Revert removing the unused imports (#385)

python/sglang/api.py

@@ -3,7 +3,11 @@
 import re
 from typing import Callable, List, Optional, Union
 
+from sglang.backend.anthropic import Anthropic
 from sglang.backend.base_backend import BaseBackend
+from sglang.backend.openai import OpenAI
+from sglang.backend.runtime_endpoint import RuntimeEndpoint
+from sglang.backend.vertexai import VertexAI
 from sglang.global_config import global_config
 from sglang.lang.ir import (
     SglExpr,

python/sglang/backend/anthropic.py

+from typing import List, Optional, Union
+
+import numpy as np
+
 from sglang.backend.base_backend import BaseBackend
 from sglang.lang.chat_template import get_chat_template
 from sglang.lang.interpreter import StreamExecutor

python/sglang/backend/base_backend.py

-from typing import List, Optional, Union
+from typing import Callable, List, Optional, Union
 
 from sglang.lang.chat_template import get_chat_template
 from sglang.lang.interpreter import StreamExecutor

python/sglang/backend/openai.py

 import logging
 import time
-from typing import List, Optional
+from typing import Callable, List, Optional, Union
 
 import numpy as np
 

python/sglang/backend/runtime_endpoint.py

 import json
-from typing import List, Optional
+from typing import Callable, List, Optional, Union
 
 import numpy as np
+import requests
 
 from sglang.backend.base_backend import BaseBackend
 from sglang.global_config import global_config
 from sglang.lang.chat_template import get_chat_template_by_model_path
 from sglang.lang.interpreter import StreamExecutor
-from sglang.lang.ir import SglSamplingParams
-from sglang.utils import find_printable_text, http_request
+from sglang.lang.ir import SglArgument, SglSamplingParams
+from sglang.utils import encode_image_base64, find_printable_text, http_request
 
 
 class RuntimeEndpoint(BaseBackend):

python/sglang/backend/vertexai.py

 import os
 import warnings
+from typing import List, Optional, Union
+
+import numpy as np
 
 from sglang.backend.base_backend import BaseBackend
 from sglang.lang.chat_template import get_chat_template

python/sglang/lang/chat_template.py

-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from enum import Enum, auto
-from typing import Callable, Dict, List, Tuple
+from typing import Callable, Dict, List, Optional, Tuple
 
 
 class ChatTemplateStyle(Enum):

python/sglang/lang/compiler.py

@@ -5,7 +5,13 @@ from typing import List, Union
 
 from sglang.global_config import global_config
 from sglang.lang.interpreter import ProgramState, StreamExecutor, pin_program
-from sglang.lang.ir import SglArgument, SglExpr, SglSamplingParams, SglVariable
+from sglang.lang.ir import (
+    SglArgument,
+    SglConstantText,
+    SglExpr,
+    SglSamplingParams,
+    SglVariable,
+)
 
 
 def compile_func(function, backend):

python/sglang/lang/interpreter.py

@@ -7,7 +7,7 @@ import threading
 import uuid
 from concurrent.futures import ThreadPoolExecutor
 from contextlib import contextmanager
-from typing import Any, Callable, Dict, List, Optional
+from typing import Any, Callable, Dict, List, Optional, Union
 
 import tqdm
 
@@ -18,6 +18,7 @@ from sglang.lang.ir import (
     SglConstantText,
     SglExpr,
     SglExprList,
+    SglFunction,
     SglGen,
     SglImage,
     SglRoleBegin,

python/sglang/lang/ir.py

@@ -472,4 +472,4 @@ class SglCommitLazy(SglExpr):
         super().__init__()
 
     def __repr__(self):
-        return "CommitLazy()"
+        return f"CommitLazy()"

python/sglang/lang/tracer.py

 """Tracing a program."""
 
 import uuid
-from typing import Any, Dict, List, Optional
+from typing import Any, Callable, Dict, List, Optional, Union
 
 from sglang.backend.base_backend import BaseBackend
+from sglang.global_config import global_config
 from sglang.lang.interpreter import ProgramState, ProgramStateGroup
 from sglang.lang.ir import (
     SglArgument,
+    SglCommitLazy,
+    SglConcateAndAppend,
     SglConstantText,
     SglExpr,
     SglExprList,
     SglFork,
+    SglFunction,
     SglGen,
     SglGetForkItem,
     SglRoleBegin,

python/sglang/srt/hf_transformers_utils.py

@@ -3,7 +3,7 @@
 import json
 import os
 import warnings
-from typing import Optional, Union
+from typing import List, Optional, Tuple, Union
 
 from huggingface_hub import snapshot_download
 from transformers import (

python/sglang/srt/managers/detokenizer_manager.py

@@ -84,7 +84,7 @@ def start_detokenizer_process(
 ):
     try:
         manager = DetokenizerManager(server_args, port_args)
-    except Exception:
+    except Exception as e:
         pipe_writer.send(get_exception_traceback())
         raise
     pipe_writer.send("init ok")

python/sglang/srt/managers/router/radix_cache.py

 import heapq
 import time
 from collections import defaultdict
+from dataclasses import dataclass
+from typing import Tuple
 
 import torch
 

python/sglang/srt/models/commandr.py

@@ -20,7 +20,7 @@
 
 # This file is based on the LLama model definition file in transformers
 """PyTorch Cohere model."""
-from typing import Optional, Tuple
+from typing import List, Optional, Tuple
 
 import torch
 import torch.utils.checkpoint

python/sglang/srt/models/llama2.py

 # Adapted from
 # https://github.com/vllm-project/vllm/blob/671af2b1c0b3ed6d856d37c21a561cc429a10701/vllm/model_executor/models/llama.py#L1
 """Inference-only LLaMA model compatible with HuggingFace weights."""
-from typing import Any, Dict, Optional, Tuple
+from typing import Any, Dict, List, Optional, Tuple
 
 import torch
 from torch import nn

python/sglang/srt/models/llava.py

@@ -5,7 +5,7 @@ from typing import List, Optional
 import numpy as np
 import torch
 from torch import nn
-from transformers import CLIPVisionModel, LlavaConfig
+from transformers import CLIPVisionModel, LlamaConfig, LlavaConfig
 from transformers.models.llava.modeling_llava import LlavaMultiModalProjector
 from vllm.model_executor.layers.linear import LinearMethodBase
 from vllm.model_executor.weight_utils import (

python/sglang/srt/models/mixtral.py

 # Adapted from
 # https://github.com/vllm-project/vllm/blob/d0215a58e78572d91dadafe9d832a2db89b09a13/vllm/model_executor/models/mixtral.py#L1
 """Inference-only Mixtral model."""
-from typing import Optional
+from typing import List, Optional, Tuple
 
 import numpy as np
 import torch

python/sglang/srt/models/qwen.py

-from typing import Any, Dict, Optional
+from typing import Any, Dict, List, Optional, Tuple
 
 import torch
 from torch import nn

python/sglang/srt/models/qwen2.py

 # Adapted from llama2.py
 # Modify details for the adaptation of Qwen2 model.
 """Inference-only Qwen2 model compatible with HuggingFace weights."""
-from typing import Any, Dict, Optional, Tuple
+from typing import Any, Dict, List, Optional, Tuple
 
 import torch
 from torch import nn