Merge branch 'main' into deploy

.github/renovate.json

@@ -0,0 +1,6 @@
+{
+  "$schema": "https://docs.renovatebot.com/renovate-schema.json",
+  "extends": [
+    "config:recommended"
+  ]
+}

frontend/app/HeroPage.svelte

@@ -2,18 +2,18 @@
   import { TokenSquare } from "@promplate/pattern";
   import { onMount } from "svelte";
   import { sineInOut } from "svelte/easing";
-  import { tweened } from "svelte/motion";
+  import { Tween } from "svelte/motion";
 
   const length = 20;
-  const i = tweened(length * 1.3, { easing: sineInOut, duration: 700 });
+  const i = new Tween(length * 1.3, { easing: sineInOut, duration: 700 });
 
-  onMount(() => ($i = 0));
+  onMount(() => i.set(0));
 </script>
 
 <div class="grid h-screen select-none place-items-center bg-neutral-950">
   <div class="flex flex-col gap-2 text-neutral-2">
     <div class="text-sm">
-      <TokenSquare {length} isHighlighted={({ x, y }) => y - x >= $i} />
+      <TokenSquare {length} isHighlighted={({ x, y }) => y - x >= i.current} />
     </div>
     <h1 class="mx-0.5 w-fit text-white font-sans">
       Promplate

frontend/app/main.ts

@@ -1,9 +1,8 @@
+import { mount } from "svelte";
 import App from "./App.svelte";
 import "@unocss/reset/tailwind-compat.css";
 import "uno.css";
 
-const app = new App({
-  target: document.getElementById("app")!,
+export default mount(App, {
+  target: document.body,
 });
-
-export default app;

frontend/package.json

@@ -12,24 +12,24 @@
     "check": "svelte-check --tsconfig ./tsconfig.json"
   },
   "devDependencies": {
-    "@antfu/eslint-config": "^3.0.0",
-    "@promplate/pattern": "^0.0.1",
-    "@sveltejs/vite-plugin-svelte": "^3.1.1",
+    "@antfu/eslint-config": "^4.1.0",
+    "@promplate/pattern": "^0.0.3",
+    "@sveltejs/vite-plugin-svelte": "^5.0.0",
     "@tsconfig/svelte": "^5.0.4",
-    "@unocss/eslint-plugin": "^0.61.0",
-    "@unocss/extractor-svelte": "^0.61.0",
-    "@unocss/reset": "^0.61.0",
+    "@unocss/eslint-plugin": "^66.0.0",
+    "@unocss/extractor-svelte": "^66.0.0",
+    "@unocss/reset": "^66.0.0",
     "eslint": "^9.4.0",
     "eslint-plugin-format": "^1.0.1",
-    "eslint-plugin-svelte": "^2.39.3",
+    "eslint-plugin-svelte": "^2.45.1",
     "prettier": "^3.3.2",
-    "prettier-plugin-svelte": "^3.2.4",
-    "svelte": "^4.2.18",
+    "prettier-plugin-svelte": "^3.2.6",
+    "svelte": "^5.0.0",
     "svelte-check": "^4.0.0",
     "svelte-eslint-parser": "^0.43.0",
     "tslib": "^2.6.3",
-    "typescript": "^5.4.5",
-    "unocss": "^0.61.0",
-    "vite": "^5.3.1"
+    "typescript": "^5.5.0",
+    "unocss": "^66.0.0",
+    "vite": "^6.0.0"
   }
 }

frontend/uno.config.ts

@@ -1,10 +1,10 @@
 import extractorSvelte from "@unocss/extractor-svelte";
-import { defineConfig, presetAttributify, presetUno, presetWebFonts, transformerDirectives, transformerVariantGroup } from "unocss";
+import { defineConfig, presetAttributify, presetWebFonts, presetWind3, transformerDirectives, transformerVariantGroup } from "unocss";
 
 const config = defineConfig({
   extractors: [extractorSvelte()],
   transformers: [transformerDirectives(), transformerVariantGroup()],
-  presets: [presetAttributify(), presetUno(), presetWebFonts({ provider: "bunny", fonts: { mono: "Fira Code" } })],
+  presets: [presetAttributify(), presetWind3({ preflight: "on-demand" }), presetWebFonts({ provider: "bunny", fonts: { mono: "Fira Code" } })],
 });
 
 export default config;

pyproject.toml

@@ -11,26 +11,26 @@ dependencies = [
     "promplate[all]~=0.3.5.0",
     "promplate-trace[langfuse,langsmith]==0.3.0dev5",
     "python-box~=7.3.0",
-    "pydantic-settings~=2.7.0",
+    "pydantic-settings~=2.8.0",
     "httpx[http2]~=0.28.0",
     "promptools[stream,validation]~=0.1.3.3",
     "fake-useragent~=2.0.0",
     "html2text~=2024.2.26",
-    "beautifulsoup4~=4.12.3",
+    "beautifulsoup4~=4.13.1",
     "rich~=13.9.0",
     "zhipuai~=2.1.5.20241203",
-    "anthropic~=0.45.0",
+    "anthropic~=0.46.0",
     "dashscope~=1.22.0",
-    "logfire[fastapi,system-metrics,httpx]~=3.3.0",
+    "logfire[fastapi,system-metrics,httpx]~=3.6.0",
 ]
 
 [tool.pdm]
 distribution = false
 
 [dependency-groups]
 dev = [
-    "isort~=5.13.2",
-    "black~=24.4.2",
+    "isort~=6.0.0",
+    "black~=25.1.0",
 ]
 
 [tool.pdm.scripts]

src/routes/openai.py

@@ -1,7 +1,10 @@
-from typing import AsyncIterable, cast
+from typing import AsyncIterable, Literal, cast, get_args
 
 from fastapi import APIRouter, Depends, Request
+from openai.types.chat import ChatCompletionContentPartTextParam
 from promplate import Message
+from pydantic import field_serializer
+from typing_extensions import TypedDict
 
 from ..utils.http import forward_headers
 from ..utils.llm import Model, openai_compatible_providers
@@ -13,23 +16,45 @@
 openai_router = APIRouter(tags=["openai"])
 
 
+class ModelItem(TypedDict):
+    id: Model
+    object: Literal["model"]
+
+
+class ModelList(TypedDict):
+    object: Literal["list"]
+    data: list[ModelItem]
+
+
 @openai_router.get("/models")
-async def get_models():
+async def get_models() -> ModelList:
     return {
         "object": "list",
         "data": [
             {
                 "id": name,
                 "object": "model",
             }
-            for name in Model.__args__
+            for name in get_args(Model)
         ],
     }
 
 
+class CompatibleMessage(Message):
+    content: str | list[ChatCompletionContentPartTextParam]  # type: ignore
+
+
 class ChatInput(ChainInput):
     stream: bool = False
-    messages: list[Message]  # type: ignore
+    messages: list[CompatibleMessage]  # type: ignore
+
+    @field_serializer("messages")
+    def serialize_messages(self, value: CompatibleMessage):
+        content = value["content"]
+        if isinstance(content, str):
+            return value
+        value["content"] = "".join(i["text"] for i in content)
+        return value
 
     @property
     def config(self):

src/utils/llm/__init__.py

@@ -39,12 +39,14 @@
     "azure:gpt-4o-mini",
     "azure:o1",
     "azure:o1-mini",
+    "azure:o3-mini",
     "azure:o1-preview",
     "Mistral-Nemo",
     "Mistral-large",
     "Mistral-large-2407",
     "Mistral-large-2411",
     "Mistral-small",
+    "Codestral-2501",
     "Ministral-3B",
     "Meta-Llama-3.1-405B-Instruct",
     "Meta-Llama-3.1-70B-Instruct",
@@ -57,6 +59,7 @@
     "Cohere-command-r-08-2024",
     "AI21-Jamba-1.5-Large",
     "AI21-Jamba-1.5-Mini",
+    "Phi-4",
     "Phi-3.5-MoE-instruct",
     "Phi-3.5-mini-instruct",
     "Phi-3-medium-128k-instruct",
@@ -65,10 +68,15 @@
     "Phi-3-mini-4k-instruct",
     "Phi-3-small-128k-instruct",
     "Phi-3-small-8k-instruct",
+    "DeepSeek-R1",
     "chatglm_turbo",
     "claude-3-haiku-20240307",
     "claude-3-sonnet-20240229",
-    "claude-3-opus-20240229",
+    "qwen-2.5-32b",
+    "qwen-2.5-coder-32b",
+    "deepseek-r1-distill-qwen-32b",
+    "deepseek-r1-distill-llama-70b",
+    "deepseek-r1-distill-llama-70b-specdec",
     "gemma-7b-it",
     "gemma2-9b-it",
     "llama3-8b-8192",
@@ -94,11 +102,6 @@
     "abab5.5-chat",
     "abab6-chat",
     "Qwen/QwQ-32B-Preview",
-    "Qwen/Qwen2-7B-Instruct",
-    "Qwen/Qwen2-1.5B-Instruct",
-    "Qwen/Qwen2-72B-Instruct",
-    "Qwen/Qwen2-57B-A14B-Instruct",
-    "Vendor-A/Qwen/Qwen2-72B-Instruct",
     "Qwen/Qwen2.5-Coder-7B-Instruct",
     "Qwen/Qwen2.5-7B-Instruct",
     "Qwen/Qwen2.5-14B-Instruct",
@@ -114,6 +117,14 @@
     "deepseek-ai/DeepSeek-V2-Chat",
     "deepseek-ai/DeepSeek-V2.5",
     "deepseek-ai/deepseek-llm-67b-chat",
+    "deepseek-ai/DeepSeek-V3",
+    "deepseek-ai/DeepSeek-R1",
+    "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
+    "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B",
+    "deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",
+    "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
+    "deepseek-ai/DeepSeek-R1-Distill-Llama-8B",
+    "deepseek-ai/DeepSeek-R1-Distill-Llama-70B",
     "internlm/internlm2_5-7b-chat",
     "internlm/internlm2_5-20b-chat",
     "yi-lightning",
@@ -122,5 +133,12 @@
     "Qwen2.5-Coder-32B-Instruct",
     "Qwen2.5-72B-Instruct",
     "QwQ-32B-Preview",
+    "Llama-3.1-Tulu-3-405B",
+    "Llama-3.2-11B-Vision-Instruct",
+    "Llama-3.2-90B-Vision-Instruct",
+    "Meta-Llama-3.2-1B-Instruct",
+    "Meta-Llama-3.2-3B-Instruct",
+    "Meta-Llama-3.3-70B-Instruct",
+    "DeepSeek-R1-Distill-Llama-70B",
     "deepseek-chat",
 ]

src/utils/llm/azure.py

@@ -14,13 +14,16 @@
 
 
 @link_llm("azure:o1")
+@link_llm("azure:o3")
 @link_llm("azure:gpt")
 @link_llm("Ministral")
+@link_llm("Codestral")
 @link_llm("Mistral")
 @link_llm("Meta")
 @link_llm("Cohere")
 @link_llm("AI21")
 @link_llm("Phi")
+@link_llm("DeepSeek")
 class AzureOpenAI(AsyncChatOpenAI):
     @staticmethod
     async def generate(prompt, **kwargs):

src/utils/llm/groq.py

@@ -18,6 +18,8 @@
 @link_llm("llama-3.2")
 @link_llm("llama-3.3-70b-")
 @link_llm("mixtral")
+@link_llm("qwen-2.5")
+@link_llm("deepseek-r1-distill")
 class Groq(AsyncChatOpenAI):
     async def complete(self, prompt: str | list[Message], /, **config):
         config = self._run_config | config

src/utils/llm/sambanova.py

@@ -8,6 +8,10 @@
 
 @link_llm("Qwen2.5")
 @link_llm("QwQ")
+@link_llm("Llama-3")
+@link_llm("Meta-Llama-3.2")
+@link_llm("Meta-Llama-3.3")
+@link_llm("DeepSeek-R1-Distill")
 class SambaNova(AsyncChatOpenAI):
     complete = staticmethod(
         patch.chat.acomplete(

src/utils/openai_api.py

@@ -3,12 +3,12 @@
 from typing import AsyncIterable
 
 
-def format_chunk(id, content, model, stop=False):
+def format_chunk(t, id, content, model, stop=False):
     if stop:
         choice = {"index": 0, "delta": {}, "finish_reason": "stop"}
     else:
         choice = {"index": 0, "delta": {"content": content, "role": "assistant"}}
-    return {"id": id, "choices": [choice], "model": model, "object": "chat.completion.chunk"}
+    return {"id": id, "choices": [choice], "model": model, "object": "chat.completion.chunk", "created": t}
 
 
 def format_response(content, model: str):
@@ -21,9 +21,10 @@ def format_response(content, model: str):
 
 
 async def stream_output(stream: AsyncIterable[str], model: str):
-    response_id = f"chatcmpl-{int(time())}"
+    created = int(time())
+    response_id = f"chatcmpl-{created}"
 
     async for delta in stream:
-        yield f"data: {dumps(format_chunk(response_id, delta, model))}\n\n"
+        yield f"data: {dumps(format_chunk(created, response_id, delta, model))}\n\n"
 
-    yield f"data: {dumps(format_chunk(response_id, None, model, stop=True))}\n\ndata: [DONE]\n\n"
+    yield f"data: {dumps(format_chunk(created, response_id, None, model, stop=True))}\n\ndata: [DONE]\n\n"