camel-ai 0.2.18__py3-none-any.whl → 0.2.20a0__py3-none-any.whl

camel/toolkits/semantic_scholar_toolkit.py

@@ -0,0 +1,308 @@
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
+
+import json
+from typing import List, Optional
+
+import requests
+
+from camel.toolkits import FunctionTool
+from camel.toolkits.base import BaseToolkit
+
+
+class SemanticScholarToolkit(BaseToolkit):
+    r"""A toolkit for interacting with the Semantic Scholar
+    API to fetch paper and author data.
+    """
+
+    def __init__(self):
+        r"""Initializes the SemanticScholarToolkit."""
+        self.base_url = "https://api.semanticscholar.org/graph/v1"
+
+    def fetch_paper_data_title(
+        self,
+        paper_title: str,
+        fields: Optional[List[str]] = None,
+    ) -> dict:
+        r"""Fetches a SINGLE paper from the Semantic Scholar
+        API based on a paper title.
+
+        Args:
+            paper_title (str): The title of the paper to fetch.
+            fields (Optional[List[str]], optional): The fields to include in
+                the response (default: :obj:`None`). If not provided defaults
+                to ["title", "abstract", "authors", "year", "citationCount",
+                "publicationTypes", "publicationDate", "openAccessPdf"].
+
+        Returns:
+            dict: The response data from the API or error information if the
+                 request fails.
+        """
+        if fields is None:
+            fields = [
+                "title",
+                "abstract",
+                "authors",
+                "year",
+                "citationCount",
+                "publicationTypes",
+                "publicationDate",
+                "openAccessPdf",
+            ]
+
+        url = f"{self.base_url}/paper/search"
+        query_params = {"query": paper_title, "fields": ",".join(fields)}
+        try:
+            response = requests.get(url, params=query_params)
+            response.raise_for_status()
+            return response.json()
+        except requests.exceptions.RequestException as e:
+            return {
+                "error": f"Request failed: {e!s}",
+                "message": str(e),
+            }
+        except ValueError:
+            return {
+                "error": "Response is not valid JSON",
+                "message": response.text,
+            }
+
+    def fetch_paper_data_id(
+        self,
+        paper_id: str,
+        fields: Optional[List[str]] = None,
+    ) -> dict:
+        r"""Fetches a SINGLE paper from the Semantic Scholar
+        API based on a paper ID.
+
+        Args:
+            paper_id (str): The ID of the paper to fetch.
+            fields (Optional[List[str]], optional): The fields to include in
+                the response (default: :obj:`None`). If not provided defaults
+                to ["title", "abstract", "authors", "year", "citationCount",
+                "publicationTypes", "publicationDate", "openAccessPdf"].
+
+        Returns:
+            dict: The response data from the API or error information
+                if the request fails.
+        """
+        if fields is None:
+            fields = [
+                "title",
+                "abstract",
+                "authors",
+                "year",
+                "citationCount",
+                "publicationTypes",
+                "publicationDate",
+                "openAccessPdf",
+            ]
+
+        url = f"{self.base_url}/paper/{paper_id}"
+        query_params = {"fields": ",".join(fields)}
+        try:
+            response = requests.get(url, params=query_params)
+            response.raise_for_status()
+            return response.json()
+        except requests.exceptions.RequestException as e:
+            return {
+                "error": f"Request failed: {e!s}",
+                "message": str(e),
+            }
+        except ValueError:
+            return {
+                "error": "Response is not valid JSON",
+                "message": response.text,
+            }
+
+    def fetch_bulk_paper_data(
+        self,
+        query: str,
+        year: str = "2023-",
+        fields: Optional[List[str]] = None,
+    ) -> dict:
+        r"""Fetches MULTIPLE papers at once from the Semantic Scholar
+        API based on a related topic.
+
+        Args:
+            query (str): The text query to match against the paper's title and
+                abstract. For example, you can use the following operators and
+                techniques to construct your query: Example 1: ((cloud
+                computing) | virtualization) +security -privacy This will
+                match papers whose title or abstract contains "cloud" and
+                "computing", or contains the word "virtualization". The papers
+                must also include the term "security" but exclude papers that
+                contain the word "privacy".
+            year (str, optional): The year filter for papers (default:
+                :obj:`"2023-"`).
+            fields (Optional[List[str]], optional): The fields to include in
+                the response (default: :obj:`None`). If not provided defaults
+                to ["title", "url", "publicationTypes", "publicationDate",
+                "openAccessPdf"].
+
+        Returns:
+            dict: The response data from the API or error information if the
+                request fails.
+        """
+        if fields is None:
+            fields = [
+                "title",
+                "url",
+                "publicationTypes",
+                "publicationDate",
+                "openAccessPdf",
+            ]
+
+        url = f"{self.base_url}/paper/search/bulk"
+        query_params = {
+            "query": query,
+            "fields": ",".join(fields),
+            "year": year,
+        }
+        try:
+            response = requests.get(url, params=query_params)
+            response.raise_for_status()
+            return response.json()
+        except requests.exceptions.RequestException as e:
+            return {
+                "error": f"Request failed: {e!s}",
+                "message": str(e),
+            }
+        except ValueError:
+            return {
+                "error": "Response is not valid JSON",
+                "message": response.text,
+            }
+
+    def fetch_recommended_papers(
+        self,
+        positive_paper_ids: List[str],
+        negative_paper_ids: List[str],
+        fields: Optional[List[str]] = None,
+        limit: int = 500,
+        save_to_file: bool = False,
+    ) -> dict:
+        r"""Fetches recommended papers from the Semantic Scholar
+        API based on the positive and negative paper IDs.
+
+        Args:
+            positive_paper_ids (list): A list of paper IDs (as strings)
+                that are positively correlated to the recommendation.
+            negative_paper_ids (list): A list of paper IDs (as strings)
+                that are negatively correlated to the recommendation.
+            fields (Optional[List[str]], optional): The fields to include in
+                the response (default: :obj:`None`). If not provided defaults
+                to ["title", "url", "citationCount", "authors",
+                "publicationTypes", "publicationDate", "openAccessPdf"].
+            limit (int, optional): The maximum number of recommended papers to
+                return (default: :obj:`500`).
+            save_to_file (bool, optional): If True, saves the response data to
+                a file (default: :obj:`False`).
+
+        Returns:
+            dict: A dictionary containing recommended papers sorted by
+                citation count.
+        """
+        if fields is None:
+            fields = [
+                "title",
+                "url",
+                "citationCount",
+                "authors",
+                "publicationTypes",
+                "publicationDate",
+                "openAccessPdf",
+            ]
+
+        url = "https://api.semanticscholar.org/recommendations/v1/papers"
+        query_params = {"fields": ",".join(fields), "limit": str(limit)}
+        data = {
+            "positive_paper_ids": positive_paper_ids,
+            "negative_paper_ids": negative_paper_ids,
+        }
+        try:
+            response = requests.post(url, params=query_params, json=data)
+            response.raise_for_status()
+            papers = response.json()
+            if save_to_file:
+                with open('recommended_papers.json', 'w') as output:
+                    json.dump(papers, output)
+            return papers
+        except requests.exceptions.RequestException as e:
+            return {"error": str(e)}
+        except ValueError:
+            return {
+                "error": "Response is not valid JSON",
+                "message": response.text,
+            }
+
+    def fetch_author_data(
+        self,
+        ids: List[str],
+        fields: Optional[List[str]] = None,
+        save_to_file: bool = False,
+    ) -> dict:
+        r"""Fetches author information from the Semantic Scholar
+        API based on author IDs.
+
+        Args:
+            ids (list): A list of author IDs (as strings) to fetch
+                data for.
+            fields (Optional[List[str]], optional): The fields to include in
+                the response (default: :obj:`None`). If not provided defaults
+                to ["name", "url", "paperCount", "hIndex", "papers"].
+            save_to_file (bool, optional): Whether to save the results to a
+                file (default: :obj:`False`).
+
+        Returns:
+            dict: The response data from the API or error information if
+                the request fails.
+        """
+        if fields is None:
+            fields = ["name", "url", "paperCount", "hIndex", "papers"]
+
+        url = f"{self.base_url}/author/batch"
+        query_params = {"fields": ",".join(fields)}
+        data = {"ids": ids}
+        try:
+            response = requests.post(url, params=query_params, json=data)
+            response.raise_for_status()
+            response_data = response.json()
+            if save_to_file:
+                with open('author_information.json', 'w') as output:
+                    json.dump(response_data, output)
+            return response_data
+        except requests.exceptions.RequestException as e:
+            return {"error": str(e)}
+        except ValueError:
+            return {
+                "error": "Response is not valid JSON",
+                "message": response.text,
+            }
+
+    def get_tools(self) -> List[FunctionTool]:
+        r"""Returns a list of FunctionTool objects representing the
+        functions in the toolkit.
+
+        Returns:
+            List[FunctionTool]: A list of FunctionTool objects
+                representing the functions in the toolkit.
+        """
+        return [
+            FunctionTool(self.fetch_paper_data_title),
+            FunctionTool(self.fetch_paper_data_id),
+            FunctionTool(self.fetch_bulk_paper_data),
+            FunctionTool(self.fetch_recommended_papers),
+            FunctionTool(self.fetch_author_data),
+        ]

camel/types/enums.py

@@ -37,21 +37,29 @@ class ModelType(UnifiedModelType, Enum):
     O1 = "o1"
     O1_PREVIEW = "o1-preview"
     O1_MINI = "o1-mini"
+    O3_MINI = "o3-mini"
 
     GLM_4 = "glm-4"
-    GLM_4V = 'glm-4v'
+    GLM_4V = "glm-4v"
+    GLM_4V_FLASH = "glm-4v-flash"
+    GLM_4V_PLUS_0111 = "glm-4v-plus-0111"
+    GLM_4_PLUS = "glm-4-plus"
+    GLM_4_AIR = "glm-4-air"
+    GLM_4_AIR_0111 = "glm-4-air-0111"
+    GLM_4_AIRX = "glm-4-airx"
+    GLM_4_LONG = "glm-4-long"
+    GLM_4_FLASHX = "glm-4-flashx"
+    GLM_4_FLASH = "glm-4-flash"
+    GLM_ZERO_PREVIEW = "glm-zero-preview"
     GLM_3_TURBO = "glm-3-turbo"
 
     # Groq platform models
     GROQ_LLAMA_3_1_8B = "llama-3.1-8b-instant"
-    GROQ_LLAMA_3_1_70B = "llama-3.1-70b-versatile"
-    GROQ_LLAMA_3_1_405B = "llama-3.1-405b-reasoning"
     GROQ_LLAMA_3_3_70B = "llama-3.3-70b-versatile"
     GROQ_LLAMA_3_3_70B_PREVIEW = "llama-3.3-70b-specdec"
     GROQ_LLAMA_3_8B = "llama3-8b-8192"
     GROQ_LLAMA_3_70B = "llama3-70b-8192"
     GROQ_MIXTRAL_8_7B = "mixtral-8x7b-32768"
-    GROQ_GEMMA_7B_IT = "gemma-7b-it"
     GROQ_GEMMA_2_9B_IT = "gemma2-9b-it"
 
     # TogetherAI platform models support tool calling
@@ -67,6 +75,17 @@ class ModelType(UnifiedModelType, Enum):
     SAMBA_LLAMA_3_1_70B = "Meta-Llama-3.1-70B-Instruct"
     SAMBA_LLAMA_3_1_405B = "Meta-Llama-3.1-405B-Instruct"
 
+    # SGLang models support tool calling
+    SGLANG_LLAMA_3_1_8B = "meta-llama/Meta-Llama-3.1-8B-Instruct"
+    SGLANG_LLAMA_3_1_70B = "meta-llama/Meta-Llama-3.1-70B-Instruct"
+    SGLANG_LLAMA_3_1_405B = "meta-llama/Meta-Llama-3.1-405B-Instruct"
+    SGLANG_LLAMA_3_2_1B = "meta-llama/Llama-3.2-1B-Instruct"
+    SGLANG_MIXTRAL_NEMO = "mistralai/Mistral-Nemo-Instruct-2407"
+    SGLANG_MISTRAL_7B = "mistralai/Mistral-7B-Instruct-v0.3"
+    SGLANG_QWEN_2_5_7B = "Qwen/Qwen2.5-7B-Instruct"
+    SGLANG_QWEN_2_5_32B = "Qwen/Qwen2.5-32B-Instruct"
+    SGLANG_QWEN_2_5_72B = "Qwen/Qwen2.5-72B-Instruct"
+
     STUB = "stub"
 
     # Legacy anthropic models
@@ -97,9 +116,12 @@ class ModelType(UnifiedModelType, Enum):
     NVIDIA_LLAMA3_3_70B_INSTRUCT = "meta/llama-3.3-70b-instruct"
 
     # Gemini models
+    GEMINI_2_0_FLASH = "gemini-2.0-flash-exp"
+    GEMINI_2_0_FLASH_THINKING = "gemini-2.0-flash-thinking-exp"
+    GEMINI_2_0_PRO_EXP = "gemini-2.0-pro-exp-02-05"
+    GEMINI_2_0_FLASH_LITE_PREVIEW = "gemini-2.0-flash-lite-preview-02-05"
     GEMINI_1_5_FLASH = "gemini-1.5-flash"
     GEMINI_1_5_PRO = "gemini-1.5-pro"
-    GEMINI_EXP_1114 = "gemini-exp-1114"
 
     # Mistral AI models
     MISTRAL_3B = "ministral-3b-latest"
@@ -136,6 +158,7 @@ class ModelType(UnifiedModelType, Enum):
     QWEN_MATH_TURBO = "qwen-math-turbo"
     QWEN_CODER_TURBO = "qwen-coder-turbo"
     QWEN_2_5_CODER_32B = "qwen2.5-coder-32b-instruct"
+    QWEN_2_5_VL_72B = "qwen2.5-vl-72b-instruct"
     QWEN_2_5_72B = "qwen2.5-72b-instruct"
     QWEN_2_5_32B = "qwen2.5-32b-instruct"
     QWEN_2_5_14B = "qwen2.5-14b-instruct"
@@ -161,6 +184,25 @@ class ModelType(UnifiedModelType, Enum):
     INTERNLM2_5_LATEST = "internlm2.5-latest"
     INTERNLM2_PRO_CHAT = "internlm2-pro-chat"
 
+    # Moonshot models
+    MOONSHOT_V1_8K = "moonshot-v1-8k"
+    MOONSHOT_V1_32K = "moonshot-v1-32k"
+    MOONSHOT_V1_128K = "moonshot-v1-128k"
+
+    # SiliconFlow models support tool calling
+    SILICONFLOW_DEEPSEEK_V2_5 = "deepseek-ai/DeepSeek-V2.5"
+    SILICONFLOW_DEEPSEEK_V3 = "deepseek-ai/DeepSeek-V3"
+    SILICONFLOW_INTERN_LM2_5_20B_CHAT = "internlm/internlm2_5-20b-chat"
+    SILICONFLOW_INTERN_LM2_5_7B_CHAT = "internlm/internlm2_5-7b-chat"
+    SILICONFLOW_PRO_INTERN_LM2_5_7B_CHAT = "Pro/internlm/internlm2_5-7b-chat"
+    SILICONFLOW_QWEN2_5_72B_INSTRUCT = "Qwen/Qwen2.5-72B-Instruct"
+    SILICONFLOW_QWEN2_5_32B_INSTRUCT = "Qwen/Qwen2.5-32B-Instruct"
+    SILICONFLOW_QWEN2_5_14B_INSTRUCT = "Qwen/Qwen2.5-14B-Instruct"
+    SILICONFLOW_QWEN2_5_7B_INSTRUCT = "Qwen/Qwen2.5-7B-Instruct"
+    SILICONFLOW_PRO_QWEN2_5_7B_INSTRUCT = "Pro/Qwen/Qwen2.5-7B-Instruct"
+    SILICONFLOW_THUDM_GLM_4_9B_CHAT = "THUDM/glm-4-9b-chat"
+    SILICONFLOW_PRO_THUDM_GLM_4_9B_CHAT = "Pro/THUDM/glm-4-9b-chat"
+
     def __str__(self):
         return self.value
 
@@ -190,6 +232,10 @@ class ModelType(UnifiedModelType, Enum):
                 self.is_internlm,
                 self.is_together,
                 self.is_sambanova,
+                self.is_groq,
+                self.is_sglang,
+                self.is_moonshot,
+                self.is_siliconflow,
             ]
         )
 
@@ -205,6 +251,7 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.O1,
             ModelType.O1_PREVIEW,
             ModelType.O1_MINI,
+            ModelType.O3_MINI,
         }
 
     @property
@@ -227,6 +274,16 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.GLM_3_TURBO,
             ModelType.GLM_4,
             ModelType.GLM_4V,
+            ModelType.GLM_4V_FLASH,
+            ModelType.GLM_4V_PLUS_0111,
+            ModelType.GLM_4_PLUS,
+            ModelType.GLM_4_AIR,
+            ModelType.GLM_4_AIR_0111,
+            ModelType.GLM_4_AIRX,
+            ModelType.GLM_4_LONG,
+            ModelType.GLM_4_FLASHX,
+            ModelType.GLM_4_FLASH,
+            ModelType.GLM_ZERO_PREVIEW,
         }
 
     @property
@@ -252,14 +309,11 @@ class ModelType(UnifiedModelType, Enum):
         r"""Returns whether this type of models is served by Groq."""
         return self in {
             ModelType.GROQ_LLAMA_3_1_8B,
-            ModelType.GROQ_LLAMA_3_1_70B,
-            ModelType.GROQ_LLAMA_3_1_405B,
             ModelType.GROQ_LLAMA_3_3_70B,
             ModelType.GROQ_LLAMA_3_3_70B_PREVIEW,
             ModelType.GROQ_LLAMA_3_8B,
             ModelType.GROQ_LLAMA_3_70B,
             ModelType.GROQ_MIXTRAL_8_7B,
-            ModelType.GROQ_GEMMA_7B_IT,
             ModelType.GROQ_GEMMA_2_9B_IT,
         }
 
@@ -277,7 +331,7 @@ class ModelType(UnifiedModelType, Enum):
 
     @property
     def is_sambanova(self) -> bool:
-        r"""Returns whether this type of models is served by SambaNova AI."""
+        r"""Returns whether this type of model is served by SambaNova AI."""
         return self in {
             ModelType.SAMBA_LLAMA_3_1_8B,
             ModelType.SAMBA_LLAMA_3_1_70B,
@@ -326,9 +380,12 @@ class ModelType(UnifiedModelType, Enum):
             bool: Whether this type of models is gemini.
         """
         return self in {
+            ModelType.GEMINI_2_0_FLASH,
             ModelType.GEMINI_1_5_FLASH,
             ModelType.GEMINI_1_5_PRO,
-            ModelType.GEMINI_EXP_1114,
+            ModelType.GEMINI_2_0_FLASH_THINKING,
+            ModelType.GEMINI_2_0_PRO_EXP,
+            ModelType.GEMINI_2_0_FLASH_LITE_PREVIEW,
         }
 
     @property
@@ -391,6 +448,7 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.QWEN_MATH_TURBO,
             ModelType.QWEN_CODER_TURBO,
             ModelType.QWEN_2_5_CODER_32B,
+            ModelType.QWEN_2_5_VL_72B,
             ModelType.QWEN_2_5_72B,
             ModelType.QWEN_2_5_32B,
             ModelType.QWEN_2_5_14B,
@@ -413,6 +471,45 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.INTERNLM2_PRO_CHAT,
         }
 
+    @property
+    def is_moonshot(self) -> bool:
+        return self in {
+            ModelType.MOONSHOT_V1_8K,
+            ModelType.MOONSHOT_V1_32K,
+            ModelType.MOONSHOT_V1_128K,
+        }
+
+    @property
+    def is_sglang(self) -> bool:
+        return self in {
+            ModelType.SGLANG_LLAMA_3_1_8B,
+            ModelType.SGLANG_LLAMA_3_1_70B,
+            ModelType.SGLANG_LLAMA_3_1_405B,
+            ModelType.SGLANG_LLAMA_3_2_1B,
+            ModelType.SGLANG_MIXTRAL_NEMO,
+            ModelType.SGLANG_MISTRAL_7B,
+            ModelType.SGLANG_QWEN_2_5_7B,
+            ModelType.SGLANG_QWEN_2_5_32B,
+            ModelType.SGLANG_QWEN_2_5_72B,
+        }
+
+    @property
+    def is_siliconflow(self) -> bool:
+        return self in {
+            ModelType.SILICONFLOW_DEEPSEEK_V2_5,
+            ModelType.SILICONFLOW_DEEPSEEK_V3,
+            ModelType.SILICONFLOW_INTERN_LM2_5_20B_CHAT,
+            ModelType.SILICONFLOW_INTERN_LM2_5_7B_CHAT,
+            ModelType.SILICONFLOW_PRO_INTERN_LM2_5_7B_CHAT,
+            ModelType.SILICONFLOW_QWEN2_5_72B_INSTRUCT,
+            ModelType.SILICONFLOW_QWEN2_5_32B_INSTRUCT,
+            ModelType.SILICONFLOW_QWEN2_5_14B_INSTRUCT,
+            ModelType.SILICONFLOW_QWEN2_5_7B_INSTRUCT,
+            ModelType.SILICONFLOW_PRO_QWEN2_5_7B_INSTRUCT,
+            ModelType.SILICONFLOW_THUDM_GLM_4_9B_CHAT,
+            ModelType.SILICONFLOW_PRO_THUDM_GLM_4_9B_CHAT,
+        }
+
     @property
     def token_limit(self) -> int:
         r"""Returns the maximum token limit for a given model.
@@ -440,13 +537,15 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.GROQ_LLAMA_3_8B,
             ModelType.GROQ_LLAMA_3_70B,
             ModelType.GROQ_LLAMA_3_3_70B_PREVIEW,
-            ModelType.GROQ_GEMMA_7B_IT,
             ModelType.GROQ_GEMMA_2_9B_IT,
             ModelType.GLM_3_TURBO,
             ModelType.GLM_4,
             ModelType.QWEN_VL_PLUS,
             ModelType.NVIDIA_LLAMA3_70B,
             ModelType.TOGETHER_MISTRAL_7B,
+            ModelType.MOONSHOT_V1_8K,
+            ModelType.GLM_4V_FLASH,
+            ModelType.GLM_4_AIRX,
         }:
             return 8_192
         elif self in {
@@ -459,6 +558,8 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.YI_LARGE_RAG,
             ModelType.SAMBA_LLAMA_3_1_8B,
             ModelType.SAMBA_LLAMA_3_1_405B,
+            ModelType.GLM_4V_PLUS_0111,
+            ModelType.GLM_ZERO_PREVIEW,
         }:
             return 16_384
         elif self in {
@@ -479,6 +580,8 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.INTERNLM2_5_LATEST,
             ModelType.INTERNLM2_PRO_CHAT,
             ModelType.TOGETHER_MIXTRAL_8_7B,
+            ModelType.SGLANG_MISTRAL_7B,
+            ModelType.MOONSHOT_V1_32K,
         }:
             return 32_768
         elif self in {
@@ -504,6 +607,7 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.MISTRAL_8B,
             ModelType.MISTRAL_3B,
             ModelType.QWEN_2_5_CODER_32B,
+            ModelType.QWEN_2_5_VL_72B,
             ModelType.QWEN_2_5_72B,
             ModelType.QWEN_2_5_32B,
             ModelType.QWEN_2_5_14B,
@@ -518,12 +622,21 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.NVIDIA_LLAMA3_3_70B_INSTRUCT,
             ModelType.GROQ_LLAMA_3_3_70B,
             ModelType.SAMBA_LLAMA_3_1_70B,
+            ModelType.SGLANG_LLAMA_3_1_8B,
+            ModelType.SGLANG_LLAMA_3_1_70B,
+            ModelType.SGLANG_LLAMA_3_1_405B,
+            ModelType.SGLANG_LLAMA_3_2_1B,
+            ModelType.SGLANG_MIXTRAL_NEMO,
+            ModelType.MOONSHOT_V1_128K,
+            ModelType.GLM_4_PLUS,
+            ModelType.GLM_4_AIR,
+            ModelType.GLM_4_AIR_0111,
+            ModelType.GLM_4_FLASHX,
+            ModelType.GLM_4_FLASH,
         }:
             return 128_000
         elif self in {
             ModelType.GROQ_LLAMA_3_1_8B,
-            ModelType.GROQ_LLAMA_3_1_70B,
-            ModelType.GROQ_LLAMA_3_1_405B,
             ModelType.QWEN_PLUS,
             ModelType.QWEN_TURBO,
             ModelType.QWEN_CODER_TURBO,
@@ -531,10 +644,14 @@ class ModelType(UnifiedModelType, Enum):
             ModelType.TOGETHER_LLAMA_3_1_70B,
             ModelType.TOGETHER_LLAMA_3_1_405B,
             ModelType.TOGETHER_LLAMA_3_3_70B,
+            ModelType.SGLANG_QWEN_2_5_7B,
+            ModelType.SGLANG_QWEN_2_5_32B,
+            ModelType.SGLANG_QWEN_2_5_72B,
         }:
             return 131_072
         elif self in {
             ModelType.O1,
+            ModelType.O3_MINI,
             ModelType.CLAUDE_2_1,
             ModelType.CLAUDE_3_OPUS,
             ModelType.CLAUDE_3_SONNET,
@@ -549,9 +666,13 @@ class ModelType(UnifiedModelType, Enum):
         }:
             return 256_000
         elif self in {
+            ModelType.GEMINI_2_0_FLASH,
             ModelType.GEMINI_1_5_FLASH,
             ModelType.GEMINI_1_5_PRO,
-            ModelType.GEMINI_EXP_1114,  # Not given in docs, assuming the same
+            ModelType.GEMINI_2_0_FLASH_THINKING,
+            ModelType.GEMINI_2_0_FLASH_LITE_PREVIEW,
+            ModelType.GEMINI_2_0_PRO_EXP,  # Not given in doc, assume the same
+            ModelType.GLM_4_LONG,
         }:
             return 1_048_576
         elif self in {
@@ -567,6 +688,11 @@ class EmbeddingModelType(Enum):
     TEXT_EMBEDDING_3_SMALL = "text-embedding-3-small"
     TEXT_EMBEDDING_3_LARGE = "text-embedding-3-large"
 
+    JINA_EMBEDDINGS_V3 = "jina-embeddings-v3"
+    JINA_CLIP_V2 = "jina-clip-v2"
+    JINA_COLBERT_V2 = "jina-colbert-v2"
+    JINA_EMBEDDINGS_V2_BASE_CODE = "jina-embeddings-v2-base-code"
+
     MISTRAL_EMBED = "mistral-embed"
 
     @property
@@ -578,6 +704,16 @@ class EmbeddingModelType(Enum):
             EmbeddingModelType.TEXT_EMBEDDING_3_LARGE,
         }
 
+    @property
+    def is_jina(self) -> bool:
+        r"""Returns whether this type of models is an Jina model."""
+        return self in {
+            EmbeddingModelType.JINA_EMBEDDINGS_V3,
+            EmbeddingModelType.JINA_CLIP_V2,
+            EmbeddingModelType.JINA_COLBERT_V2,
+            EmbeddingModelType.JINA_EMBEDDINGS_V2_BASE_CODE,
+        }
+
     @property
     def is_mistral(self) -> bool:
         r"""Returns whether this type of models is an Mistral-released
@@ -589,7 +725,20 @@ class EmbeddingModelType(Enum):
 
     @property
     def output_dim(self) -> int:
-        if self is EmbeddingModelType.TEXT_EMBEDDING_ADA_2:
+        if self in {
+            EmbeddingModelType.JINA_COLBERT_V2,
+        }:
+            return 128
+        elif self in {
+            EmbeddingModelType.JINA_EMBEDDINGS_V2_BASE_CODE,
+        }:
+            return 768
+        elif self in {
+            EmbeddingModelType.JINA_EMBEDDINGS_V3,
+            EmbeddingModelType.JINA_CLIP_V2,
+        }:
+            return 1024
+        elif self is EmbeddingModelType.TEXT_EMBEDDING_ADA_2:
             return 1536
         elif self is EmbeddingModelType.TEXT_EMBEDDING_3_SMALL:
             return 1536
@@ -709,6 +858,8 @@ class ModelPlatformType(Enum):
     DEEPSEEK = "deepseek"
     SGLANG = "sglang"
     INTERNLM = "internlm"
+    MOONSHOT = "moonshot"
+    SILICONFLOW = "siliconflow"
 
     @property
     def is_openai(self) -> bool:
@@ -816,6 +967,16 @@ class ModelPlatformType(Enum):
         r"""Returns whether this platform is InternLM."""
         return self is ModelPlatformType.INTERNLM
 
+    @property
+    def is_moonshot(self) -> bool:
+        r"""Returns whether this platform is Moonshot model."""
+        return self is ModelPlatformType.MOONSHOT
+
+    @property
+    def is_siliconflow(self) -> bool:
+        r"""Returns whether this platform is SiliconFlow."""
+        return self is ModelPlatformType.SILICONFLOW
+
 
 class AudioModelType(Enum):
     TTS_1 = "tts-1"