flowllm 0.1.2__py3-none-any.whl → 0.1.5__py3-none-any.whl

{flowllm-0.1.2.dist-info → flowllm-0.1.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: flowllm
-Version: 0.1.2
+Version: 0.1.5
 Summary: A flexible framework for building LLM-powered flows and mcp services
 Author-email: FlowLLM Team <flowllm@example.com>
 Maintainer-email: FlowLLM Team <flowllm@example.com>
@@ -219,7 +219,7 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
 Classifier: Typing :: Typed
-Requires-Python: >=3.12
+Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: akshare
@@ -246,9 +246,10 @@ Requires-Dist: urllib3
 Requires-Dist: uvicorn[standard]
 Requires-Dist: chromadb
 Requires-Dist: elasticsearch
-Requires-Dist: ray
-Provides-Extra: distributed
-Requires-Dist: ray; extra == "distributed"
+Requires-Dist: pyfiglet
+Requires-Dist: rich
+Provides-Extra: dist
+Requires-Dist: ray; extra == "dist"
 Provides-Extra: all
-Requires-Dist: flowllm[distributed]; extra == "all"
+Requires-Dist: flowllm[dist]; extra == "all"
 Dynamic: license-file

flowllm-0.1.5.dist-info/RECORD

@@ -0,0 +1,98 @@
+flowllm/__init__.py,sha256=pZ88ULKiYoL0K634bdCPZwXzM4qWtaLp0w6CH9QqxCo,489
+flowllm/app.py,sha256=dLeIbsXRztVNG-Upd2F61xSuOJsh8jy8TDhTv7jROv4,252
+flowllm/client/__init__.py,sha256=ruBU6hBC7LWoVg5oNdZBTqi3MhNUhkCYAHhPj-UJKEA,815
+flowllm/client/async_http_client.py,sha256=w3YUsCTlART3HvAIpf30ZIX4m7L7k9f9iZejkNBq_h8,3026
+flowllm/client/http_client.py,sha256=rk6-hF3lqwn8a2igQjBuRz7JOV9lJJrNYlC7Xk18Zwo,2930
+flowllm/client/mcp_client.py,sha256=dFJvHqqVpUv7icIUNKdJpfQI1FXx3zF76Lgnbyyp8Go,4867
+flowllm/client/sync_mcp_client.py,sha256=Csv5hwlPDDCktM9qqLqdQkSG039ooiSurN5DMcJfrq8,4305
+flowllm/config/__init__.py,sha256=fUjhShdU7qN6hfj7blDt7nHBQGlWwbhaa_p9_6vDpg8,57
+flowllm/config/base.yaml,sha256=Zc9jkE-2lrDMg2qS-Cu-Nj1vtxlp5SHVHmWMD2ZuuxM,1379
+flowllm/config/fin_supply.yaml,sha256=QRxw-9COoZFzTwV4bx8Y0h7cUOae6vjmaT7OSeTyTzU,1318
+flowllm/config/pydantic_config_parser.py,sha256=cRfhw5fk5p1ZKx9ijMsMkmmvEu-M15xqFj8pnvTFtvs,7584
+flowllm/context/__init__.py,sha256=dM7ttmA_QuLddTz1JU8QgHtnaL9K3SgQ8OwTkqXEj6Q,69
+flowllm/context/base_context.py,sha256=jm4JBCl8IPUboP4hPnO1llFtJyU3jVg-rtGiHUNXXyA,1742
+flowllm/context/flow_context.py,sha256=NGfB6oFVx0MNClHjhA_mrv66wHee5jvUjz_eSfEYRfo,2244
+flowllm/context/prompt_handler.py,sha256=LbMJDwrzz8HF7HVDyphVSWI1X4QHMebJqLP4XeXkXOE,2667
+flowllm/context/registry.py,sha256=zJf8SYi8ziFTdf2EG9Ut2_3sxEffchrl81W-VbcoEns,1110
+flowllm/context/service_context.py,sha256=VsF5Dsh_L4vdEZKcDLqNV1JvhfivrI_7Env0YCOBoC8,8043
+flowllm/embedding_model/__init__.py,sha256=jHxnI2xySqYcu7Y2wxmlmpIyYeKH_4PuNCBNvouXhEE,78
+flowllm/embedding_model/base_embedding_model.py,sha256=oIySUu6_ROEXw13jtKYUtnL0bgBQdHtNZqnPPxatG9I,8021
+flowllm/embedding_model/openai_compatible_embedding_model.py,sha256=tzjGkHfV9qOlkWZfYyg8SKHdj3GPs1cuYr1f7_SIEZE,5481
+flowllm/enumeration/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flowllm/enumeration/chunk_enum.py,sha256=pGh5ppp9UHCuQYxiNfPcsh9X7tI2Dt6D8HMSLiYi2Io,170
+flowllm/enumeration/http_enum.py,sha256=ddIYeVz2J8CInzEkhWbz1rPe0qh2FVfryNNE1uc1ioM,141
+flowllm/enumeration/role.py,sha256=SFMGXmwYFndGIwbr9GxS3OIoHIZ_nO23DeRnqpV4U3g,133
+flowllm/flow/__init__.py,sha256=Nnwgz85jYpTQWYSTOpvY8tppuovpzFHGtROefiBUeDY,264
+flowllm/flow/base_flow.py,sha256=xIR5nVdxveW_57jz2ZpWr6nQcUgp59uizPlI_salInY,3606
+flowllm/flow/base_tool_flow.py,sha256=eC-V9kafecsXJixqXGfmD1DF70tA8ClM2f4rnW2K1vU,368
+flowllm/flow/tool_op_flow.py,sha256=Uhlgezk8LP0y0tOIdPnHTCbjr_SN0nCmaonXzVTQ1Ww,2996
+flowllm/flow/expression/__init__.py,sha256=WTCjGmimDp86DBmOgOW5CcPd1-iKHwlD2GUbfkG-74w,53
+flowllm/flow/expression/expression_parser.py,sha256=ioYcC3VEpWwUkQZmDvGQCh1SqUp9SBloRxh7rqAuQP0,6109
+flowllm/flow/expression/expression_tool_flow.py,sha256=-5oKEfgCmJ8lMwdWaT79HbF4M9q74lBU4RdS7JbXtD0,990
+flowllm/flow/gallery/__init__.py,sha256=debywmZxht5JlZYn15ZQ40t-z2FWuDe-RtALMNiR-Us,60
+flowllm/flow/gallery/mock_tool_flow.py,sha256=0-lV7bGw0ChCYXCoMZ-CCyl1DWpRu31L7pIxBMvQocw,2759
+flowllm/llm/__init__.py,sha256=c2akU1k4IVT6PoW3FdKlFmQNZj7MVCzV_VTUw5VpDrc,99
+flowllm/llm/base_llm.py,sha256=80YvTLZQtEwCwtwgkb-r56CzmxMlO5BAiH_wZ9rl-zg,9439
+flowllm/llm/litellm_llm.py,sha256=fZpbThfRRyOBPASdZfGHUr_twwzgombbEEhWDQtuzus,19422
+flowllm/llm/openai_compatible_llm.py,sha256=MKbe4-p4NkCBCCkrty0mFsBTRpXctI-7rz0-2kzvNQo,18534
+flowllm/op/__init__.py,sha256=quaB1nSOt-dz7WMryBhmcJDIcbv4AqxvgAmoKvjJucg,210
+flowllm/op/base_llm_op.py,sha256=mzRKgROambE1xI4JSwNrCLdkN6y-PemCGzZERqVpJr8,2682
+flowllm/op/base_op.py,sha256=Gg5k64YftGI3UP5jc7gn44QQ2x83E--yh0mDml3JMUo,12059
+flowllm/op/base_tool_op.py,sha256=5ULOE_-23RnYy6-PWLfxrAHpP5d2KoG3qPiIgWC9pP4,1482
+flowllm/op/parallel_op.py,sha256=wea_zrzZXG22du9U8S5Ka4Weo7AxMR1AiRkmZER_yzQ,730
+flowllm/op/sequential_op.py,sha256=vgsWcsQm6VoLFRgzFvy6SSz-4nBHMAnQFkj4_bqpBuY,587
+flowllm/op/akshare/__init__.py,sha256=PaCPzBv-csRgmtmBjoYqxvOgAB3Vp8EeSGP3rGw6rc8,178
+flowllm/op/akshare/get_ak_a_code_op.py,sha256=_WQWqGZV9c2pWO9SqDNTcVtMeVUJwRkXJoFAPPtxy7U,3847
+flowllm/op/akshare/get_ak_a_code_prompt.yaml,sha256=NwGJygecUwspQvrg98GdMpcmQESpZ2Kf5YprhM2PyZU,608
+flowllm/op/akshare/get_ak_a_info_op.py,sha256=yaFjy6sZuVJvELWHnv0EwyAeKUp6Y1ThXoUn3adicUg,4489
+flowllm/op/gallery/__init__.py,sha256=0FDEv_jNUsg1E9GDMTb6jjUdYCEVKjHPjTGR3CwMHF8,74
+flowllm/op/gallery/mock_op.py,sha256=uDj0CJS9Yy0g3ICCJHItDzUc7Q0dY_ZuUm5RmHloHFA,950
+flowllm/op/llm/__init__.py,sha256=Rb-LfmQwMsQMVn8wa1BOythYMvL4F4-2ypEmVeshFL0,115
+flowllm/op/llm/react_llm_op.py,sha256=Ckdp7P6FEyi3blECwupjzTgoStx3FDl32lRAMjhYQW4,4123
+flowllm/op/llm/react_llm_prompt.yaml,sha256=UnBvR_2ph7AG-HNDkn-SJvZiym-UcbcizKdc9w9NImg,2344
+flowllm/op/llm/simple_llm_op.py,sha256=YQrZwCQfAG9F6b9Onr90mM8wR-h5Tpj0E1__g3q9RJo,1373
+flowllm/op/llm/stream_llm_op.py,sha256=QCWb0KoRfFATa0IgUI8JEpOnioqn-IVALcLFk8cEt-g,1766
+flowllm/op/mcp/__init__.py,sha256=UkJR7eKHsUGwaD8_tbx0hrsxBTbPX63VSwFI4gHytl8,91
+flowllm/op/mcp/ant_op.py,sha256=atUhYFwbd7IH_7qgJV2AZJSzrFwRXbFw_JWDkzhZpOs,1253
+flowllm/op/mcp/base_sse_mcp_op.py,sha256=MwHI5dJjntCFdXt8XvszW8eX1m-pbSeaYtwEg2s5obg,1087
+flowllm/op/search/__init__.py,sha256=rmfkzcizghmRQmfsX0UD1eoLZLC2xRDjjdkmbdkaMXw,95
+flowllm/op/search/dashscope_search_op.py,sha256=0AAdhGqvbHP0ibeLPvQFqzlV9nI6eg5G-zyJpPvZZEA,4694
+flowllm/op/search/dashscope_search_prompt.yaml,sha256=NPsS3QCo8xlRpWw7fyMpuYBeOXSAtXscwGO-cVgUNFw,355
+flowllm/op/search/tavily_search_op.py,sha256=iJNQuG7PQNFLUoW_ksPbdN_WDia19cMcd-I_zcgOkjM,3287
+flowllm/schema/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flowllm/schema/flow_request.py,sha256=jaOgEJ7XRWNgS_QTCBOOsa6hKtdgqnXVC7S4Vi6lMBg,338
+flowllm/schema/flow_response.py,sha256=BusrfkJ8OCGk9Tp1KCWVSt33L2t_pYpn9QRqS0bjDcw,323
+flowllm/schema/flow_stream_chunk.py,sha256=6NZDfbTt-zdh1Jh4x5sos2EnzZhBYW2J2fBEcgo9m4I,355
+flowllm/schema/message.py,sha256=MvTKhUX5yUY9mcP9vEFRZ4RJMObhUPQ3iQjOC_OpTvU,1345
+flowllm/schema/service_config.py,sha256=YOmOoTaljsaAnfr9JULmDQ92JA6U0MIpD7dkF0Fs3T0,2614
+flowllm/schema/tool_call.py,sha256=pmglMr-iE36if2Zc-mX8oU4AepgBqbBfHHdrvK1IoG8,5460
+flowllm/schema/vector_node.py,sha256=Urs9EyzzjuQVYPCB_Pfp0iK2npWWCJSXdT6uDLdT05w,362
+flowllm/service/__init__.py,sha256=wUV5Q_cCP1PaBJy1EZYz6YqfUOE8qWM7ZqMGKY3d1sg,74
+flowllm/service/base_service.py,sha256=JW29_guk47-oISz-ROGShO7o4VVYY8iFxOqlDQFd4h4,2969
+flowllm/service/http_service.py,sha256=QQDQKtYrCfJgpr_zxP4HowgFHa4poqgcWhxl_dbF2Ns,3430
+flowllm/service/mcp_service.py,sha256=ss0NWP_UDPqkPPzsVK-kfvBZf7wNuzDdZ8VvQIecTVA,1662
+flowllm/storage/__init__.py,sha256=wlGDDqp6nM4TLoEAXtSglgB0-K2YHAq_DNnsHTBQj8s,41
+flowllm/storage/cache/__init__.py,sha256=KyLfUg3gJ4SvHMRagtZDL8cp5s5SRsN9H-_y4uRf2II,34
+flowllm/storage/cache/cache_data_handler.py,sha256=fNnVwKOTOL3fZfukIN2gy3txlxiwr3CR9t3tVnIusvw,2700
+flowllm/storage/cache/data_cache.py,sha256=cVz20JUkGgXzB5YgZiyeiBYpA0nRFyBtxhgVVQkFeA4,12144
+flowllm/storage/vector_store/__init__.py,sha256=ZRv-O-uhpb8hUOmsxKk5VqlSY15d27VqvJP3pqUgUWw,194
+flowllm/storage/vector_store/base_vector_store.py,sha256=ISntxjjXGv4gnOSzZ7mF3sLvhokJsd7ATcDcpD295a4,6321
+flowllm/storage/vector_store/chroma_vector_store.py,sha256=gRsgzblSI8-_Mnq6Kq1PUrOxTk1XxqDNwRcTwBbuN_U,16641
+flowllm/storage/vector_store/es_vector_store.py,sha256=FTeXRGdSDosnmDEpNHTfs4kY0Az22mWBql037BWkYDc,18664
+flowllm/storage/vector_store/local_vector_store.py,sha256=wsngpKlbfosMTDhiJlGSv5MP47y1XSU3nJJPWefxx1E,17835
+flowllm/storage/vector_store/memory_vector_store.py,sha256=ALyU_q_ufrObCZDT4F3s4PLFfXQ7QYnWuRFYvXYLY4M,21081
+flowllm/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flowllm/utils/common_utils.py,sha256=-3WxJV7SgfI5vtw9Uj2BupUBzdB5vwsqCysErF-QC-E,3209
+flowllm/utils/fetch_url.py,sha256=pOYiliL4kTGQKERCHrFhammBsbv9XzSvVyBQDMgXKEY,4404
+flowllm/utils/llm_utils.py,sha256=ywhf1LGe2aKEaL5PyOpoOAbiekrQVOcpxUJ4ARPioQQ,1121
+flowllm/utils/logger_utils.py,sha256=cBgBKO_oQlN0O19AR0tMxmMo1TTaDdjPI3_g2Xy8Csw,703
+flowllm/utils/miner_u_pdf_processor.py,sha256=wYcKmNxFNwOMmDyO790G3uqBsFtZgHy65VZE2pM9A10,28956
+flowllm/utils/ridge_v2.py,sha256=XIn6nu4jUV7_QUCeyhSEhm-4ltueaS7JdbDQmSQFnRE,1802
+flowllm/utils/singleton.py,sha256=No3otyPDRHu6wQuFRC-w28MkbommVFTLd7H4mT6-Zos,213
+flowllm/utils/timer.py,sha256=8aj3dIYOyxNDNdlcitezdepxEptqkx69aw6JNFWsr30,1492
+flowllm-0.1.5.dist-info/licenses/LICENSE,sha256=kFfPsL7YvEW4jPATpyvUPdtEgftLK53zQrVYJ0eBASY,11337
+flowllm-0.1.5.dist-info/METADATA,sha256=omJ1490ENhn4C8yavYbTvHVC7tMN530pmRaxY9Otyu0,14760
+flowllm-0.1.5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+flowllm-0.1.5.dist-info/entry_points.txt,sha256=PcKC61HCKtF4ONb4HnrIY6J2JeV1bsi_l0O2m99A2Wg,45
+flowllm-0.1.5.dist-info/top_level.txt,sha256=a2tZBwbrnw4uiydMI5kAEGz-cxG25rf6v0QM1sYIrjs,8
+flowllm-0.1.5.dist-info/RECORD,,

flowllm/config/default.yaml

@@ -1,77 +0,0 @@
-# default config.yaml
-backend: mcp
-language: ""
-thread_pool_max_workers: 32
-ray_max_workers: 1
-
-mcp:
-  transport: sse
-  host: "0.0.0.0"
-  port: 8001
-
-http:
-  host: "0.0.0.0"
-  port: 8001
-  timeout_keep_alive: 600
-  limit_concurrency: 64
-
-flow:
-  get_a_stock_infos:
-    flow_content: get_ak_a_code_op >> get_ak_a_info_op >> get_ak_a_spot_op >> get_ak_a_money_flow_op >> get_ak_a_financial_info_op >> merge_ak_a_info_op
-    description: "Retrieve the A-share stock codes from the query, and fetch information about these stocks, including company basic information, current stock price and its change percentage, capital inflow and outflow data for the most recent day, and financial information from the latest quarter."
-    input_schema:
-      query:
-        type: "str"
-        description: "user question"
-
-  get_a_stock_news:
-    flow_content: get_ak_a_code_op >> get_ak_a_news_op >> merge_ak_a_info_op
-    description: "Retrieve the A-share stock codes from the query, and obtain the latest news information about these stocks."
-    input_schema:
-      query:
-        type: "str"
-        description: "user question"
-
-  mock_expression_flow:
-    flow_content: mock1_op>>((mock4_op>>mock2_op)|mock5_op)>>(mock3_op|mock6_op)
-    description: "mock flow"
-    input_schema:
-      a:
-        type: "str"
-        description: "mock attr a"
-        required: true
-      b:
-        type: "str"
-        description: "mock attr b"
-        required: true
-
-op:
-  mock1_op:
-    backend: mock1_op
-    llm: default
-    vector_store: default
-
-llm:
-  default:
-    backend: openai_compatible
-    model_name: qwen3-30b-a3b-thinking-2507
-    params:
-      temperature: 0.6
-
-  qwen3_30b_instruct:
-    backend: openai_compatible
-    model_name: qwen3-30b-a3b-instruct-2507
-
-embedding_model:
-  default:
-    backend: openai_compatible
-    model_name: text-embedding-v4
-    params:
-      dimensions: 1024
-
-vector_store:
-  default:
-    backend: elasticsearch
-    embedding_model: default
-    params:
-      hosts: "http://localhost:9200"

flowllm/config/empty.yaml

@@ -1,37 +0,0 @@
-# default config.yaml
-backend: http
-language: ""
-thread_pool_max_workers: 32
-ray_max_workers: 1
-
-mcp:
-  transport: sse
-  host: "0.0.0.0"
-  port: 8001
-
-http:
-  host: "0.0.0.0"
-  port: 8001
-  timeout_keep_alive: 600
-  limit_concurrency: 64
-
-llm:
-  default:
-    backend: openai_compatible
-    model_name: qwen3-30b-a3b-thinking-2507
-    params:
-      temperature: 0.6
-
-embedding_model:
-  default:
-    backend: openai_compatible
-    model_name: text-embedding-v4
-    params:
-      dimensions: 1024
-
-vector_store:
-  default:
-    backend: elasticsearch
-    embedding_model: default
-    params:
-      hosts: "http://localhost:9200"

flowllm/flow/gallery/cmd_flow.py

@@ -1,11 +0,0 @@
-from flowllm.flow.base_flow import BaseFlow
-from flowllm.flow.parser.expression_parser import ExpressionParser
-
-
-class CmdFlow(BaseFlow):
-
-    def build_flow(self):
-        flow: str = self.flow_params["flow"]
-        assert flow, "add `flow=<op_flow>` in cmd!"
-        parser = ExpressionParser(flow)
-        return parser.parse_flow()

flowllm/flow/gallery/code_tool_flow.py

@@ -1,30 +0,0 @@
-from flowllm.context.flow_context import FlowContext
-from flowllm.context.service_context import C
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.op.code.execute_code_op import ExecuteCodeOp
-from flowllm.schema.tool_call import ToolCall
-
-
-@C.register_tool_flow()
-class CodeToolFlow(BaseToolFlow):
-
-    def build_flow(self):
-        return ExecuteCodeOp()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**{
-            "name": "python_execute",
-            "description": "Execute python code can be used in scenarios such as analysis or calculation, and the final result can be printed using the `print` function.",
-            "input_schema": {
-                "code": {
-                    "type": "str",
-                    "description": "code to be executed. Please do not execute any matplotlib code here.",
-                    "required": True
-                }
-            }
-        })
-
-    def return_callback(self, context: FlowContext):
-        context.response.answer = context.code_result
-        return context.response
-

flowllm/flow/gallery/dashscope_search_tool_flow.py

@@ -1,34 +0,0 @@
-from flowllm.context.flow_context import FlowContext
-from flowllm.context.service_context import C
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.op.search import DashscopeSearchOp
-from flowllm.schema.tool_call import ToolCall
-
-
-@C.register_tool_flow()
-class DashscopeSearchToolFlow(BaseToolFlow):
-
-    def build_flow(self):
-        return DashscopeSearchOp()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**{
-            "name": "web_search",
-            "description": "Use search keywords to retrieve relevant information from the internet. If there are multiple search keywords, please use each keyword separately to call this tool.",
-            "input_schema": {
-                "query": {
-                    "type": "str",
-                    "description": "search keyword",
-                    "required": True
-                }
-            }
-        })
-
-    def return_callback(self, context: FlowContext):
-        context.response.answer = context.dashscope_search_result
-        return context.response
-
-
-if __name__ == "__main__":
-    flow = DashscopeSearchToolFlow()
-    flow(query="what is AI?")

flowllm/flow/gallery/deepsearch_tool_flow.py

@@ -1,39 +0,0 @@
-from flowllm.context.flow_context import FlowContext
-from flowllm.context.service_context import C
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.op.search.dashscope_deep_research_op import DashscopeDeepResearchOp
-from flowllm.schema.tool_call import ToolCall
-
-
-@C.register_tool_flow()
-class DeepSearchToolFlow(BaseToolFlow):
-
-    def build_flow(self):
-        return DashscopeDeepResearchOp()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**{
-            "name": "deep_search",
-            "description": "Perform deep research on a topic using Dashscope's qwen-deep-research model. This tool will conduct multi-phase research including model questioning, web research, and result generation.",
-            "input_schema": {
-                "query": {
-                    "type": "str",
-                    "description": "Research topic or question",
-                    "required": True
-                }
-            }
-        })
-
-    def return_callback(self, context: FlowContext):
-        context.response.answer = context.dashscope_deep_research_result
-        return context.response
-
-
-if __name__ == "__main__":
-    from flowllm.utils.common_utils import load_env
-
-    load_env()
-
-    flow = DeepSearchToolFlow()
-    result = flow(query="中国电解铝行业值得投资吗，有哪些值得投资的标的，各个标的之间需要对比优劣势")
-    print(result.answer)

flowllm/flow/gallery/expression_tool_flow.py

@@ -1,18 +0,0 @@
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.flow.parser.expression_parser import ExpressionParser
-from flowllm.schema.service_config import FlowConfig
-from flowllm.schema.tool_call import ToolCall
-
-
-class ExpressionToolFlow(BaseToolFlow):
-
-    def __init__(self, flow_config: FlowConfig = None, **kwargs):
-        self.flow_config: FlowConfig = flow_config
-        super().__init__(name=flow_config.name, **kwargs)
-
-    def build_flow(self):
-        parser = ExpressionParser(self.flow_config.flow_content)
-        return parser.parse_flow()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**self.flow_config.model_dump())

flowllm/flow/gallery/tavily_search_tool_flow.py

@@ -1,30 +0,0 @@
-from flowllm.context.flow_context import FlowContext
-from flowllm.context.service_context import C
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.op.search import TavilySearchOp
-from flowllm.schema.tool_call import ToolCall
-
-
-@C.register_tool_flow()
-class TavilySearchToolFlow(BaseToolFlow):
-
-    def build_flow(self):
-        return TavilySearchOp()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**{
-            "name": "web_search",
-            "description": "Use search keywords to retrieve relevant information from the internet. If there are multiple search keywords, please use each keyword separately to call this tool.",
-            "input_schema": {
-                "query": {
-                    "type": "str",
-                    "description": "search keyword",
-                    "required": True
-                }
-            }
-        })
-
-    def return_callback(self, context: FlowContext):
-        context.response.answer = context.tavily_search_result
-        return context.response
-

flowllm/flow/gallery/terminate_tool_flow.py

@@ -1,30 +0,0 @@
-from flowllm.context.flow_context import FlowContext
-from flowllm.context.service_context import C
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.op.gallery.terminate_op import TerminateOp
-from flowllm.schema.tool_call import ToolCall
-
-
-@C.register_tool_flow()
-class TerminateToolFlow(BaseToolFlow):
-
-    def build_flow(self):
-        return TerminateOp()
-
-    def build_tool_call(self) -> ToolCall:
-        return ToolCall(**{
-            "name": "terminate",
-            "description": "If you can answer the user's question based on the context, be sure to use the **terminate** tool.",
-            "input_schema": {
-                "status": {
-                    "type": "str",
-                    "description": "Please determine whether the user's question has been completed. (success / failure)",
-                    "required": True,
-                    "enum": ["success", "failure"],
-                }
-            }
-        })
-
-    def return_callback(self, context: FlowContext):
-        context.response.answer = context.terminate_answer
-        return context.response

flowllm/op/agent/react_op.py

@@ -1,83 +0,0 @@
-import datetime
-import time
-from typing import List, Dict
-
-from loguru import logger
-
-from flowllm import C, BaseLLMOp
-from flowllm.flow.base_tool_flow import BaseToolFlow
-from flowllm.flow.gallery import DashscopeSearchToolFlow, CodeToolFlow, TerminateToolFlow
-from flowllm.schema.message import Message, Role
-
-
-@C.register_op()
-class ReactOp(BaseLLMOp):
-    # TODO: test react op
-    file_path: str = __file__
-
-    def execute(self):
-        query: str = self.context.query
-
-        max_steps: int = int(self.op_params.get("max_steps", 10))
-        tools: List[BaseToolFlow] = [DashscopeSearchToolFlow(), CodeToolFlow(), TerminateToolFlow()]
-        tool_dict: Dict[str, BaseToolFlow] = {x.name: x for x in tools}
-        now_time = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
-        has_terminate_tool = False
-
-        user_prompt = self.prompt_format(prompt_name="role_prompt",
-                                         time=now_time,
-                                         tools=",".join([x.name for x in tools]),
-                                         query=query)
-        messages: List[Message] = [Message(role=Role.USER, content=user_prompt)]
-        logger.info(f"step.0 user_prompt={user_prompt}")
-
-        for i in range(max_steps):
-            if has_terminate_tool:
-                assistant_message: Message = self.llm.chat(messages)
-            else:
-                assistant_message: Message = self.llm.chat(messages, tools=[x.tool_call for x in tools])
-
-            messages.append(assistant_message)
-            logger.info(f"assistant.{i}.reasoning_content={assistant_message.reasoning_content}\n"
-                        f"content={assistant_message.content}\n"
-                        f"tool.size={len(assistant_message.tool_calls)}")
-
-            if has_terminate_tool:
-                break
-
-            for tool in assistant_message.tool_calls:
-                if tool.name == "terminate":
-                    has_terminate_tool = True
-                    logger.info(f"step={i} find terminate tool, break.")
-                    break
-
-            if not has_terminate_tool and not assistant_message.tool_calls:
-                logger.warning(f"【bugfix】step={i} no tools, break.")
-                has_terminate_tool = True
-
-            for j, tool_call in enumerate(assistant_message.tool_calls):
-                logger.info(f"submit step={i} tool_calls.name={tool_call.name} argument_dict={tool_call.argument_dict}")
-
-                if tool_call.name not in tool_dict:
-                    continue
-
-                self.submit_task(tool_dict[tool_call.name].__call__, **tool_call.argument_dict)
-                time.sleep(1)
-
-            if not has_terminate_tool:
-                user_content_list = []
-                for tool_result, tool_call in zip(self.join_task(), assistant_message.tool_calls):
-                    logger.info(f"submit step={i} tool_calls.name={tool_call.name} tool_result={tool_result}")
-                    assert isinstance(tool_result, str)
-                    user_content_list.append(f"<tool_response>\n{tool_result}\n</tool_response>")
-                user_content_list.append(self.prompt_format(prompt_name="next_prompt"))
-                assistant_message.tool_calls.clear()
-                messages.append(Message(role=Role.USER, content="\n".join(user_content_list)))
-
-            else:
-                assistant_message.tool_calls.clear()
-                messages.append(Message(role=Role.USER, content=self.prompt_format(prompt_name="final_prompt")))
-
-        # Store results in context instead of response
-        self.context.messages = messages
-        self.context.answer = messages[-1].content