Merge pull request #2569 from hlohaus/12Jan

Read FinishReason and Usage from Gemini API

g4f/Provider/Jmuz.py

@@ -5,7 +5,7 @@
 
 class Jmuz(OpenaiAPI):
     label = "Jmuz"
-    url = "https://jmuz.me"
+    url = "https://discord.gg/qXfu24JmsB"
     login_url = None
     api_base = "https://jmuz.me/gpt/api/v2"
     api_key = "prod"
@@ -15,7 +15,7 @@ class Jmuz(OpenaiAPI):
     supports_stream = True
     supports_system_message = False
 
-    default_model = 'gpt-4o'
+    default_model = "gpt-4o"
     model_aliases = {
         "gemini": "gemini-exp",
         "deepseek-chat": "deepseek-2.5",
@@ -29,13 +29,7 @@ def get_models(cls):
         return cls.models
 
     @classmethod
-    def get_model(cls, model: str, **kwargs) -> str:
-        if model in cls.get_models():
-            return model
-        return cls.default_model
-
-    @classmethod
-    def create_async_generator(
+    async def create_async_generator(
             cls,
             model: str,
             messages: Messages,
@@ -52,12 +46,20 @@ def create_async_generator(
             "cache-control": "no-cache",
             "user-agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/129.0.0.0 Safari/537.36"
         }
-        return super().create_async_generator(
+        started = False
+        async for chunk in super().create_async_generator(
             model=model,
             messages=messages,
             api_base=cls.api_base,
             api_key=cls.api_key,
             stream=cls.supports_stream,
             headers=headers,
             **kwargs
-        )
+        ):
+            if isinstance(chunk, str) and cls.url in chunk:
+                continue
+            if isinstance(chunk, str) and not started:
+                chunk = chunk.lstrip()
+            if chunk:
+                started = True
+                yield chunk

g4f/Provider/Pizzagpt.py

@@ -10,7 +10,7 @@
 class Pizzagpt(AsyncGeneratorProvider, ProviderModelMixin):
     url = "https://www.pizzagpt.it"
     api_endpoint = "/api/chatx-completion"
-    working = True
+    working = False
     default_model = 'gpt-4o-mini'
 
     @classmethod
@@ -46,6 +46,6 @@ async def create_async_generator(
                 response_json = await response.json()
                 content = response_json.get("answer", response_json).get("content")
                 if content:
-                    if "misuse detected. please get in touch" in content:
+                    if "Misuse detected. please get in touch" in content:
                         raise ValueError(content)
                     yield content

g4f/Provider/needs_auth/Custom.py

@@ -3,9 +3,10 @@
 from .OpenaiAPI import OpenaiAPI
 
 class Custom(OpenaiAPI):
-    label = "Custom"
+    label = "Custom Provider"
     url = None
-    login_url = "http://localhost:8080"
+    login_url = None
     working = True
     api_base = "http://localhost:8080/v1"
-    needs_auth = False
+    needs_auth = False
+    sort_models = False

g4f/Provider/needs_auth/GeminiPro.py

@@ -3,12 +3,14 @@
 import base64
 import json
 import requests
+from typing import Optional
 from aiohttp import ClientSession, BaseConnector
 
 from ...typing import AsyncResult, Messages, ImagesType
 from ...image import to_bytes, is_accepted_format
 from ...errors import MissingAuthError
 from ...requests.raise_for_status import raise_for_status
+from ...providers.response import Usage, FinishReason
 from ..base_provider import AsyncGeneratorProvider, ProviderModelMixin
 from ..helper import get_connector
 from ... import debug
@@ -62,6 +64,7 @@ async def create_async_generator(
         api_base: str = api_base,
         use_auth_header: bool = False,
         images: ImagesType = None,
+        tools: Optional[list] = None,
         connector: BaseConnector = None,
         **kwargs
     ) -> AsyncResult:
@@ -104,7 +107,10 @@ async def create_async_generator(
                     "maxOutputTokens": kwargs.get("max_tokens"),
                     "topP": kwargs.get("top_p"),
                     "topK": kwargs.get("top_k"),
-                }
+                },
+                 "tools": [{
+                    "functionDeclarations": tools
+                 }] if tools else None
             }
             system_prompt = "\n".join(
                 message["content"]
@@ -128,6 +134,15 @@ async def create_async_generator(
                                 data = b"".join(lines)
                                 data = json.loads(data)
                                 yield data["candidates"][0]["content"]["parts"][0]["text"]
+                                if "finishReason" in data["candidates"][0]:
+                                    yield FinishReason(data["candidates"][0]["finishReason"].lower())
+                                usage = data.get("usageMetadata")
+                                if usage:
+                                    yield Usage(
+                                        prompt_tokens=usage.get("promptTokenCount"),
+                                        completion_tokens=usage.get("candidatesTokenCount"),
+                                        total_tokens=usage.get("totalTokenCount")
+                                    )
                             except:
                                 data = data.decode(errors="ignore") if isinstance(data, bytes) else data
                                 raise RuntimeError(f"Read chunk failed: {data}")

g4f/Provider/needs_auth/HuggingFace.py

@@ -23,21 +23,25 @@ class HuggingFace(AsyncGeneratorProvider, ProviderModelMixin):
     default_model = HuggingChat.default_model
     default_image_model = HuggingChat.default_image_model
     model_aliases = HuggingChat.model_aliases
+    extra_models = [
+        "meta-llama/Llama-3.2-11B-Vision-Instruct",
+        "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF",
+        "NousResearch/Hermes-3-Llama-3.1-8B",
+    ]
 
     @classmethod
     def get_models(cls) -> list[str]:
         if not cls.models:
             url = "https://huggingface.co/api/models?inference=warm&pipeline_tag=text-generation"
             models = [model["id"] for model in requests.get(url).json()]
-            models.append("meta-llama/Llama-3.2-11B-Vision-Instruct")
-            models.append("nvidia/Llama-3.1-Nemotron-70B-Instruct-HF")
+            models.extend(cls.extra_models)
             models.sort()
             if not cls.image_models:
                 url = "https://huggingface.co/api/models?pipeline_tag=text-to-image"
                 cls.image_models = [model["id"] for model in requests.get(url).json() if model["trendingScore"] >= 20]
                 cls.image_models.sort()
                 models.extend(cls.image_models)
-            cls.models = models
+            cls.models = list(set(models))
         return cls.models
 
     @classmethod

g4f/Provider/needs_auth/OpenaiAPI.py

@@ -23,6 +23,7 @@ class OpenaiAPI(AsyncGeneratorProvider, ProviderModelMixin, RaiseErrorMixin):
     supports_system_message = True
     default_model = ""
     fallback_models = []
+    sort_models = True
 
     @classmethod
     def get_models(cls, api_key: str = None, api_base: str = None) -> list[str]:
@@ -36,8 +37,11 @@ def get_models(cls, api_key: str = None, api_base: str = None) -> list[str]:
                 response = requests.get(f"{api_base}/models", headers=headers)
                 raise_for_status(response)
                 data = response.json()
-                cls.models = [model.get("id") for model in (data.get("data") if isinstance(data, dict) else data)]
-                cls.models.sort()
+                data = data.get("data") if isinstance(data, dict) else data
+                cls.image_models = [model.get("id") for model in data if model.get("image")]
+                cls.models = [model.get("id") for model in data]
+                if cls.sort_models:
+                    cls.models.sort()
             except Exception as e:
                 debug.log(e)
                 cls.models = cls.fallback_models

g4f/api/__init__.py

@@ -215,12 +215,16 @@ async def read_root_v1():
             HTTP_200_OK: {"model": List[ModelResponseModel]},
         })
         async def models():
-            return [{
-                'id': model_id,
-                'object': 'model',
-                'created': 0,
-                'owned_by': model.base_provider
-            } for model_id, model in g4f.models.ModelUtils.convert.items()]
+            return {
+                "object": "list",
+                "data": [{
+                    "id": model_id,
+                    "object": "model",
+                    "created": 0,
+                    "owned_by": model.base_provider,
+                    "image": isinstance(model, g4f.models.ImageModel),
+                } for model_id, model in g4f.models.ModelUtils.convert.items()]
+            }
 
         @self.app.get("/v1/models/{model_name}", responses={
             HTTP_200_OK: {"model": ModelResponseModel},

g4f/gui/client/index.html

@@ -143,7 +143,7 @@ <h3>Settings</h3>
                 <label for="refine" class="toogle" title=""></label>
             </div>
             <div class="field box">
-                <label for="systemPrompt" class="label" title="">Default for System prompt</label>
+                <label for="systemPrompt" class="label" title="">System prompt</label>
                 <textarea id="systemPrompt" placeholder="You are a helpful assistant."></textarea>
             </div>
             <div class="field box">
@@ -157,6 +157,14 @@ <h3>Settings</h3>
                     document.getElementById('recognition-language').placeholder = navigator.language;
                 </script>
             </div>
+            <div class="field box">
+                <label for="Custom-api_base" class="label" title="">Custom Provider (Base Url):</label>
+                <input type="text" id="Custom-api_base" name="Custom[api_base]" placeholder="http://localhost:8080/v1"/>
+            </div>
+            <div class="field box hidden">
+                <label for="Custom-api_key" class="label" title="">Custom Provider:</label>
+                <input type="text" id="Custom-api_key" name="Custom[api_key]" placeholder="api_key"/>
+            </div>
             <div class="field box hidden">
                 <label for="BingCreateImages-api_key" class="label" title="">Microsoft Designer in Bing:</label>
                 <input type="text" id="BingCreateImages-api_key" name="BingCreateImages[api_key]" placeholder="&quot;_U&quot; cookie"/>
@@ -254,10 +262,7 @@ <h3>Settings</h3>
                      <option value="gpt-4o">gpt-4o</option>
                      <option value="gpt-4o-mini">gpt-4o-mini</option>
                      <option value="llama-3.1-70b">llama-3.1-70b</option>
-                     <option value="llama-3.1-405b">llama-3.1-405b</option>
                      <option value="mixtral-8x7b">mixtral-8x7b</option>
-                     <option value="gemini-pro">gemini-pro</option>
-                     <option value="gemini-flash">gemini-flash</option>
                      <option value="claude-3.5-sonnet">claude-3.5-sonnet</option>
                      <option value="flux">flux (Image Generation)</option>
                      <option value="dall-e-3">dall-e-3 (Image Generation)</option>

g4f/gui/client/static/js/chat.v1.js

@@ -351,11 +351,6 @@ const handle_ask = async () => {
     await count_input()
     await add_conversation(window.conversation_id);
 
-    if ("text" in fileInput.dataset) {
-        message += '\n```' + fileInput.dataset.type + '\n'; 
-        message += fileInput.dataset.text;
-        message += '\n```'
-    }
     let message_index = await add_message(window.conversation_id, "user", message);
     let message_id = get_message_id();
 
@@ -799,6 +794,7 @@ const ask_gpt = async (message_id, message_index = -1, regenerate = false, provi
         const files = input && input.files.length > 0 ? input.files : null;
         const download_images = document.getElementById("download_images")?.checked;
         const api_key = get_api_key_by_provider(provider);
+        const api_base = provider == "Custom" ? document.getElementById(`${provider}-api_base`).value : null;
         const ignored = Array.from(settings.querySelectorAll("input.provider:not(:checked)")).map((el)=>el.value);
         await api("conversation", {
             id: message_id,
@@ -811,6 +807,7 @@ const ask_gpt = async (message_id, message_index = -1, regenerate = false, provi
             action: action,
             download_images: download_images,
             api_key: api_key,
+            api_base: api_base,
             ignored: ignored,
         }, files, message_id, scroll);
         content_map.update_timeouts.forEach((timeoutId)=>clearTimeout(timeoutId));
@@ -1066,7 +1063,7 @@ const load_conversation = async (conversation_id, scroll=true) => {
         }
         buffer = buffer.replace(/ \[aborted\]$/g, "").replace(/ \[error\]$/g, "");
         new_content = item.content.replace(/ \[aborted\]$/g, "").replace(/ \[error\]$/g, "");
-        buffer += merge_messages(buffer, new_content);
+        buffer = merge_messages(buffer, new_content);
         last_model = item.provider?.model;
         providers.push(item.provider?.name);
         let next_i = parseInt(i) + 1;
@@ -1176,9 +1173,6 @@ const load_conversation = async (conversation_id, scroll=true) => {
     if (window.GPTTokenizer_cl100k_base) {
         const filtered = prepare_messages(messages, null, true, false);
         if (filtered.length > 0) {
-            if (GPTTokenizer_o200k_base && last_model?.startsWith("gpt-4o") || last_model?.startsWith("o1")) {
-                return GPTTokenizer_o200k_base?.encodeChat(filtered, last_model).length;
-            }
             last_model = last_model?.startsWith("gpt-3") ? "gpt-3.5-turbo" : "gpt-4"
             let count_total = GPTTokenizer_cl100k_base?.encodeChat(filtered, last_model).length
             if (count_total > 0) {
@@ -1890,7 +1884,6 @@ setTimeout(load_version, 100);
 
 fileInput.addEventListener('click', async (event) => {
     fileInput.value = '';
-    delete fileInput.dataset.text;
 });
 
 async function upload_cookies() {
@@ -1920,7 +1913,6 @@ function formatFileSize(bytes) {
 async function upload_files(fileInput) {
     const paperclip = document.querySelector(".user-input .fa-paperclip");
     const bucket_id = uuid();
-    delete fileInput.dataset.text;
     paperclip.classList.add("blink");
 
     const formData = new FormData();
@@ -1980,8 +1972,7 @@ fileInput.addEventListener('change', async (event) => {
         if (type == "json") {
             const reader = new FileReader();
             reader.addEventListener('load', async (event) => {
-                fileInput.dataset.text = event.target.result;
-                const data = JSON.parse(fileInput.dataset.text);
+                const data = JSON.parse(event.target.result);
                 if (data.options && "g4f" in data.options) {
                     let count = 0;
                     Object.keys(data).forEach(key => {
@@ -1990,7 +1981,6 @@ fileInput.addEventListener('change', async (event) => {
                             count += 1;
                         }
                     });
-                    delete fileInput.dataset.text;
                     await load_conversations();
                     fileInput.value = "";
                     inputCount.innerText = `${count} Conversations were imported successfully`;
@@ -2012,8 +2002,6 @@ fileInput.addEventListener('change', async (event) => {
             });
             reader.readAsText(fileInput.files[0]);
         }
-    } else {
-        delete fileInput.dataset.text;
     }
 });
 
@@ -2033,16 +2021,19 @@ function get_selected_model() {
 }
 
 async function api(ressource, args=null, files=null, message_id=null, scroll=true) {
-    let api_key;
+    const headers = {};
     if (ressource == "models" && args) {
         api_key = get_api_key_by_provider(args);
+        if (api_key) {
+            headers.x_api_key = api_key;
+        }
+        api_base = args == "Custom" ? document.getElementById(`${args}-api_base`).value : null;
+        if (api_base) {
+            headers.x_api_base = api_base;
+        }
         ressource = `${ressource}/${args}`;
     }
     const url = `/backend-api/v2/${ressource}`;
-    const headers = {};
-    if (api_key) {
-        headers.x_api_key = api_key;
-    }
     if (ressource == "conversation") {
         let body = JSON.stringify(args);
         headers.accept = 'text/event-stream';
@@ -2224,7 +2215,7 @@ if (SpeechRecognition) {
     };
     recognition.onend = function() {
         messageInput.value = `${startValue ? startValue + "\n" : ""}${buffer}`;
-        if (!microLabel.classList.contains("recognition")) {
+        if (microLabel.classList.contains("recognition")) {
             recognition.start();
         } else {
             messageInput.readOnly = false;

g4f/gui/server/api.py

@@ -37,12 +37,12 @@ def get_models():
         for model, providers in models.__models__.values()]
 
     @staticmethod
-    def get_provider_models(provider: str, api_key: str = None):
+    def get_provider_models(provider: str, api_key: str = None, api_base: str = None):
         if provider in ProviderUtils.convert:
             provider = ProviderUtils.convert[provider]
             if issubclass(provider, ProviderModelMixin):
                 if api_key is not None and "api_key" in signature(provider.get_models).parameters:
-                    models = provider.get_models(api_key=api_key)
+                    models = provider.get_models(api_key=api_key, api_base=api_base)
                 else:
                     models = provider.get_models()
                 return [
@@ -90,6 +90,9 @@ def _prepare_conversation_kwargs(self, json_data: dict, kwargs: dict):
         api_key = json_data.get("api_key")
         if api_key is not None:
             kwargs["api_key"] = api_key
+        api_base = json_data.get("api_base")
+        if api_base is not None:
+            kwargs["api_base"] = api_base
         kwargs["tool_calls"] = [{
             "function": {
                 "name": "bucket_tool"