auto-coder 0.1.281__tar.gz → 0.1.283__tar.gz

{auto_coder-0.1.281 → auto_coder-0.1.283}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.281
+Version: 0.1.283
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -26,7 +26,7 @@ Requires-Dist: tabulate
 Requires-Dist: jupyter_client
 Requires-Dist: prompt-toolkit
 Requires-Dist: tokenizers
-Requires-Dist: byzerllm[saas]>=0.1.170
+Requires-Dist: byzerllm[saas]>=0.1.171
 Requires-Dist: patch
 Requires-Dist: diff_match_patch
 Requires-Dist: GitPython

{auto_coder-0.1.281 → auto_coder-0.1.283}/src/auto_coder.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.281
+Version: 0.1.283
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -26,7 +26,7 @@ Requires-Dist: tabulate
 Requires-Dist: jupyter_client
 Requires-Dist: prompt-toolkit
 Requires-Dist: tokenizers
-Requires-Dist: byzerllm[saas]>=0.1.170
+Requires-Dist: byzerllm[saas]>=0.1.171
 Requires-Dist: patch
 Requires-Dist: diff_match_patch
 Requires-Dist: GitPython

{auto_coder-0.1.281 → auto_coder-0.1.283}/src/auto_coder.egg-info/SOURCES.txt

@@ -138,6 +138,7 @@ src/autocoder/rag/cache/__init__.py
 src/autocoder/rag/cache/base_cache.py
 src/autocoder/rag/cache/byzer_storage_cache.py
 src/autocoder/rag/cache/file_monitor_cache.py
+src/autocoder/rag/cache/local_byzer_storage_cache.py
 src/autocoder/rag/cache/simple_cache.py
 src/autocoder/rag/loaders/__init__.py
 src/autocoder/rag/loaders/docx_loader.py

{auto_coder-0.1.281 → auto_coder-0.1.283}/src/auto_coder.egg-info/requires.txt

@@ -16,7 +16,7 @@ tabulate
 jupyter_client
 prompt-toolkit
 tokenizers
-byzerllm[saas]>=0.1.170
+byzerllm[saas]>=0.1.171
 patch
 diff_match_patch
 GitPython

{auto_coder-0.1.281 → auto_coder-0.1.283}/src/autocoder/auto_coder_rag.py

@@ -22,7 +22,7 @@ from loguru import logger
 import asyncio
 from datetime import datetime
 
-from autocoder.rag.document_retriever import process_file_local
+from autocoder.rag.utils import process_file_local
 import pkg_resources
 from autocoder.rag.token_counter import TokenCounter
 from autocoder.rag.types import RAGServiceInfo
@@ -186,6 +186,11 @@ def main(input_args: Optional[List[str]] = None):
     build_index_parser.add_argument(
         "--model", default="v3_chat", help=desc["model"]
     )
+    
+    build_index_parser.add_argument(
+        "--on_ray", action="store_true", help="Run on Ray"
+    )
+
     build_index_parser.add_argument(
         "--index_model", default="", help=desc["index_model"]
     )
@@ -216,8 +221,7 @@ def main(input_args: Optional[List[str]] = None):
     )
     serve_parser.add_argument("--file", default="", help=desc["file"])
     serve_parser.add_argument("--model", default="v3_chat", help=desc["model"])
-    serve_parser.add_argument("--index_model", default="", help=desc["index_model"])
-    serve_parser.add_argument("--emb_model", default="", help=desc["emb_model"])
+    serve_parser.add_argument("--index_model", default="", help=desc["index_model"])    
     serve_parser.add_argument("--ray_address", default="auto", help=desc["ray_address"])
     serve_parser.add_argument(
         "--index_filter_workers",
@@ -377,6 +381,12 @@ def main(input_args: Optional[List[str]] = None):
         help="The model used for question answering",
     )
 
+    serve_parser.add_argument(
+        "--emb_model",
+        default="",
+        help="The model used for embedding documents",
+    )
+
     # Benchmark command
     benchmark_parser = subparsers.add_parser(
         "benchmark", help="Benchmark LLM client performance"
@@ -510,10 +520,7 @@ def main(input_args: Optional[List[str]] = None):
         # Generate unique name for RAG build if doc_dir exists
         if server_args.doc_dir:
             auto_coder_args.rag_build_name = generate_unique_name_from_path(server_args.doc_dir)
-            logger.info(f"Generated RAG build name: {auto_coder_args.rag_build_name}")
-
-        if auto_coder_args.enable_hybrid_index and args.product_mode == "lite":
-            raise Exception("Hybrid index is not supported in lite mode")
+            logger.info(f"Generated RAG build name: {auto_coder_args.rag_build_name}")        
 
         if auto_coder_args.enable_hybrid_index and args.product_mode == "pro":
             # 尝试连接storage
@@ -555,16 +562,22 @@ def main(input_args: Optional[List[str]] = None):
                 qa_model.skip_nontext_check = True
                 llm.setup_sub_client("qa_model", qa_model)
 
+            if args.emb_model:
+                emb_model = byzerllm.ByzerLLM()
+                emb_model.setup_default_model_name(args.emb_model)
+                emb_model.skip_nontext_check = True
+                llm.setup_sub_client("emb_model", emb_model)
+
             # 当启用hybrid_index时,检查必要的组件
             if auto_coder_args.enable_hybrid_index:
-                if not llm.is_model_exist("emb"):
+                if not args.emb_model and not llm.is_model_exist("emb"):
                     logger.error(
                         "When enable_hybrid_index is true, an 'emb' model must be deployed"
                     )
                     return
-                llm.setup_default_emb_model_name("emb")
+                llm.setup_default_emb_model_name(args.emb_model or "emb")
 
-        elif args.product_mode == "lite":
+        if args.product_mode == "lite":
             from autocoder import models as models_module
             model_info = models_module.get_model_by_name(args.model)
             llm = byzerllm.SimpleByzerLLM(default_model_name=args.model)
@@ -629,6 +642,26 @@ def main(input_args: Optional[List[str]] = None):
                 )
                 llm.setup_sub_client("qa_model", qa_model)
 
+            if args.emb_model:
+                model_info = models_module.get_model_by_name(args.emb_model)
+                emb_model = byzerllm.SimpleByzerLLM(default_model_name=args.emb_model)
+                emb_model.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.emb_model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": False
+                    }
+                )
+                llm.setup_sub_client("emb_model", emb_model)
+
+            if args.enable_hybrid_index:
+                if not args.emb_model:
+                    raise Exception("When enable_hybrid_index is true, an 'emb' model must be specified")                
+
         if server_args.doc_dir:
             auto_coder_args.rag_type = "simple"
             auto_coder_args.rag_build_name = generate_unique_name_from_path(server_args.doc_dir)
@@ -675,31 +708,62 @@ def main(input_args: Optional[List[str]] = None):
         auto_coder_args.enable_hybrid_index = True
         auto_coder_args.rag_type = "simple"
 
-        try:
-            from byzerllm.apps.byzer_storage.simple_api import ByzerStorage
-
-            storage = ByzerStorage("byzerai_store", "rag", "files")
-            storage.retrieval.cluster_info("byzerai_store")
-        except Exception as e:
-            logger.error(
-                "When enable_hybrid_index is true, ByzerStorage must be started"
-            )
-            logger.error("Please run 'byzerllm storage start' first")
-            return
+        if args.on_ray:
 
-        llm = byzerllm.ByzerLLM()
-        llm.setup_default_model_name(args.model)
+            try:
+                from byzerllm.apps.byzer_storage.simple_api import ByzerStorage
 
-        # 当启用hybrid_index时,检查必要的组件
-        if auto_coder_args.enable_hybrid_index:
-            if not llm.is_model_exist("emb"):
+                storage = ByzerStorage("byzerai_store", "rag", "files")
+                storage.retrieval.cluster_info("byzerai_store")
+            except Exception as e:
                 logger.error(
-                    "When enable_hybrid_index is true, an 'emb' model must be deployed"
+                    "When enable_hybrid_index is true, ByzerStorage must be started"
                 )
+                logger.error("Please run 'byzerllm storage start' first")
                 return
-            llm.setup_default_emb_model_name("emb")
 
-        auto_coder_args.rag_build_name = generate_unique_name_from_path(args.doc_dir)
+            llm = byzerllm.ByzerLLM()
+            llm.setup_default_model_name(args.model)
+
+            # 当启用hybrid_index时,检查必要的组件
+            if auto_coder_args.enable_hybrid_index:
+                if not llm.is_model_exist("emb"):
+                    logger.error(
+                        "When enable_hybrid_index is true, an 'emb' model must be deployed"
+                    )
+                    return
+                llm.setup_default_emb_model_name("emb")
+        else:
+            from autocoder import models as models_module
+            model_info = models_module.get_model_by_name(args.model)
+            llm = byzerllm.SimpleByzerLLM(default_model_name=args.model)
+            llm.deploy(
+                model_path="",
+                pretrained_model_type=model_info["model_type"],
+                udf_name=args.model,
+                infer_params={
+                    "saas.base_url": model_info["base_url"],
+                    "saas.api_key": model_info["api_key"],
+                    "saas.model": model_info["model_name"],
+                    "saas.is_reasoning": model_info["is_reasoning"]
+                }
+            )
+
+            model_info = models_module.get_model_by_name(args.emb_model)
+            emb_model = byzerllm.SimpleByzerLLM(default_model_name=args.emb_model)
+            emb_model.deploy(
+                model_path="",
+                pretrained_model_type=model_info["model_type"],
+                udf_name=args.emb_model,
+                infer_params={
+                    "saas.base_url": model_info["base_url"],
+                    "saas.api_key": model_info["api_key"],
+                    "saas.model": model_info["model_name"],
+                    "saas.is_reasoning": False
+                }
+            )
+            llm.setup_sub_client("emb_model", emb_model)
+        
         rag = RAGFactory.get_rag(
             llm=llm,
             args=auto_coder_args,