fix issues with optional dependencies (#204)

* fix issues with optional dependencies

* format document

* Tree: Format and comment

backends/exllamav2/utils.py

@@ -5,34 +5,6 @@ from importlib.metadata import PackageNotFoundError, version as package_version
 from loguru import logger
 
 
-def check_exllama_version():
-    """Verifies the exllama version"""
-
-    required_version = version.parse("0.2.2")
-    current_version = version.parse(package_version("exllamav2").split("+")[0])
-
-    unsupported_message = (
-        f"ERROR: TabbyAPI requires ExLlamaV2 {required_version} "
-        f"or greater. Your current version is {current_version}.\n"
-        "Please update your environment by running an update script "
-        "(update_scripts/"
-        f"update_deps.{'bat' if platform.system() == 'Windows' else 'sh'})\n\n"
-        "Or you can manually run a requirements update "
-        "using the following command:\n\n"
-        "For CUDA 12.1:\n"
-        "pip install --upgrade .[cu121]\n\n"
-        "For CUDA 11.8:\n"
-        "pip install --upgrade .[cu118]\n\n"
-        "For ROCm:\n"
-        "pip install --upgrade .[amd]\n\n"
-    )
-
-    if current_version < required_version:
-        raise SystemExit(unsupported_message)
-    else:
-        logger.info(f"ExllamaV2 version: {current_version}")
-
-
 def hardware_supports_flash_attn(gpu_device_list: list[int]):
     """
     Check whether all GPUs in list support FA2

backends/exllamav2/version.py

@@ -0,0 +1,39 @@
+import platform
+from packaging import version
+from importlib.metadata import version as package_version
+from loguru import logger
+from common.optional_dependencies import dependencies
+
+
+def check_exllama_version():
+    """Verifies the exllama version"""
+
+    install_message = (
+        "Please update your environment by running an update script "
+        "(update_scripts/"
+        f"update_deps.{'bat' if platform.system() == 'Windows' else 'sh'})\n\n"
+        "Or you can manually run a requirements update "
+        "using the following command:\n\n"
+        "For CUDA 12.1:\n"
+        "pip install --upgrade .[cu121]\n\n"
+        "For CUDA 11.8:\n"
+        "pip install --upgrade .[cu118]\n\n"
+        "For ROCm:\n"
+        "pip install --upgrade .[amd]\n\n"
+    )
+
+    if not dependencies.exl2:
+        raise SystemExit(("Exllamav2 is not installed.\n" + install_message))
+
+    required_version = version.parse("0.2.2")
+    current_version = version.parse(package_version("exllamav2").split("+")[0])
+
+    unsupported_message = (
+        f"ERROR: TabbyAPI requires ExLlamaV2 {required_version} "
+        f"or greater. Your current version is {current_version}.\n" + install_message
+    )
+
+    if current_version < required_version:
+        raise SystemExit(unsupported_message)
+    else:
+        logger.info(f"ExllamaV2 version: {current_version}")

backends/infinity/model.py

@@ -5,16 +5,12 @@ from loguru import logger
 from typing import List, Optional
 
 from common.utils import unwrap
+from common.optional_dependencies import dependencies
 
 # Conditionally import infinity to sidestep its logger
-has_infinity_emb: bool = False
-try:
+if dependencies.extras:
     from infinity_emb import EngineArgs, AsyncEmbeddingEngine
 
-    has_infinity_emb = True
-except ImportError:
-    pass
-
 
 class InfinityContainer:
     model_dir: pathlib.Path
@@ -23,7 +19,7 @@ class InfinityContainer:
 
     # Conditionally set the type hint based on importablity
     # TODO: Clean this up
-    if has_infinity_emb:
+    if dependencies.extras:
         engine: Optional[AsyncEmbeddingEngine] = None
     else:
         engine = None