -> TextEmbeddingResult:
         """
         Invoke large language model
 
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         if not isinstance(self.model_type_instance, TextEmbeddingModel):
@@ -212,6 +216,7 @@ class ModelInstance:
             credentials=self.credentials,
             texts=texts,
             user=user,
+            input_type=input_type,
         )
 
     def get_text_embedding_num_tokens(self, texts: list[str]) -> int:

api/core/model_runtime/model_providers/__base/text_embedding_model.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from pydantic import ConfigDict
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import ModelPropertyKey, ModelType
 from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult
 from core.model_runtime.model_providers.__base.ai_model import AIModel
@@ -20,35 +21,47 @@ class TextEmbeddingModel(AIModel):
     model_config = ConfigDict(protected_namespaces=())
 
     def invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
-        Invoke large language model
+        Invoke text embedding model
 
         :param model: model name
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         self.started_at = time.perf_counter()
 
         try:
-            return self._invoke(model, credentials, texts, user)
+            return self._invoke(model, credentials, texts, user, input_type)
         except Exception as e:
             raise self._transform_invoke_error(e)
 
     @abstractmethod
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
-        Invoke large language model
+        Invoke text embedding model
 
         :param model: model name
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         raise NotImplementedError

api/core/model_runtime/model_providers/azure_openai/text_embedding/text_embedding.py

@@ -7,6 +7,7 @@ import numpy as np
 import tiktoken
 from openai import AzureOpenAI
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import AIModelEntity, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.validate import CredentialsValidateFailedError
@@ -17,8 +18,23 @@ from core.model_runtime.model_providers.azure_openai._constant import EMBEDDING_
 
 class AzureOpenAITextEmbeddingModel(_CommonAzureOpenAI, TextEmbeddingModel):
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
+        """
+        Invoke text embedding model
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
+        :return: embeddings result
+        """
         base_model_name = credentials["base_model_name"]
         credentials_kwargs = self._to_credential_kwargs(credentials)
         client = AzureOpenAI(**credentials_kwargs)

api/core/model_runtime/model_providers/baichuan/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from requests import post
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -35,7 +36,12 @@ class BaichuanTextEmbeddingModel(TextEmbeddingModel):
     api_base: str = "http://api.baichuan-ai.com/v1/embeddings"
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -44,6 +50,7 @@ class BaichuanTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         api_key = credentials["api_key"]

api/core/model_runtime/model_providers/bedrock/text_embedding/text_embedding.py

@@ -13,6 +13,7 @@ from botocore.exceptions import (
     UnknownServiceError,
 )
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -30,7 +31,12 @@ logger = logging.getLogger(__name__)
 
 class BedrockTextEmbeddingModel(TextEmbeddingModel):
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -39,6 +45,7 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         client_config = Config(region_name=credentials["aws_region"])

api/core/model_runtime/model_providers/cohere/text_embedding/text_embedding.py

@@ -5,6 +5,7 @@ import cohere
 import numpy as np
 from cohere.core import RequestOptions
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -25,7 +26,12 @@ class CohereTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -34,6 +40,7 @@ class CohereTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         # get model properties

api/core/model_runtime/model_providers/fireworks/fireworks.yaml

@@ -15,6 +15,7 @@ help:
     en_US: https://fireworks.ai/account/api-keys
 supported_model_types:
   - llm
+  - text-embedding
 configurate_methods:
   - predefined-model
 provider_credential_schema:

api/core/model_runtime/model_providers/fireworks/text_embedding/UAE-Large-V1.yaml

@@ -0,0 +1,12 @@
+model: WhereIsAI/UAE-Large-V1
+label:
+  zh_Hans: UAE-Large-V1
+  en_US: UAE-Large-V1
+model_type: text-embedding
+model_properties:
+  context_size: 512
+  max_chunks: 1
+pricing:
+  input: '0.008'
+  unit: '0.000001'
+  currency: 'USD'

api/core/model_runtime/model_providers/fireworks/text_embedding/gte-base.yaml

@@ -0,0 +1,12 @@
+model: thenlper/gte-base
+label:
+  zh_Hans: GTE-base
+  en_US: GTE-base
+model_type: text-embedding
+model_properties:
+  context_size: 512
+  max_chunks: 1
+pricing:
+  input: '0.008'
+  unit: '0.000001'
+  currency: 'USD'

api/core/model_runtime/model_providers/fireworks/text_embedding/gte-large.yaml

@@ -0,0 +1,12 @@
+model: thenlper/gte-large
+label:
+  zh_Hans: GTE-large
+  en_US: GTE-large
+model_type: text-embedding
+model_properties:
+  context_size: 512
+  max_chunks: 1
+pricing:
+  input: '0.008'
+  unit: '0.000001'
+  currency: 'USD'

api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.5.yaml

@@ -0,0 +1,12 @@
+model: nomic-ai/nomic-embed-text-v1.5
+label:
+  zh_Hans: nomic-embed-text-v1.5
+  en_US: nomic-embed-text-v1.5
+model_type: text-embedding
+model_properties:
+  context_size: 8192
+  max_chunks: 16
+pricing:
+  input: '0.008'
+  unit: '0.000001'
+  currency: 'USD'

api/core/model_runtime/model_providers/fireworks/text_embedding/nomic-embed-text-v1.yaml

@@ -0,0 +1,12 @@
+model: nomic-ai/nomic-embed-text-v1
+label:
+  zh_Hans: nomic-embed-text-v1
+  en_US: nomic-embed-text-v1
+model_type: text-embedding
+model_properties:
+  context_size: 8192
+  max_chunks: 16
+pricing:
+  input: '0.008'
+  unit: '0.000001'
+  currency: 'USD'

api/core/model_runtime/model_providers/fireworks/text_embedding/text_embedding.py

@@ -0,0 +1,151 @@
+import time
+from collections.abc import Mapping
+from typing import Optional, Union
+
+import numpy as np
+from openai import OpenAI
+
+from core.embedding.embedding_constant import EmbeddingInputType
+from core.model_runtime.entities.model_entities import PriceType
+from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.__base.text_embedding_model import TextEmbeddingModel
+from core.model_runtime.model_providers.fireworks._common import _CommonFireworks
+
+
+class FireworksTextEmbeddingModel(_CommonFireworks, TextEmbeddingModel):
+    """
+    Model class for Fireworks text embedding model.
+    """
+
+    def _invoke(
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
+    ) -> TextEmbeddingResult:
+        """
+        Invoke text embedding model
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
+        :return: embeddings result
+        """
+
+        credentials_kwargs = self._to_credential_kwargs(credentials)
+        client = OpenAI(**credentials_kwargs)
+
+        extra_model_kwargs = {}
+        if user:
+            extra_model_kwargs["user"] = user
+
+        extra_model_kwargs["encoding_format"] = "float"
+
+        context_size = self._get_context_size(model, credentials)
+        max_chunks = self._get_max_chunks(model, credentials)
+
+        inputs = []
+        indices = []
+        used_tokens = 0
+
+        for i, text in enumerate(texts):
+            # Here token count is only an approximation based on the GPT2 tokenizer
+            # TODO: Optimize for better token estimation and chunking
+            num_tokens = self._get_num_tokens_by_gpt2(text)
+
+            if num_tokens >= context_size:
+                cutoff = int(np.floor(len(text) * (context_size / num_tokens)))
+                # if num tokens is larger than context length, only use the start
+                inputs.append(text[0:cutoff])
+            else:
+                inputs.append(text)
+            indices += [i]
+
+        batched_embeddings = []
+        _iter = range(0, len(inputs), max_chunks)
+
+        for i in _iter:
+            embeddings_batch, embedding_used_tokens = self._embedding_invoke(
+                model=model,
+                client=client,
+                texts=inputs[i : i + max_chunks],
+                extra_model_kwargs=extra_model_kwargs,
+            )
+            used_tokens += embedding_used_tokens
+            batched_embeddings += embeddings_batch
+
+        usage = self._calc_response_usage(model=model, credentials=credentials, tokens=used_tokens)
+        return TextEmbeddingResult(embeddings=batched_embeddings, usage=usage, model=model)
+
+    def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> int:
+        """
+        Get number of tokens for given prompt messages
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :return:
+        """
+        return sum(self._get_num_tokens_by_gpt2(text) for text in texts)
+
+    def validate_credentials(self, model: str, credentials: Mapping) -> None:
+        """
+        Validate model credentials
+
+        :param model: model name
+        :param credentials: model credentials
+        :return:
+        """
+        try:
+            # transform credentials to kwargs for model instance
+            credentials_kwargs = self._to_credential_kwargs(credentials)
+            client = OpenAI(**credentials_kwargs)
+
+            # call embedding model
+            self._embedding_invoke(model=model, client=client, texts=["ping"], extra_model_kwargs={})
+        except Exception as ex:
+            raise CredentialsValidateFailedError(str(ex))
+
+    def _embedding_invoke(
+        self, model: str, client: OpenAI, texts: Union[list[str], str], extra_model_kwargs: dict
+    ) -> tuple[list[list[float]], int]:
+        """
+        Invoke embedding model
+        :param model: model name
+        :param client: model client
+        :param texts: texts to embed
+        :param extra_model_kwargs: extra model kwargs
+        :return: embeddings and used tokens
+        """
+        response = client.embeddings.create(model=model, input=texts, **extra_model_kwargs)
+        return [data.embedding for data in response.data], response.usage.total_tokens
+
+    def _calc_response_usage(self, model: str, credentials: dict, tokens: int) -> EmbeddingUsage:
+        """
+        Calculate response usage
+
+        :param model: model name
+        :param credentials: model credentials
+        :param tokens: input tokens
+        :return: usage
+        """
+        input_price_info = self.get_price(
+            model=model, credentials=credentials, tokens=tokens, price_type=PriceType.INPUT
+        )
+
+        usage = EmbeddingUsage(
+            tokens=tokens,
+            total_tokens=tokens,
+            unit_price=input_price_info.unit_price,
+            price_unit=input_price_info.unit,
+            total_price=input_price_info.total_amount,
+            currency=input_price_info.currency,
+            latency=time.perf_counter() - self.started_at,
+        )
+
+        return usage

api/core/model_runtime/model_providers/huggingface_hub/text_embedding/text_embedding.py

@@ -6,6 +6,7 @@ import numpy as np
 import requests
 from huggingface_hub import HfApi, InferenceClient
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -18,8 +19,23 @@ HUGGINGFACE_ENDPOINT_API = "https://api.endpoints.huggingface.cloud/v2/endpoint/
 
 class HuggingfaceHubTextEmbeddingModel(_CommonHuggingfaceHub, TextEmbeddingModel):
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
+        """
+        Invoke text embedding model
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
+        :return: embeddings result
+        """
         client = InferenceClient(token=credentials["huggingfacehub_api_token"])
 
         execute_model = model

api/core/model_runtime/model_providers/huggingface_tei/text_embedding/text_embedding.py

@@ -1,6 +1,7 @@
 import time
 from typing import Optional
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -23,7 +24,12 @@ class HuggingfaceTeiTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -38,6 +44,7 @@ class HuggingfaceTeiTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         server_url = credentials["server_url"]

api/core/model_runtime/model_providers/hunyuan/text_embedding/text_embedding.py

@@ -9,6 +9,7 @@ from tencentcloud.common.profile.client_profile import ClientProfile
 from tencentcloud.common.profile.http_profile import HttpProfile
 from tencentcloud.hunyuan.v20230901 import hunyuan_client, models
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -26,7 +27,12 @@ class HunyuanTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -35,6 +41,7 @@ class HunyuanTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/jina/jina.yaml

@@ -67,46 +67,3 @@ model_credential_schema:
       required: false
       type: text-input
       default: '8192'
-    - variable: task
-      label:
-        zh_Hans: 下游任务
-        en_US: Downstream task
-      placeholder:
-        zh_Hans: 选择将使用向量模型的下游任务。模型将返回针对该任务优化的向量。
-        en_US: Select the downstream task for which the embeddings will be used. The model will return the optimized embeddings for that task.
-      required: false
-      type: select
-      options:
-        - value: retrieval.query
-          label:
-            en_US: retrieval.query
-        - value: retrieval.passage
-          label:
-            en_US: retrieval.passage
-        - value: separation
-          label:
-            en_US: separation
-        - value: classification
-          label:
-            en_US: classification
-        - value: text-matching
-          label:
-            en_US: text-matching
-    - variable: dimensions
-      label:
-        zh_Hans: 输出维度
-        en_US: Output dimensions
-      placeholder:
-        zh_Hans: 输入您的输出维度
-        en_US: Enter output dimensions
-      required: false
-      type: text-input
-    - variable: late_chunking
-      label:
-        zh_Hans: 后期分块
-        en_US: Late chunking
-      placeholder:
-        zh_Hans: 应用后期分块技术来利用模型的长上下文功能来生成上下文块向量化。
-        en_US: Apply the late chunking technique to leverage the model's long-context capabilities for generating contextual chunk embeddings.
-      required: false
-      type: switch

api/core/model_runtime/model_providers/jina/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from requests import post
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -27,7 +28,7 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
 
     api_base: str = "https://api.jina.ai/v1"
 
-    def _to_payload(self, model: str, texts: list[str], credentials: dict) -> dict:
+    def _to_payload(self, model: str, texts: list[str], credentials: dict, input_type: EmbeddingInputType) -> dict:
         """
         Parse model credentials
 
@@ -44,23 +45,20 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
 
         data = {"model": model, "input": [transform_jina_input_text(model, text) for text in texts]}
 
-        task = credentials.get("task")
-        dimensions = credentials.get("dimensions")
-        late_chunking = credentials.get("late_chunking")
-
-        if task is not None:
-            data["task"] = task
-
-        if dimensions is not None:
-            data["dimensions"] = int(dimensions)
-
-        if late_chunking is not None:
-            data["late_chunking"] = late_chunking
+        # model specific parameters
+        if model == "jina-embeddings-v3":
+            # set `task` type according to input type for the best performance
+            data["task"] = "retrieval.query" if input_type == EmbeddingInputType.QUERY else "retrieval.passage"
 
         return data
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -69,6 +67,7 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         api_key = credentials["api_key"]
@@ -81,7 +80,7 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
         url = base_url + "/embeddings"
         headers = {"Authorization": "Bearer " + api_key, "Content-Type": "application/json"}
 
-        data = self._to_payload(model=model, texts=texts, credentials=credentials)
+        data = self._to_payload(model=model, texts=texts, credentials=credentials, input_type=input_type)
 
         try:
             response = post(url, headers=headers, data=dumps(data))

api/core/model_runtime/model_providers/localai/text_embedding/text_embedding.py

@@ -5,6 +5,7 @@ from typing import Optional
 from requests import post
 from yarl import URL
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -22,11 +23,16 @@ from core.model_runtime.model_providers.__base.text_embedding_model import TextE
 
 class LocalAITextEmbeddingModel(TextEmbeddingModel):
     """
-    Model class for Jina text embedding model.
+    Model class for LocalAI text embedding model.
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -35,6 +41,7 @@ class LocalAITextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         if len(texts) != 1:

api/core/model_runtime/model_providers/minimax/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from requests import post
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -34,7 +35,12 @@ class MinimaxTextEmbeddingModel(TextEmbeddingModel):
     api_base: str = "https://api.minimax.chat/v1/embeddings"
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -43,6 +49,7 @@ class MinimaxTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         api_key = credentials["minimax_api_key"]

api/core/model_runtime/model_providers/mixedbread/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 import requests
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -27,7 +28,12 @@ class MixedBreadTextEmbeddingModel(TextEmbeddingModel):
     api_base: str = "https://api.mixedbread.ai/v1"
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -36,6 +42,7 @@ class MixedBreadTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         api_key = credentials["api_key"]

api/core/model_runtime/model_providers/nomic/text_embedding/text_embedding.py

@@ -5,6 +5,7 @@ from typing import Optional
 from nomic import embed
 from nomic import login as nomic_login
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import (
     EmbeddingUsage,
@@ -46,6 +47,7 @@ class NomicTextEmbeddingModel(_CommonNomic, TextEmbeddingModel):
         credentials: dict,
         texts: list[str],
         user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -54,6 +56,7 @@ class NomicTextEmbeddingModel(_CommonNomic, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         embeddings, prompt_tokens, total_tokens = self.embed_text(

api/core/model_runtime/model_providers/nvidia/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from requests import post
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -27,7 +28,12 @@ class NvidiaTextEmbeddingModel(TextEmbeddingModel):
     models: list[str] = ["NV-Embed-QA"]
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -36,6 +42,7 @@ class NvidiaTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         api_key = credentials["api_key"]

api/core/model_runtime/model_providers/oci/text_embedding/text_embedding.py

@@ -6,6 +6,7 @@ from typing import Optional
 import numpy as np
 import oci
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -41,7 +42,12 @@ class OCITextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -50,6 +56,7 @@ class OCITextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         # get model properties

api/core/model_runtime/model_providers/ollama/text_embedding/text_embedding.py

@@ -8,6 +8,7 @@ from urllib.parse import urljoin
 import numpy as np
 import requests
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import (
     AIModelEntity,
@@ -38,7 +39,12 @@ class OllamaEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -47,6 +53,7 @@ class OllamaEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/openai/text_embedding/text_embedding.py

@@ -6,6 +6,7 @@ import numpy as np
 import tiktoken
 from openai import OpenAI
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.validate import CredentialsValidateFailedError
@@ -19,7 +20,12 @@ class OpenAITextEmbeddingModel(_CommonOpenAI, TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -28,6 +34,7 @@ class OpenAITextEmbeddingModel(_CommonOpenAI, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         # transform credentials to kwargs for model instance

api/core/model_runtime/model_providers/openai_api_compatible/text_embedding/text_embedding.py

@@ -7,6 +7,7 @@ from urllib.parse import urljoin
 import numpy as np
 import requests
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import (
     AIModelEntity,
@@ -28,7 +29,12 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -37,6 +43,7 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/openllm/text_embedding/text_embedding.py

@@ -5,6 +5,7 @@ from typing import Optional
 from requests import post
 from requests.exceptions import ConnectionError, InvalidSchema, MissingSchema
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import (
@@ -25,7 +26,12 @@ class OpenLLMTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -34,6 +40,7 @@ class OpenLLMTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         server_url = credentials["server_url"]

api/core/model_runtime/model_providers/perfxcloud/text_embedding/text_embedding.py

@@ -7,6 +7,7 @@ from urllib.parse import urljoin
 import numpy as np
 import requests
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import (
     AIModelEntity,
@@ -28,7 +29,12 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -37,6 +43,7 @@ class OAICompatEmbeddingModel(_CommonOaiApiCompat, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/replicate/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 
 from replicate import Client as ReplicateClient
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -14,8 +15,23 @@ from core.model_runtime.model_providers.replicate._common import _CommonReplicat
 
 class ReplicateEmbeddingModel(_CommonReplicate, TextEmbeddingModel):
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
+        """
+        Invoke text embedding model
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
+        :return: embeddings result
+        """
         client = ReplicateClient(api_token=credentials["replicate_api_token"], timeout=30)
 
         if "model_version" in credentials:

api/core/model_runtime/model_providers/sagemaker/llm/llm.py

@@ -84,8 +84,9 @@ class SageMakerLargeLanguageModel(LargeLanguageModel):
     Model class for Cohere large language model.
     """
 
-    sagemaker_client: Any = None
+    sagemaker_session: Any = None
     predictor: Any = None
+    sagemaker_endpoint: str = None
 
     def _handle_chat_generate_response(
         self,
@@ -211,7 +212,7 @@ class SageMakerLargeLanguageModel(LargeLanguageModel):
         :param user: unique user id
         :return: full response or stream response chunk generator result
         """
-        if not self.sagemaker_client:
+        if not self.sagemaker_session:
             access_key = credentials.get("aws_access_key_id")
             secret_key = credentials.get("aws_secret_access_key")
             aws_region = credentials.get("aws_region")
@@ -226,11 +227,14 @@ class SageMakerLargeLanguageModel(LargeLanguageModel):
             else:
                 boto_session = boto3.Session()
 
-            self.sagemaker_client = boto_session.client("sagemaker")
-            sagemaker_session = Session(boto_session=boto_session, sagemaker_client=self.sagemaker_client)
+            sagemaker_client = boto_session.client("sagemaker")
+            self.sagemaker_session = Session(boto_session=boto_session, sagemaker_client=sagemaker_client)
+
+        if self.sagemaker_endpoint != credentials.get("sagemaker_endpoint"):
+            self.sagemaker_endpoint = credentials.get("sagemaker_endpoint")
             self.predictor = Predictor(
-                endpoint_name=credentials.get("sagemaker_endpoint"),
-                sagemaker_session=sagemaker_session,
+                endpoint_name=self.sagemaker_endpoint,
+                sagemaker_session=self.sagemaker_session,
                 serializer=serializers.JSONSerializer(),
             )
 

api/core/model_runtime/model_providers/sagemaker/text_embedding/text_embedding.py

@@ -6,6 +6,7 @@ from typing import Any, Optional
 
 import boto3
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -53,7 +54,12 @@ class SageMakerEmbeddingModel(TextEmbeddingModel):
         return embeddings
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -62,6 +68,7 @@ class SageMakerEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         # get model properties

api/core/model_runtime/model_providers/siliconflow/llm/_position.yaml

@@ -1,25 +1,38 @@
-- Qwen/Qwen2.5-7B-Instruct
-- Qwen/Qwen2.5-14B-Instruct
-- Qwen/Qwen2.5-32B-Instruct
 - Qwen/Qwen2.5-72B-Instruct
+- Qwen/Qwen2.5-Math-72B-Instruct
+- Qwen/Qwen2.5-32B-Instruct
+- Qwen/Qwen2.5-14B-Instruct
+- Qwen/Qwen2.5-7B-Instruct
+- Qwen/Qwen2.5-Coder-7B-Instruct
+- deepseek-ai/DeepSeek-V2.5
 - Qwen/Qwen2-72B-Instruct
 - Qwen/Qwen2-57B-A14B-Instruct
 - Qwen/Qwen2-7B-Instruct
 - Qwen/Qwen2-1.5B-Instruct
-- 01-ai/Yi-1.5-34B-Chat
-- 01-ai/Yi-1.5-9B-Chat-16K
-- 01-ai/Yi-1.5-6B-Chat
-- THUDM/glm-4-9b-chat
-- deepseek-ai/DeepSeek-V2.5
 - deepseek-ai/DeepSeek-V2-Chat
 - deepseek-ai/DeepSeek-Coder-V2-Instruct
+- THUDM/glm-4-9b-chat
+- THUDM/chatglm3-6b
+- 01-ai/Yi-1.5-34B-Chat-16K
+- 01-ai/Yi-1.5-9B-Chat-16K
+- 01-ai/Yi-1.5-6B-Chat
+- internlm/internlm2_5-20b-chat
 - internlm/internlm2_5-7b-chat
-- google/gemma-2-27b-it
-- google/gemma-2-9b-it
-- meta-llama/Meta-Llama-3-70B-Instruct
-- meta-llama/Meta-Llama-3-8B-Instruct
 - meta-llama/Meta-Llama-3.1-405B-Instruct
 - meta-llama/Meta-Llama-3.1-70B-Instruct
 - meta-llama/Meta-Llama-3.1-8B-Instruct
-- mistralai/Mixtral-8x7B-Instruct-v0.1
+- meta-llama/Meta-Llama-3-70B-Instruct
+- meta-llama/Meta-Llama-3-8B-Instruct
+- google/gemma-2-27b-it
+- google/gemma-2-9b-it
 - mistralai/Mistral-7B-Instruct-v0.2
+- Pro/Qwen/Qwen2-7B-Instruct
+- Pro/Qwen/Qwen2-1.5B-Instruct
+- Pro/THUDM/glm-4-9b-chat
+- Pro/THUDM/chatglm3-6b
+- Pro/01-ai/Yi-1.5-9B-Chat-16K
+- Pro/01-ai/Yi-1.5-6B-Chat
+- Pro/internlm/internlm2_5-7b-chat
+- Pro/meta-llama/Meta-Llama-3.1-8B-Instruct
+- Pro/meta-llama/Meta-Llama-3-8B-Instruct
+- Pro/google/gemma-2-9b-it

api/core/model_runtime/model_providers/siliconflow/text_embedding/text_embedding.py

@@ -1,5 +1,6 @@
 from typing import Optional
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult
 from core.model_runtime.model_providers.openai_api_compatible.text_embedding.text_embedding import (
     OAICompatEmbeddingModel,
@@ -16,8 +17,23 @@ class SiliconflowTextEmbeddingModel(OAICompatEmbeddingModel):
         super().validate_credentials(model, credentials)
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
+        """
+        Invoke text embedding model
+
+        :param model: model name
+        :param credentials: model credentials
+        :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
+        :return: embeddings result
+        """
         self._add_custom_parameters(credentials)
         return super()._invoke(model, credentials, texts, user)
 

api/core/model_runtime/model_providers/spark/llm/llm.py

@@ -213,18 +213,21 @@ class SparkLargeLanguageModel(LargeLanguageModel):
         :param prompt_messages: prompt messages
         :return: llm response chunk generator result
         """
+        completion = ""
         for index, content in enumerate(client.subscribe()):
             if isinstance(content, dict):
                 delta = content["data"]
             else:
                 delta = content
-
+            completion += delta
             assistant_prompt_message = AssistantPromptMessage(
                 content=delta or "",
             )
-
+            temp_assistant_prompt_message = AssistantPromptMessage(
+                content=completion,
+            )
             prompt_tokens = self.get_num_tokens(model, credentials, prompt_messages)
-            completion_tokens = self.get_num_tokens(model, credentials, [assistant_prompt_message])
+            completion_tokens = self.get_num_tokens(model, credentials, [temp_assistant_prompt_message])
 
             # transform usage
             usage = self._calc_response_usage(model, credentials, prompt_tokens, completion_tokens)

api/core/model_runtime/model_providers/tongyi/text_embedding/text_embedding.py

@@ -4,6 +4,7 @@ from typing import Optional
 import dashscope
 import numpy as np
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import (
     EmbeddingUsage,
@@ -27,6 +28,7 @@ class TongyiTextEmbeddingModel(_CommonTongyi, TextEmbeddingModel):
         credentials: dict,
         texts: list[str],
         user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -35,6 +37,7 @@ class TongyiTextEmbeddingModel(_CommonTongyi, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         credentials_kwargs = self._to_credential_kwargs(credentials)

api/core/model_runtime/model_providers/upstage/text_embedding/text_embedding.py

@@ -7,6 +7,7 @@ import numpy as np
 from openai import OpenAI
 from tokenizers import Tokenizer
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.validate import CredentialsValidateFailedError
@@ -22,7 +23,14 @@ class UpstageTextEmbeddingModel(_CommonUpstage, TextEmbeddingModel):
     def _get_tokenizer(self) -> Tokenizer:
         return Tokenizer.from_pretrained("upstage/solar-1-mini-tokenizer")
 
-    def _invoke(self, model: str, credentials: dict, texts: list[str], user: str | None = None) -> TextEmbeddingResult:
+    def _invoke(
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: str | None = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
+    ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
 
@@ -30,6 +38,7 @@ class UpstageTextEmbeddingModel(_CommonUpstage, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-001.yaml

@@ -1,6 +1,6 @@
 model: gemini-1.5-flash-001
 label:
-  en_US: Gemini 1.5 Flash
+  en_US: Gemini 1.5 Flash 001
 model_type: llm
 features:
   - agent-thought

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-flash-002.yaml

@@ -0,0 +1,37 @@
+model: gemini-1.5-flash-002
+label:
+  en_US: Gemini 1.5 Flash 002
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 1048576
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      en_US: Top k
+    type: int
+    help:
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+  - name: max_output_tokens
+    use_template: max_tokens
+    required: true
+    default: 8192
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00'
+  output: '0.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-001.yaml

@@ -1,6 +1,6 @@
 model: gemini-1.5-pro-001
 label:
-  en_US: Gemini 1.5 Pro
+  en_US: Gemini 1.5 Pro 001
 model_type: llm
 features:
   - agent-thought

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-1.5-pro-002.yaml

@@ -0,0 +1,37 @@
+model: gemini-1.5-pro-002
+label:
+  en_US: Gemini 1.5 Pro 002
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 1048576
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      en_US: Top k
+    type: int
+    help:
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+  - name: max_output_tokens
+    use_template: max_tokens
+    required: true
+    default: 8192
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00'
+  output: '0.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-flash-experimental.yaml

@@ -0,0 +1,37 @@
+model: gemini-flash-experimental
+label:
+  en_US: Gemini Flash Experimental
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 1048576
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      en_US: Top k
+    type: int
+    help:
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+  - name: max_output_tokens
+    use_template: max_tokens
+    required: true
+    default: 8192
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00'
+  output: '0.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/vertex_ai/llm/gemini-pro-experimental.yaml

@@ -0,0 +1,37 @@
+model: gemini-pro-experimental
+label:
+  en_US: Gemini Pro Experimental
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 1048576
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      en_US: Top k
+    type: int
+    help:
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+  - name: max_output_tokens
+    use_template: max_tokens
+    required: true
+    default: 8192
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00'
+  output: '0.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/vertex_ai/llm/llm.py

@@ -2,6 +2,7 @@ import base64
 import io
 import json
 import logging
+import time
 from collections.abc import Generator
 from typing import Optional, Union, cast
 
@@ -20,7 +21,6 @@ from google.api_core import exceptions
 from google.cloud import aiplatform
 from google.oauth2 import service_account
 from PIL import Image
-from vertexai.generative_models import HarmBlockThreshold, HarmCategory
 
 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta, LLMUsage
 from core.model_runtime.entities.message_entities import (
@@ -34,6 +34,7 @@ from core.model_runtime.entities.message_entities import (
     ToolPromptMessage,
     UserPromptMessage,
 )
+from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.errors.invoke import (
     InvokeAuthorizationError,
     InvokeBadRequestError,
@@ -503,20 +504,12 @@ class VertexAiLargeLanguageModel(LargeLanguageModel):
                     else:
                         history.append(content)
 
-        safety_settings = {
-            HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE,
-            HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE,
-            HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
-            HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
-        }
-
         google_model = glm.GenerativeModel(model_name=model, system_instruction=system_instruction)
 
         response = google_model.generate_content(
             contents=history,
             generation_config=glm.GenerationConfig(**config_kwargs),
             stream=stream,
-            safety_settings=safety_settings,
             tools=self._convert_tools_to_glm_tool(tools) if tools else None,
         )
 

api/core/model_runtime/model_providers/vertex_ai/text_embedding/text_embedding.py

@@ -9,6 +9,7 @@ from google.cloud import aiplatform
 from google.oauth2 import service_account
 from vertexai.language_models import TextEmbeddingModel as VertexTextEmbeddingModel
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import (
     AIModelEntity,
@@ -30,7 +31,12 @@ class VertexAiTextEmbeddingModel(_CommonVertexAi, TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -38,6 +44,8 @@ class VertexAiTextEmbeddingModel(_CommonVertexAi, TextEmbeddingModel):
         :param model: model name
         :param credentials: model credentials
         :param texts: texts to embed
+        :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         service_account_info = json.loads(base64.b64decode(credentials["vertex_service_account_key"]))

api/core/model_runtime/model_providers/volcengine_maas/text_embedding/text_embedding.py

@@ -2,6 +2,7 @@ import time
 from decimal import Decimal
 from typing import Optional
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import (
     AIModelEntity,
@@ -41,7 +42,12 @@ class VolcengineMaaSTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -50,6 +56,7 @@ class VolcengineMaaSTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         if ArkClientV3.is_legacy(credentials):

api/core/model_runtime/model_providers/wenxin/text_embedding/text_embedding.py

@@ -7,6 +7,7 @@ from typing import Any, Optional
 import numpy as np
 from requests import Response, post
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.invoke import InvokeError
@@ -70,7 +71,12 @@ class WenxinTextEmbeddingModel(TextEmbeddingModel):
         return WenxinTextEmbedding(api_key, secret_key)
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -79,6 +85,7 @@ class WenxinTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
 

api/core/model_runtime/model_providers/xinference/text_embedding/text_embedding.py

@@ -3,6 +3,7 @@ from typing import Optional
 
 from xinference_client.client.restful.restful_client import Client, RESTfulEmbeddingModelHandle
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.common_entities import I18nObject
 from core.model_runtime.entities.model_entities import AIModelEntity, FetchFrom, ModelPropertyKey, ModelType, PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
@@ -25,7 +26,12 @@ class XinferenceTextEmbeddingModel(TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -40,6 +46,7 @@ class XinferenceTextEmbeddingModel(TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         server_url = credentials["server_url"]

api/core/model_runtime/model_providers/zhipuai/text_embedding/text_embedding.py

@@ -1,6 +1,7 @@
 import time
 from typing import Optional
 
+from core.embedding.embedding_constant import EmbeddingInputType
 from core.model_runtime.entities.model_entities import PriceType
 from core.model_runtime.entities.text_embedding_entities import EmbeddingUsage, TextEmbeddingResult
 from core.model_runtime.errors.validate import CredentialsValidateFailedError
@@ -15,7 +16,12 @@ class ZhipuAITextEmbeddingModel(_CommonZhipuaiAI, TextEmbeddingModel):
     """
 
     def _invoke(
-        self, model: str, credentials: dict, texts: list[str], user: Optional[str] = None
+        self,
+        model: str,
+        credentials: dict,
+        texts: list[str],
+        user: Optional[str] = None,
+        input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
     ) -> TextEmbeddingResult:
         """
         Invoke text embedding model
@@ -24,6 +30,7 @@ class ZhipuAITextEmbeddingModel(_CommonZhipuaiAI, TextEmbeddingModel):
         :param credentials: model credentials
         :param texts: texts to embed
         :param user: unique user id
+        :param input_type: input type
         :return: embeddings result
         """
         credentials_kwargs = self._to_credential_kwargs(credentials)

api/core/rag/extractor/extract_processor.py

@@ -124,7 +124,7 @@ class ExtractProcessor:
                         extractor = UnstructuredPPTXExtractor(file_path, unstructured_api_url)
                     elif file_extension == ".xml":
                         extractor = UnstructuredXmlExtractor(file_path, unstructured_api_url)
-                    elif file_extension == "epub":
+                    elif file_extension == ".epub":
                         extractor = UnstructuredEpubExtractor(file_path, unstructured_api_url)
                     else:
                         # txt
@@ -146,7 +146,7 @@ class ExtractProcessor:
                         extractor = WordExtractor(file_path, upload_file.tenant_id, upload_file.created_by)
                     elif file_extension == ".csv":
                         extractor = CSVExtractor(file_path, autodetect_encoding=True)
-                    elif file_extension == "epub":
+                    elif file_extension == ".epub":
                         extractor = UnstructuredEpubExtractor(file_path)
                     else:
                         # txt

api/core/tools/__base/tool_provider.py

@@ -64,6 +64,9 @@ class ToolProviderController(ABC):
 
             # check type
             credential_schema = credentials_need_to_validate[credential_name]
+            if not credential_schema.required and credentials[credential_name] is None:
+                continue
+
             if credential_schema.type in {ProviderConfig.Type.SECRET_INPUT, ProviderConfig.Type.TEXT_INPUT}:
                 if not isinstance(credentials[credential_name], str):
                     raise ToolProviderCredentialValidationError(f"credential {credential_name} should be string")

api/poetry.lock

@@ -2333,13 +2333,13 @@ develop = ["aiohttp", "furo", "httpx", "opentelemetry-api", "opentelemetry-sdk",
 
 [[package]]
 name = "elasticsearch"
-version = "8.14.0"
+version = "8.15.1"
 description = "Python client for Elasticsearch"
 optional = false
-python-versions = ">=3.7"
+python-versions = ">=3.8"
 files = [
-    {file = "elasticsearch-8.14.0-py3-none-any.whl", hash = "sha256:cef8ef70a81af027f3da74a4f7d9296b390c636903088439087b8262a468c130"},
-    {file = "elasticsearch-8.14.0.tar.gz", hash = "sha256:aa2490029dd96f4015b333c1827aa21fd6c0a4d223b00dfb0fe933b8d09a511b"},
+    {file = "elasticsearch-8.15.1-py3-none-any.whl", hash = "sha256:02a0476e98768a30d7926335fc0d305c04fdb928eea1354c6e6040d8c2814569"},
+    {file = "elasticsearch-8.15.1.tar.gz", hash = "sha256:40c0d312f8adf8bdc81795bc16a0b546ddf544cb1f90e829a244e4780c4dbfd8"},
 ]
 
 [package.dependencies]
@@ -2347,7 +2347,10 @@ elastic-transport = ">=8.13,<9"
 
 [package.extras]
 async = ["aiohttp (>=3,<4)"]
+dev = ["aiohttp", "black", "build", "coverage", "isort", "jinja2", "mapbox-vector-tile", "nox", "numpy", "orjson", "pandas", "pyarrow", "pytest", "pytest-asyncio", "pytest-cov", "python-dateutil", "pyyaml (>=5.4)", "requests (>=2,<3)", "simsimd", "twine", "unasync"]
+docs = ["sphinx", "sphinx-autodoc-typehints", "sphinx-rtd-theme (>=2.0)"]
 orjson = ["orjson (>=3)"]
+pyarrow = ["pyarrow (>=1)"]
 requests = ["requests (>=2.4.0,!=2.32.2,<3.0.0)"]
 vectorstore-mmr = ["numpy (>=1)", "simsimd (>=3)"]
 
@@ -10498,4 +10501,4 @@ cffi = ["cffi (>=1.11)"]
 [metadata]
 lock-version = "2.0"
 python-versions = ">=3.10,<3.13"
-content-hash = "17c4108d92c415d987f8b437ea3e0484c5601a05bfe175339a8546c93c159bc5"
+content-hash = "69b42bb1ff033f14e199fee8335356275099421d72bbd7037b7a991ea65cae08"

api/pyproject.toml

@@ -253,7 +253,7 @@ alibabacloud_gpdb20160503 = "~3.8.0"
 alibabacloud_tea_openapi = "~0.3.9"
 chromadb = "0.5.1"
 clickhouse-connect = "~0.7.16"
-elasticsearch = "8.14.0"
+elasticsearch = "~8.15.1"
 oracledb = "~2.2.1"
 pgvecto-rs = { version = "~0.2.1", extras = ['sqlalchemy'] }
 pgvector = "0.2.5"

api/tests/integration_tests/model_runtime/fireworks/test_text_embedding.py

@@ -0,0 +1,54 @@
+import os
+
+import pytest
+
+from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.fireworks.text_embedding.text_embedding import FireworksTextEmbeddingModel
+from tests.integration_tests.model_runtime.__mock.openai import setup_openai_mock
+
+
+@pytest.mark.parametrize("setup_openai_mock", [["text_embedding"]], indirect=True)
+def test_validate_credentials(setup_openai_mock):
+    model = FireworksTextEmbeddingModel()
+
+    with pytest.raises(CredentialsValidateFailedError):
+        model.validate_credentials(
+            model="nomic-ai/nomic-embed-text-v1.5", credentials={"fireworks_api_key": "invalid_key"}
+        )
+
+    model.validate_credentials(
+        model="nomic-ai/nomic-embed-text-v1.5", credentials={"fireworks_api_key": os.environ.get("FIREWORKS_API_KEY")}
+    )
+
+
+@pytest.mark.parametrize("setup_openai_mock", [["text_embedding"]], indirect=True)
+def test_invoke_model(setup_openai_mock):
+    model = FireworksTextEmbeddingModel()
+
+    result = model.invoke(
+        model="nomic-ai/nomic-embed-text-v1.5",
+        credentials={
+            "fireworks_api_key": os.environ.get("FIREWORKS_API_KEY"),
+        },
+        texts=["hello", "world", " ".join(["long_text"] * 100), " ".join(["another_long_text"] * 100)],
+        user="foo",
+    )
+
+    assert isinstance(result, TextEmbeddingResult)
+    assert len(result.embeddings) == 4
+    assert result.usage.total_tokens == 2
+
+
+def test_get_num_tokens():
+    model = FireworksTextEmbeddingModel()
+
+    num_tokens = model.get_num_tokens(
+        model="nomic-ai/nomic-embed-text-v1.5",
+        credentials={
+            "fireworks_api_key": os.environ.get("FIREWORKS_API_KEY"),
+        },
+        texts=["hello", "world"],
+    )
+
+    assert num_tokens == 2

docker/.env.example

@@ -568,6 +568,10 @@ WORKFLOW_MAX_EXECUTION_STEPS=500
 WORKFLOW_MAX_EXECUTION_TIME=1200
 WORKFLOW_CALL_MAX_DEPTH=5
 
+# HTTP request node in workflow configuration
+HTTP_REQUEST_NODE_MAX_BINARY_SIZE=10485760
+HTTP_REQUEST_NODE_MAX_TEXT_SIZE=1048576
+
 # SSRF Proxy server HTTP URL
 SSRF_PROXY_HTTP_URL=http://ssrf_proxy:3128
 # SSRF Proxy server HTTPS URL

docker/docker-compose.yaml

@@ -207,6 +207,8 @@ x-shared-env: &shared-api-worker-env
   WORKFLOW_CALL_MAX_DEPTH: ${WORKFLOW_MAX_EXECUTION_TIME:-5}
   SSRF_PROXY_HTTP_URL: ${SSRF_PROXY_HTTP_URL:-http://ssrf_proxy:3128}
   SSRF_PROXY_HTTPS_URL: ${SSRF_PROXY_HTTPS_URL:-http://ssrf_proxy:3128}
+  HTTP_REQUEST_NODE_MAX_BINARY_SIZE: ${HTTP_REQUEST_NODE_MAX_BINARY_SIZE:-10485760}
+  HTTP_REQUEST_NODE_MAX_TEXT_SIZE: ${HTTP_REQUEST_NODE_MAX_TEXT_SIZE:-1048576}
 
 services:
   # API service
@@ -628,7 +630,7 @@ services:
   # https://www.elastic.co/guide/en/elasticsearch/reference/current/settings.html
   # https://www.elastic.co/guide/en/elasticsearch/reference/current/docker.html#docker-prod-prerequisites
   elasticsearch:
-    image: docker.elastic.co/elasticsearch/elasticsearch:8.14.3
+    image: docker.elastic.co/elasticsearch/elasticsearch:8.15.1
     container_name: elasticsearch
     profiles:
       - elasticsearch
@@ -655,7 +657,7 @@ services:
   # https://www.elastic.co/guide/en/kibana/current/docker.html
   # https://www.elastic.co/guide/en/kibana/current/settings.html
   kibana:
-    image: docker.elastic.co/kibana/kibana:8.14.3
+    image: docker.elastic.co/kibana/kibana:8.15.1
     container_name: kibana
     profiles:
       - elasticsearch

sdks/python-client/dify_client/client.py

@@ -1,103 +1,80 @@
 import json
+
 import requests
 
 
 class DifyClient:
-    def __init__(self, api_key, base_url: str = 'https://api.dify.ai/v1'):
+    def __init__(self, api_key, base_url: str = "https://api.dify.ai/v1"):
         self.api_key = api_key
         self.base_url = base_url
 
     def _send_request(self, method, endpoint, json=None, params=None, stream=False):
-        headers = {
-            "Authorization": f"Bearer {self.api_key}",
-            "Content-Type": "application/json"
-        }
+        headers = {"Authorization": f"Bearer {self.api_key}", "Content-Type": "application/json"}
 
         url = f"{self.base_url}{endpoint}"
         response = requests.request(method, url, json=json, params=params, headers=headers, stream=stream)
 
         return response
-    
 
     def _send_request_with_files(self, method, endpoint, data, files):
-        headers = {
-            "Authorization": f"Bearer {self.api_key}"
-        }
+        headers = {"Authorization": f"Bearer {self.api_key}"}
 
         url = f"{self.base_url}{endpoint}"
         response = requests.request(method, url, data=data, headers=headers, files=files)
 
         return response
-    
+
     def message_feedback(self, message_id, rating, user):
-        data = {
-            "rating": rating,
-            "user": user
-        }
+        data = {"rating": rating, "user": user}
         return self._send_request("POST", f"/messages/{message_id}/feedbacks", data)
-    
+
     def get_application_parameters(self, user):
         params = {"user": user}
         return self._send_request("GET", "/parameters", params=params)
-    
+
     def file_upload(self, user, files):
-        data = {
-            "user": user
-        }
+        data = {"user": user}
         return self._send_request_with_files("POST", "/files/upload", data=data, files=files)
 
-    def text_to_audio(self, text:str, user:str, streaming:bool=False):
-        data = {
-            "text": text,
-            "user": user,
-            "streaming": streaming
-        }
+    def text_to_audio(self, text: str, user: str, streaming: bool = False):
+        data = {"text": text, "user": user, "streaming": streaming}
         return self._send_request("POST", "/text-to-audio", data=data)
-    
-    def get_meta(self,user):
-        params = { "user": user}
-        return self._send_request("GET", f"/meta", params=params)
+
+    def get_meta(self, user):
+        params = {"user": user}
+        return self._send_request("GET", "/meta", params=params)
 
 
 class CompletionClient(DifyClient):
     def create_completion_message(self, inputs, response_mode, user, files=None):
-        data = {
-            "inputs": inputs,
-            "response_mode": response_mode,
-            "user": user,
-            "files": files
-        }
-        return self._send_request("POST", "/completion-messages", data,
-                                  stream=True if response_mode == "streaming" else False)
+        data = {"inputs": inputs, "response_mode": response_mode, "user": user, "files": files}
+        return self._send_request(
+            "POST", "/completion-messages", data, stream=True if response_mode == "streaming" else False
+        )
 
 
 class ChatClient(DifyClient):
     def create_chat_message(self, inputs, query, user, response_mode="blocking", conversation_id=None, files=None):
-        data = {
-            "inputs": inputs,
-            "query": query,
-            "user": user,
-            "response_mode": response_mode,
-            "files": files
-        }
+        data = {"inputs": inputs, "query": query, "user": user, "response_mode": response_mode, "files": files}
         if conversation_id:
             data["conversation_id"] = conversation_id
 
-        return self._send_request("POST", "/chat-messages", data,
-                                  stream=True if response_mode == "streaming" else False)
-    
-    def get_suggested(self, message_id, user:str):
+        return self._send_request(
+            "POST", "/chat-messages", data, stream=True if response_mode == "streaming" else False
+        )
+
+    def get_suggested(self, message_id, user: str):
         params = {"user": user}
         return self._send_request("GET", f"/messages/{message_id}/suggested", params=params)
-    
+
     def stop_message(self, task_id, user):
         data = {"user": user}
-        return self._send_request("POST", f"/chat-messages/{task_id}/stop", data)   
+        return self._send_request("POST", f"/chat-messages/{task_id}/stop", data)
 
     def get_conversations(self, user, last_id=None, limit=None, pinned=None):
         params = {"user": user, "last_id": last_id, "limit": limit, "pinned": pinned}
         return self._send_request("GET", "/conversations", params=params)
-    
+
     def get_conversation_messages(self, user, conversation_id=None, first_id=None, limit=None):
         params = {"user": user}
 
@@ -109,15 +86,15 @@ class ChatClient(DifyClient):
             params["limit"] = limit
 
         return self._send_request("GET", "/messages", params=params)
-    
-    def rename_conversation(self, conversation_id, name,auto_generate:bool, user:str):
-        data = {"name": name, "auto_generate": auto_generate,"user": user}
+
+    def rename_conversation(self, conversation_id, name, auto_generate: bool, user: str):
+        data = {"name": name, "auto_generate": auto_generate, "user": user}
         return self._send_request("POST", f"/conversations/{conversation_id}/name", data)
 
     def delete_conversation(self, conversation_id, user):
         data = {"user": user}
         return self._send_request("DELETE", f"/conversations/{conversation_id}", data)
-    
+
     def audio_to_text(self, audio_file, user):
         data = {"user": user}
         files = {"audio_file": audio_file}
@@ -125,10 +102,10 @@ class ChatClient(DifyClient):
 
 
 class WorkflowClient(DifyClient):
-    def run(self, inputs:dict, response_mode:str="streaming", user:str="abc-123"):
+    def run(self, inputs: dict, response_mode: str = "streaming", user: str = "abc-123"):
         data = {"inputs": inputs, "response_mode": response_mode, "user": user}
         return self._send_request("POST", "/workflows/run", data)
-    
+
     def stop(self, task_id, user):
         data = {"user": user}
         return self._send_request("POST", f"/workflows/tasks/{task_id}/stop", data)
@@ -137,10 +114,8 @@ class WorkflowClient(DifyClient):
         return self._send_request("GET", f"/workflows/run/{workflow_run_id}")
 
 
-
 class KnowledgeBaseClient(DifyClient):
-
-    def __init__(self, api_key, base_url: str = 'https://api.dify.ai/v1', dataset_id: str = None):
+    def __init__(self, api_key, base_url: str = "https://api.dify.ai/v1", dataset_id: str = None):
         """
         Construct a KnowledgeBaseClient object.
 
@@ -150,10 +125,7 @@ class KnowledgeBaseClient(DifyClient):
             dataset_id (str, optional): ID of the dataset. Defaults to None. You don't need this if you just want to
                 create a new dataset. or list datasets. otherwise you need to set this.
         """
-        super().__init__(
-            api_key=api_key,
-            base_url=base_url
-        )
+        super().__init__(api_key=api_key, base_url=base_url)
         self.dataset_id = dataset_id
 
     def _get_dataset_id(self):
@@ -162,10 +134,10 @@ class KnowledgeBaseClient(DifyClient):
         return self.dataset_id
 
     def create_dataset(self, name: str, **kwargs):
-        return self._send_request('POST', '/datasets', {'name': name}, **kwargs)
+        return self._send_request("POST", "/datasets", {"name": name}, **kwargs)
 
     def list_datasets(self, page: int = 1, page_size: int = 20, **kwargs):
-        return self._send_request('GET', f'/datasets?page={page}&limit={page_size}', **kwargs)
+        return self._send_request("GET", f"/datasets?page={page}&limit={page_size}", **kwargs)
 
     def create_document_by_text(self, name, text, extra_params: dict = None, **kwargs):
         """
@@ -193,14 +165,7 @@ class KnowledgeBaseClient(DifyClient):
         }
         :return: Response from the API
         """
-        data = {
-            'indexing_technique': 'high_quality',
-            'process_rule': {
-                'mode': 'automatic'
-            },
-            'name': name,
-            'text': text
-        }
+        data = {"indexing_technique": "high_quality", "process_rule": {"mode": "automatic"}, "name": name, "text": text}
         if extra_params is not None and isinstance(extra_params, dict):
             data.update(extra_params)
         url = f"/datasets/{self._get_dataset_id()}/document/create_by_text"
@@ -233,10 +198,7 @@ class KnowledgeBaseClient(DifyClient):
         }
         :return: Response from the API
         """
-        data = {
-            'name': name,
-            'text': text
-        }
+        data = {"name": name, "text": text}
         if extra_params is not None and isinstance(extra_params, dict):
             data.update(extra_params)
         url = f"/datasets/{self._get_dataset_id()}/documents/{document_id}/update_by_text"
@@ -269,16 +231,11 @@ class KnowledgeBaseClient(DifyClient):
         :return: Response from the API
         """
         files = {"file": open(file_path, "rb")}
-        data = {
-            'process_rule': {
-                'mode': 'automatic'
-            },
-            'indexing_technique': 'high_quality'
-        }
+        data = {"process_rule": {"mode": "automatic"}, "indexing_technique": "high_quality"}
         if extra_params is not None and isinstance(extra_params, dict):
             data.update(extra_params)
         if original_document_id is not None:
-            data['original_document_id'] = original_document_id
+            data["original_document_id"] = original_document_id
         url = f"/datasets/{self._get_dataset_id()}/document/create_by_file"
         return self._send_request_with_files("POST", url, {"data": json.dumps(data)}, files)
 
@@ -352,11 +309,11 @@ class KnowledgeBaseClient(DifyClient):
         """
         params = {}
         if page is not None:
-            params['page'] = page
+            params["page"] = page
         if page_size is not None:
-            params['limit'] = page_size
+            params["limit"] = page_size
         if keyword is not None:
-            params['keyword'] = keyword
+            params["keyword"] = keyword
         url = f"/datasets/{self._get_dataset_id()}/documents"
         return self._send_request("GET", url, params=params, **kwargs)
 
@@ -383,9 +340,9 @@ class KnowledgeBaseClient(DifyClient):
         url = f"/datasets/{self._get_dataset_id()}/documents/{document_id}/segments"
         params = {}
         if keyword is not None:
-            params['keyword'] = keyword
+            params["keyword"] = keyword
         if status is not None:
-            params['status'] = status
+            params["status"] = status
         if "params" in kwargs:
             params.update(kwargs["params"])
         return self._send_request("GET", url, params=params, **kwargs)

web/app/components/workflow/nodes/http/node.tsx

@@ -15,7 +15,7 @@ const Node: FC<NodeProps<HttpNodeType>> = ({
     <div className='mb-1 px-3 py-1'>
       <div className='flex items-start p-1 rounded-md bg-gray-100'>
         <div className='flex items-center h-4 shrink-0 px-1 rounded bg-gray-25 text-xs font-semibold text-gray-700 uppercase'>{method}</div>
-        <div className='pl-1'>
+        <div className='pl-1 pt-1'>
           <ReadonlyInputWithSelectVar
             value={url}
             nodeId={id}

web/i18n/es-ES/common.ts

@@ -202,7 +202,7 @@ const translation = {
     invitationLink: 'Enlace de invitación',
     failedInvitationEmails: 'Los siguientes usuarios no fueron invitados exitosamente',
     ok: 'OK',
-    removeFromTeam: 'Eliminar del equipo',
+    removeFromTeam: 'Eliminar del espacio de trabajo',
     removeFromTeamTip: 'Se eliminará el acceso al equipo',
     setAdmin: 'Establecer como administrador',
     setMember: 'Establecer como miembro ordinario',

web/i18n/zh-Hans/common.ts

@@ -200,7 +200,7 @@ const translation = {
     invitationLink: '邀请链接',
     failedInvitationEmails: '邀请以下邮箱失败',
     ok: '好的',
-    removeFromTeam: '移除团队',
+    removeFromTeam: '移出团队',
     removeFromTeamTip: '将取消团队访问',
     setAdmin: '设为管理员',
     setMember: '设为普通成员',

web/i18n/zh-Hant/common.ts

@@ -194,7 +194,7 @@ const translation = {
     invitationLink: '邀請連結',
     failedInvitationEmails: '邀請以下郵箱失敗',
     ok: '好的',
-    removeFromTeam: '移除團隊',
+    removeFromTeam: '移出團隊',
     removeFromTeamTip: '將取消團隊訪問',
     setAdmin: '設為管理員',
     setMember: '設為普通成員',

web/public/embed.js

@@ -69,38 +69,47 @@
       iframe.id = iframeId;
       iframe.src = iframeUrl;
       iframe.style.cssText = `
-        border: none; position: fixed; flex-direction: column; justify-content: space-between;
+        border: none; position: absolute; flex-direction: column; justify-content: space-between;
         box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px;
-        bottom: 5rem; right: 1rem; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem;
+        bottom: 55px; right: 0; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem;
         max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647;
         overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none;
       `;
 
-      document.body.appendChild(iframe);
+      return iframe;
     }
 
     // Function to reset the iframe position
     function resetIframePosition() {
+      if (window.innerWidth <= 640)
+        return
+
       const targetIframe = document.getElementById(iframeId);
       const targetButton = document.getElementById(buttonId);
       if (targetIframe && targetButton) {
         const buttonRect = targetButton.getBoundingClientRect();
-        const buttonBottom = window.innerHeight - buttonRect.bottom;
-        const buttonRight = window.innerWidth - buttonRect.right;
-        const buttonLeft = buttonRect.left;
 
-        // Adjust iframe position to stay within viewport
-        targetIframe.style.bottom = `${
-          buttonBottom + buttonRect.height + 5 + targetIframe.clientHeight > window.innerHeight
-            ? buttonBottom - targetIframe.clientHeight - 5
-            : buttonBottom + buttonRect.height + 5
-        }px`;
+        const buttonInBottom = buttonRect.top - 5 > targetIframe.clientHeight
 
-        targetIframe.style.right = `${
-          buttonRight + targetIframe.clientWidth > window.innerWidth
-            ? window.innerWidth - buttonLeft - targetIframe.clientWidth
-            : buttonRight
-        }px`;
+        if (buttonInBottom) {
+          targetIframe.style.bottom = `${buttonRect.height + 5}px`;
+          targetIframe.style.top = 'unset';
+        }
+        else {
+          targetIframe.style.bottom = 'unset';
+          targetIframe.style.top = `${buttonRect.height + 5}px`;
+        }
+
+        const buttonInRight = buttonRect.right > targetIframe.clientWidth;
+
+        if (buttonInRight) {
+          targetIframe.style.right = '0';
+          targetIframe.style.left = 'unset';
+        }
+        else {
+          targetIframe.style.right = 'unset';
+          targetIframe.style.left = 0;
+        }
       }
     }
 
@@ -146,12 +155,6 @@
           box-shadow: var(--${containerDiv.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px);
           cursor: pointer;
           z-index: 2147483647;
-          transition: all 0.2s ease-in-out 0s;
-        }
-      `);
-      styleSheet.sheet.insertRule(`
-        #${containerDiv.id}:hover {
-          transform: var(--${containerDiv.id}-hover-transform, scale(1.1));
         }
       `);
 
@@ -167,7 +170,7 @@
       containerDiv.addEventListener("click", function () {
         const targetIframe = document.getElementById(iframeId);
         if (!targetIframe) {
-          createIframe();
+          containerDiv.appendChild(createIframe());
           resetIframePosition();
           this.title = "Exit (ESC)";
           displayDiv.innerHTML = svgIcons.close;
@@ -255,9 +258,6 @@
     if (!document.getElementById(buttonId)) {
       createButton();
     }
-
-    createIframe();
-    document.getElementById(iframeId).style.display = 'none';
   }
 
   // Add esc Exit keyboard event triggered
@@ -279,4 +279,4 @@
   } else {
     document.body.onload = embedChatbot;
   }
-})();
+})();

web/public/embed.min.js

@@ -1,31 +1,26 @@
-(()=>{let t="difyChatbotConfig",a="dify-chatbot-bubble-button",c="dify-chatbot-bubble-window",h=window[t],p={open:`<svg id="openIcon" width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-    <path fill-rule="evenodd" clip-rule="evenodd" d="M7.7586 2L16.2412 2C17.0462 1.99999 17.7105 1.99998 18.2517 2.04419C18.8138 2.09012 19.3305 2.18868 19.8159 2.43598C20.5685 2.81947 21.1804 3.43139 21.5639 4.18404C21.8112 4.66937 21.9098 5.18608 21.9557 5.74818C21.9999 6.28937 21.9999 6.95373 21.9999 7.7587L22 14.1376C22.0004 14.933 22.0007 15.5236 21.8636 16.0353C21.4937 17.4156 20.4155 18.4938 19.0352 18.8637C18.7277 18.9461 18.3917 18.9789 17.9999 18.9918L17.9999 20.371C18 20.6062 18 20.846 17.9822 21.0425C17.9651 21.2305 17.9199 21.5852 17.6722 21.8955C17.3872 22.2525 16.9551 22.4602 16.4983 22.4597C16.1013 22.4593 15.7961 22.273 15.6386 22.1689C15.474 22.06 15.2868 21.9102 15.1031 21.7632L12.69 19.8327C12.1714 19.4178 12.0174 19.3007 11.8575 19.219C11.697 19.137 11.5262 19.0771 11.3496 19.0408C11.1737 19.0047 10.9803 19 10.3162 19H7.75858C6.95362 19 6.28927 19 5.74808 18.9558C5.18598 18.9099 4.66928 18.8113 4.18394 18.564C3.43129 18.1805 2.81937 17.5686 2.43588 16.816C2.18859 16.3306 2.09002 15.8139 2.0441 15.2518C1.99988 14.7106 1.99989 14.0463 1.9999 13.2413V7.75868C1.99989 6.95372 1.99988 6.28936 2.0441 5.74818C2.09002 5.18608 2.18859 4.66937 2.43588 4.18404C2.81937 3.43139 3.43129 2.81947 4.18394 2.43598C4.66928 2.18868 5.18598 2.09012 5.74808 2.04419C6.28927 1.99998 6.95364 1.99999 7.7586 2ZM10.5073 7.5C10.5073 6.67157 9.83575 6 9.00732 6C8.1789 6 7.50732 6.67157 7.50732 7.5C7.50732 8.32843 8.1789 9 9.00732 9C9.83575 9 10.5073 8.32843 10.5073 7.5ZM16.6073 11.7001C16.1669 11.3697 15.5426 11.4577 15.2105 11.8959C15.1488 11.9746 15.081 12.0486 15.0119 12.1207C14.8646 12.2744 14.6432 12.4829 14.3566 12.6913C13.7796 13.111 12.9818 13.5001 12.0073 13.5001C11.0328 13.5001 10.235 13.111 9.65799 12.6913C9.37138 12.4829 9.15004 12.2744 9.00274 12.1207C8.93366 12.0486 8.86581 11.9745 8.80418 11.8959C8.472 11.4577 7.84775 11.3697 7.40732 11.7001C6.96549 12.0314 6.87595 12.6582 7.20732 13.1001C7.20479 13.0968 7.21072 13.1043 7.22094 13.1171C7.24532 13.1478 7.29407 13.2091 7.31068 13.2289C7.36932 13.2987 7.45232 13.3934 7.55877 13.5045C7.77084 13.7258 8.08075 14.0172 8.48165 14.3088C9.27958 14.8891 10.4818 15.5001 12.0073 15.5001C13.5328 15.5001 14.735 14.8891 15.533 14.3088C15.9339 14.0172 16.2438 13.7258 16.4559 13.5045C16.5623 13.3934 16.6453 13.2987 16.704 13.2289C16.7333 13.1939 16.7567 13.165 16.7739 13.1432C17.1193 12.6969 17.0729 12.0493 16.6073 11.7001ZM15.0073 6C15.8358 6 16.5073 6.67157 16.5073 7.5C16.5073 8.32843 15.8358 9 15.0073 9C14.1789 9 13.5073 8.32843 13.5073 7.5C13.5073 6.67157 14.1789 6 15.0073 6Z" fill="white"/>
-  </svg>`,close:`<svg id="closeIcon" width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-    <path d="M18 18L6 6M6 18L18 6" stroke="white" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"/>
-  </svg>`};async function e(){if(h&&h.token){var e=new URLSearchParams(await(async()=>{var e=h?.inputs||{};let n={};return await Promise.all(Object.entries(e).map(async([e,t])=>{n[e]=(e=t,e=(new TextEncoder).encode(e),e=new Response(new Blob([e]).stream().pipeThrough(new CompressionStream("gzip"))).arrayBuffer(),e=new Uint8Array(await e),await btoa(String.fromCharCode(...e)))})),n})());let t=`${h.baseUrl||`https://${h.isDev?"dev.":""}udify.app`}/chatbot/${h.token}?`+e;function o(){var e=document.createElement("iframe");e.allow="fullscreen;microphone",e.title="dify chatbot bubble window",e.id=c,e.src=t,e.style.cssText=`
-      border: none; position: fixed; flex-direction: column; justify-content: space-between;
-      box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px;
-      bottom: 5rem; right: 1rem; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem;
-      max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647;
-      overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none;
-    `,document.body.appendChild(e)}function i(){var e,t,n,o=document.getElementById(c),i=document.getElementById(a);o&&i&&(i=i.getBoundingClientRect(),e=window.innerHeight-i.bottom,t=window.innerWidth-i.right,n=i.left,o.style.bottom=`${e+i.height+5+o.clientHeight>window.innerHeight?e-o.clientHeight-5:e+i.height+5}px`,o.style.right=`${t+o.clientWidth>window.innerWidth?window.innerWidth-n-o.clientWidth:t}px`)}function n(){let n=document.createElement("div");Object.entries(h.containerProps||{}).forEach(([e,t])=>{"className"===e?n.classList.add(...t.split(" ")):"style"===e?"object"==typeof t?Object.assign(n.style,t):n.style.cssText=t:"function"==typeof t?n.addEventListener(e.replace(/^on/,"").toLowerCase(),t):n[e]=t}),n.id=a;var e=document.createElement("style");document.head.appendChild(e),e.sheet.insertRule(`
-      #${n.id} {
-        position: fixed;
-        bottom: var(--${n.id}-bottom, 1rem);
-        right: var(--${n.id}-right, 1rem);
-        left: var(--${n.id}-left, unset);
-        top: var(--${n.id}-top, unset);
-        width: var(--${n.id}-width, 50px);
-        height: var(--${n.id}-height, 50px);
-        border-radius: var(--${n.id}-border-radius, 25px);
-        background-color: var(--${n.id}-bg-color, #155EEF);
-        box-shadow: var(--${n.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px);
-        cursor: pointer;
-        z-index: 2147483647;
-        transition: all 0.2s ease-in-out 0s;
-      }
-    `),e.sheet.insertRule(`
-      #${n.id}:hover {
-        transform: var(--${n.id}-hover-transform, scale(1.1));
-      }
-    `);let t=document.createElement("div");if(t.style.cssText="display: flex; align-items: center; justify-content: center; width: 100%; height: 100%; z-index: 2147483647;",t.innerHTML=p.open,n.appendChild(t),document.body.appendChild(n),n.addEventListener("click",function(){var e=document.getElementById(c);e?(e.style.display="none"===e.style.display?"block":"none",t.innerHTML="none"===e.style.display?p.open:p.close,"none"===e.style.display?document.removeEventListener("keydown",d):document.addEventListener("keydown",d),i()):(o(),i(),this.title="Exit (ESC)",t.innerHTML=p.close,document.addEventListener("keydown",d))}),h.draggable){var s=n;var l=h.dragAxis||"both";let i=!1,d,r;s.addEventListener("mousedown",function(e){i=!0,d=e.clientX-s.offsetLeft,r=e.clientY-s.offsetTop}),document.addEventListener("mousemove",function(e){var t,n,o;i&&(s.style.transition="none",s.style.cursor="grabbing",(t=document.getElementById(c))&&(t.style.display="none",s.querySelector("div").innerHTML=p.open),t=e.clientX-d,e=window.innerHeight-e.clientY-r,o=s.getBoundingClientRect(),n=window.innerWidth-o.width,o=window.innerHeight-o.height,"x"!==l&&"both"!==l||s.style.setProperty(`--${a}-left`,Math.max(0,Math.min(t,n))+"px"),"y"!==l&&"both"!==l||s.style.setProperty(`--${a}-bottom`,Math.max(0,Math.min(e,o))+"px"))}),document.addEventListener("mouseup",function(){i=!1,s.style.transition="",s.style.cursor="pointer"})}}2048<t.length&&console.error("The URL is too long, please reduce the number of inputs to prevent the bot from failing to load"),document.getElementById(a)||n(),o(),document.getElementById(c).style.display="none"}else console.error(t+" is empty or token is not provided")}function d(e){var t;"Escape"===e.key&&(e=document.getElementById(c),t=document.getElementById(a),e)&&"none"!==e.style.display&&(e.style.display="none",t.querySelector("div").innerHTML=p.open)}document.addEventListener("keydown",d),h?.dynamicScript?e():document.body.onload=e})();
+!function(){const n="difyChatbotConfig",a="dify-chatbot-bubble-button",c="dify-chatbot-bubble-window",p=window[n],h={open:`<svg id="openIcon" width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+      <path fill-rule="evenodd" clip-rule="evenodd" d="M7.7586 2L16.2412 2C17.0462 1.99999 17.7105 1.99998 18.2517 2.04419C18.8138 2.09012 19.3305 2.18868 19.8159 2.43598C20.5685 2.81947 21.1804 3.43139 21.5639 4.18404C21.8112 4.66937 21.9098 5.18608 21.9557 5.74818C21.9999 6.28937 21.9999 6.95373 21.9999 7.7587L22 14.1376C22.0004 14.933 22.0007 15.5236 21.8636 16.0353C21.4937 17.4156 20.4155 18.4938 19.0352 18.8637C18.7277 18.9461 18.3917 18.9789 17.9999 18.9918L17.9999 20.371C18 20.6062 18 20.846 17.9822 21.0425C17.9651 21.2305 17.9199 21.5852 17.6722 21.8955C17.3872 22.2525 16.9551 22.4602 16.4983 22.4597C16.1013 22.4593 15.7961 22.273 15.6386 22.1689C15.474 22.06 15.2868 21.9102 15.1031 21.7632L12.69 19.8327C12.1714 19.4178 12.0174 19.3007 11.8575 19.219C11.697 19.137 11.5262 19.0771 11.3496 19.0408C11.1737 19.0047 10.9803 19 10.3162 19H7.75858C6.95362 19 6.28927 19 5.74808 18.9558C5.18598 18.9099 4.66928 18.8113 4.18394 18.564C3.43129 18.1805 2.81937 17.5686 2.43588 16.816C2.18859 16.3306 2.09002 15.8139 2.0441 15.2518C1.99988 14.7106 1.99989 14.0463 1.9999 13.2413V7.75868C1.99989 6.95372 1.99988 6.28936 2.0441 5.74818C2.09002 5.18608 2.18859 4.66937 2.43588 4.18404C2.81937 3.43139 3.43129 2.81947 4.18394 2.43598C4.66928 2.18868 5.18598 2.09012 5.74808 2.04419C6.28927 1.99998 6.95364 1.99999 7.7586 2ZM10.5073 7.5C10.5073 6.67157 9.83575 6 9.00732 6C8.1789 6 7.50732 6.67157 7.50732 7.5C7.50732 8.32843 8.1789 9 9.00732 9C9.83575 9 10.5073 8.32843 10.5073 7.5ZM16.6073 11.7001C16.1669 11.3697 15.5426 11.4577 15.2105 11.8959C15.1488 11.9746 15.081 12.0486 15.0119 12.1207C14.8646 12.2744 14.6432 12.4829 14.3566 12.6913C13.7796 13.111 12.9818 13.5001 12.0073 13.5001C11.0328 13.5001 10.235 13.111 9.65799 12.6913C9.37138 12.4829 9.15004 12.2744 9.00274 12.1207C8.93366 12.0486 8.86581 11.9745 8.80418 11.8959C8.472 11.4577 7.84775 11.3697 7.40732 11.7001C6.96549 12.0314 6.87595 12.6582 7.20732 13.1001C7.20479 13.0968 7.21072 13.1043 7.22094 13.1171C7.24532 13.1478 7.29407 13.2091 7.31068 13.2289C7.36932 13.2987 7.45232 13.3934 7.55877 13.5045C7.77084 13.7258 8.08075 14.0172 8.48165 14.3088C9.27958 14.8891 10.4818 15.5001 12.0073 15.5001C13.5328 15.5001 14.735 14.8891 15.533 14.3088C15.9339 14.0172 16.2438 13.7258 16.4559 13.5045C16.5623 13.3934 16.6453 13.2987 16.704 13.2289C16.7333 13.1939 16.7567 13.165 16.7739 13.1432C17.1193 12.6969 17.0729 12.0493 16.6073 11.7001ZM15.0073 6C15.8358 6 16.5073 6.67157 16.5073 7.5C16.5073 8.32843 15.8358 9 15.0073 9C14.1789 9 13.5073 8.32843 13.5073 7.5C13.5073 6.67157 14.1789 6 15.0073 6Z" fill="white"/>
+    </svg>`,close:`<svg id="closeIcon" width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+      <path d="M18 18L6 6M6 18L18 6" stroke="white" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"/>
+    </svg>`};async function e(){if(p&&p.token){var e=new URLSearchParams(await async function(){var e=p?.inputs||{};const n={};return await Promise.all(Object.entries(e).map(async([e,t])=>{n[e]=(e=t,e=(new TextEncoder).encode(e),e=new Response(new Blob([e]).stream().pipeThrough(new CompressionStream("gzip"))).arrayBuffer(),e=new Uint8Array(await e),await btoa(String.fromCharCode(...e)))})),n}());const i=`${p.baseUrl||`https://${p.isDev?"dev.":""}udify.app`}/chatbot/${p.token}?`+e;function o(){var e,t;window.innerWidth<=640||(e=document.getElementById(c),t=document.getElementById(a),e&&t&&((t=t.getBoundingClientRect()).top-5>e.clientHeight?(e.style.bottom=t.height+5+"px",e.style.top="unset"):(e.style.bottom="unset",e.style.top=t.height+5+"px"),t.right>e.clientWidth?(e.style.right="0",e.style.left="unset"):(e.style.right="unset",e.style.left=0)))}function t(){const n=document.createElement("div");Object.entries(p.containerProps||{}).forEach(([e,t])=>{"className"===e?n.classList.add(...t.split(" ")):"style"===e?"object"==typeof t?Object.assign(n.style,t):n.style.cssText=t:"function"==typeof t?n.addEventListener(e.replace(/^on/,"").toLowerCase(),t):n[e]=t}),n.id=a;var e=document.createElement("style");document.head.appendChild(e),e.sheet.insertRule(`
+        #${n.id} {
+          position: fixed;
+          bottom: var(--${n.id}-bottom, 1rem);
+          right: var(--${n.id}-right, 1rem);
+          left: var(--${n.id}-left, unset);
+          top: var(--${n.id}-top, unset);
+          width: var(--${n.id}-width, 50px);
+          height: var(--${n.id}-height, 50px);
+          border-radius: var(--${n.id}-border-radius, 25px);
+          background-color: var(--${n.id}-bg-color, #155EEF);
+          box-shadow: var(--${n.id}-box-shadow, rgba(0, 0, 0, 0.2) 0px 4px 8px 0px);
+          cursor: pointer;
+          z-index: 2147483647;
+        }
+      `);const t=document.createElement("div");if(t.style.cssText="display: flex; align-items: center; justify-content: center; width: 100%; height: 100%; z-index: 2147483647;",t.innerHTML=h.open,n.appendChild(t),document.body.appendChild(n),n.addEventListener("click",function(){var e=document.getElementById(c);e?(e.style.display="none"===e.style.display?"block":"none",t.innerHTML="none"===e.style.display?h.open:h.close,"none"===e.style.display?document.removeEventListener("keydown",d):document.addEventListener("keydown",d),o()):(n.appendChild(((e=document.createElement("iframe")).allow="fullscreen;microphone",e.title="dify chatbot bubble window",e.id=c,e.src=i,e.style.cssText=`
+        border: none; position: absolute; flex-direction: column; justify-content: space-between;
+        box-shadow: rgba(150, 150, 150, 0.2) 0px 10px 30px 0px, rgba(150, 150, 150, 0.2) 0px 0px 0px 1px;
+        bottom: 55px; right: 0; width: 24rem; max-width: calc(100vw - 2rem); height: 40rem;
+        max-height: calc(100vh - 6rem); border-radius: 0.75rem; display: flex; z-index: 2147483647;
+        overflow: hidden; left: unset; background-color: #F3F4F6;user-select: none;
+      `,e)),o(),this.title="Exit (ESC)",t.innerHTML=h.close,document.addEventListener("keydown",d))}),p.draggable){var s=n;var l=p.dragAxis||"both";let i=!1,d,r;s.addEventListener("mousedown",function(e){i=!0,d=e.clientX-s.offsetLeft,r=e.clientY-s.offsetTop}),document.addEventListener("mousemove",function(e){var t,n,o;i&&(s.style.transition="none",s.style.cursor="grabbing",(t=document.getElementById(c))&&(t.style.display="none",s.querySelector("div").innerHTML=h.open),t=e.clientX-d,e=window.innerHeight-e.clientY-r,o=s.getBoundingClientRect(),n=window.innerWidth-o.width,o=window.innerHeight-o.height,"x"!==l&&"both"!==l||s.style.setProperty(`--${a}-left`,Math.max(0,Math.min(t,n))+"px"),"y"!==l&&"both"!==l||s.style.setProperty(`--${a}-bottom`,Math.max(0,Math.min(e,o))+"px"))}),document.addEventListener("mouseup",function(){i=!1,s.style.transition="",s.style.cursor="pointer"})}}2048<i.length&&console.error("The URL is too long, please reduce the number of inputs to prevent the bot from failing to load"),document.getElementById(a)||t()}else console.error(n+" is empty or token is not provided")}function d(e){var t;"Escape"===e.key&&(e=document.getElementById(c),t=document.getElementById(a),e)&&"none"!==e.style.display&&(e.style.display="none",t.querySelector("div").innerHTML=h.open)}document.addEventListener("keydown",d),p?.dynamicScript?e():document.body.onload=e}();