ws-bom-robot-app 0.0.57__py3-none-any.whl → 0.0.59__py3-none-any.whl

ws_bom_robot_app/config.py

@@ -28,6 +28,7 @@ class Settings(BaseSettings):
     GOOGLE_API_KEY: str = ''
     NEBULY_API_URL: str =''
     GOOGLE_APPLICATION_CREDENTIALS: str = '' # path to google credentials iam file, e.d. ./.secrets/google-credentials.json
+    TAVILY_API_KEY: str = '' #TODO DELETE
     model_config = ConfigDict(
         env_file='./.env',
         extra='ignore',
@@ -43,6 +44,8 @@ class Settings(BaseSettings):
         os.environ["GOOGLE_API_KEY"] = self.GOOGLE_API_KEY
         os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = self.GOOGLE_APPLICATION_CREDENTIALS
         os.environ["NEBULY_API_URL"] = self.NEBULY_API_URL
+        # TODO DELETE
+        os.environ["TAVILY_API_KEY"] = self.TAVILY_API_KEY
 
     class RuntimeOptions(BaseModel):
         @staticmethod

ws_bom_robot_app/llm/agent_handler.py

@@ -41,7 +41,6 @@ class AgentHandler(AsyncCallbackHandler):
         self.json_start_regex = re.compile(r'(`{1,3}\s*json\b)') # detect a potential json start fence.
         self.json_end_regex = re.compile(r'(`{1,3})')         # an end fence (one to three backticks).
         self.stream_cut_last_output_chunk_size = 16  # safe cut last chunk size to output if no markers are found
-
     async def on_chat_model_start(self, serialized, messages, *, run_id, parent_run_id = None, tags = None, metadata = None, **kwargs):
         if not self.__started:
           self.__started = True
@@ -102,7 +101,7 @@ class AgentHandler(AsyncCallbackHandler):
                       finally:
                           self.json_buffer = ""
                       # remove the end fence from pending.
-                      self.stream_buffer = self.stream_buffer[end_match.end():].replace('`','').strip()
+                      self.stream_buffer = self.stream_buffer[end_match.end():].strip()
                       self.in_json_block = False
                   else:
                       # no end marker found

ws_bom_robot_app/llm/providers/llm_manager.py

@@ -48,7 +48,7 @@ class OpenAI(LlmInterface):
             api_key=self.config.api_key or os.getenv("OPENAI_API_KEY"),
             model=self.config.model,
             stream_usage=True)
-        if not any(self.config.model.startswith(prefix) for prefix in ["o1", "o3"]):
+        if not (any(self.config.model.startswith(prefix) for prefix in ["o1", "o3"]) or "search" in self.config.model):
             chat.temperature = self.config.temperature
             chat.streaming = True
         return chat

ws_bom_robot_app/llm/tools/models/main.py

@@ -7,3 +7,7 @@ class DocumentRetrieverInput(BaseModel):
 class ImageGeneratorInput(BaseModel):
   query: str = Field(description="description of the image to generate.")
   language: str = Field(description="Language of the query. Default is 'it'", default="it")
+class LlmChainInput(BaseModel):
+  input: str = Field(description="Input to the LLM chain")
+class SearchOnlineInput(BaseModel):
+  query: str = Field(description="The search query string")

ws_bom_robot_app/llm/tools/tool_manager.py

@@ -4,7 +4,7 @@ from ws_bom_robot_app.llm.models.api import LlmAppTool
 from ws_bom_robot_app.llm.providers.llm_manager import LlmInterface
 from ws_bom_robot_app.llm.vector_store.db.manager import VectorDbManager
 from ws_bom_robot_app.llm.tools.utils import getRandomWaitingMessage, translate_text
-from ws_bom_robot_app.llm.tools.models.main import NoopInput,DocumentRetrieverInput,ImageGeneratorInput
+from ws_bom_robot_app.llm.tools.models.main import NoopInput,DocumentRetrieverInput,ImageGeneratorInput,LlmChainInput,SearchOnlineInput
 from pydantic import BaseModel, ConfigDict
 
 class ToolConfig(BaseModel):
@@ -41,6 +41,39 @@ class ToolManager:
         self.callbacks = callbacks
         self.queue = queue
 
+    async def __extract_documents(self, query: str, app_tool: LlmAppTool):
+        search_type = "similarity"
+        search_kwargs = {"k": 4}
+        if app_tool.search_settings:
+            search_settings = app_tool.search_settings # type: ignore
+            if search_settings.search_type == "similarityScoreThreshold":
+                search_type = "similarity_score_threshold"
+                search_kwargs = {
+                    "score_threshold": search_settings.score_threshold_id if search_settings.score_threshold_id else  0.5,
+                    "k": search_settings.search_k if search_settings.search_k else 100
+                }
+            elif search_settings.search_type == "mmr":
+                search_type = "mmr"
+                search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
+            elif search_settings.search_type == "default":
+                search_type = "similarity"
+                search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
+            else:
+                search_type = "mixed"
+                search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
+        if self.queue:
+          await self.queue.put(getRandomWaitingMessage(app_tool.waiting_message, traduction=False))
+
+        return await VectorDbManager.get_strategy(app_tool.vector_type).invoke(
+            self.llm.get_embeddings(),
+            app_tool.vector_db,
+            query,
+            search_type,
+            search_kwargs,
+            app_tool=app_tool,
+            llm=self.llm.get_llm(),
+            source=app_tool.function_id,
+            )
 
     #region functions
     async def document_retriever(self, query: str) -> list:
@@ -65,40 +98,7 @@ class ToolManager:
             self.app_tool.type == "function" and self.app_tool.vector_db
             #and self.settings.get("dataSource") == "knowledgebase"
         ):
-            search_type = "similarity"
-            search_kwargs = {"k": 4}
-            if self.app_tool.search_settings:
-                search_settings = self.app_tool.search_settings # type: ignore
-                if search_settings.search_type == "similarityScoreThreshold":
-                    search_type = "similarity_score_threshold"
-                    search_kwargs = {
-                        "score_threshold": search_settings.score_threshold_id if search_settings.score_threshold_id else  0.5,
-                        "k": search_settings.search_k if search_settings.search_k else 100
-                    }
-                elif search_settings.search_type == "mmr":
-                    search_type = "mmr"
-                    search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
-                elif search_settings.search_type == "default":
-                    search_type = "similarity"
-                    search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
-                else:
-                    search_type = "mixed"
-                    search_kwargs = {"k": search_settings.search_k if search_settings.search_k else 4}
-            if self.queue:
-              await self.queue.put(getRandomWaitingMessage(self.app_tool.waiting_message, traduction=False))
-
-            return await VectorDbManager.get_strategy(self.app_tool.vector_type).invoke(
-                self.llm.get_embeddings(),
-                self.app_tool.vector_db,
-                query,
-                search_type,
-                search_kwargs,
-                app_tool=self.app_tool,
-                llm=self.llm.get_llm(),
-                source=self.app_tool.function_id,
-                )
-        return []
-        #raise ValueError(f"Invalid configuration for {self.settings.name} tool of type {self.settings.type}. Must be a function or vector db not found.")
+            return await self.__extract_documents(query, self.app_tool)
 
     async def image_generator(self, query: str, language: str = "it"):
         """
@@ -120,12 +120,66 @@ class ToolManager:
         except Exception as e:
             return f"Error: {str(e)}"
 
+    async def llm_chain(self, input: str):
+        if self.app_tool.type == "llmChain":
+          from langchain_core.prompts import ChatPromptTemplate
+          from langchain_core.output_parsers import StrOutputParser
+          system_message = self.app_tool.llm_chain_settings.prompt
+          context = []
+          if self.app_tool.data_source == "knowledgebase":
+            context = await self.__extract_documents(input, self.app_tool)
+          if len(context) > 0:
+            for doc in context:
+              system_message += f"\n\nContext:\n{doc.metadata.get("source", "")}: {doc.page_content}"
+          prompt = ChatPromptTemplate.from_messages(
+              [ ("system", system_message),
+                ("user", "{input}")],
+          )
+          model = self.app_tool.llm_chain_settings.model
+          self.llm.config.model = model
+          llm = self.llm.get_llm()
+          chain = prompt | llm | StrOutputParser()
+          result = await chain.ainvoke({"input": input})
+          return result
+
+
+    async def search_online(self, query: str):
+        from ws_bom_robot_app.llm.tools.utils import fetch_page, extract_content_with_trafilatura
+        from langchain_community.utilities import DuckDuckGoSearchAPIWrapper
+        import aiohttp, asyncio, ast
+        # Wrapper DuckDuckGo
+        search = DuckDuckGoSearchAPIWrapper(max_results=10)
+        try:
+          raw_results = search.results(query, max_results=10)
+        except Exception as e:
+            print(f"[!] Errore ricerca: {e}")
+        urls = [r["link"] for r in raw_results]
+        async with aiohttp.ClientSession() as session:
+            tasks = [fetch_page(session, url) for url in urls]
+            responses = await asyncio.gather(*tasks)
+        final_results = []
+        for item in responses:
+            url = item["url"]
+            html = item["html"]
+            if html:
+                content = await extract_content_with_trafilatura(html)
+                if content:
+                    final_results.append({"url": url, "content": content})
+                else:
+                    final_results.append({"url": url, "content": "No content found"})
+            else:
+                final_results.append({"url": url, "content": "Page not found"})
+        return final_results
+
+
     #endregion
 
     #class variables (static)
     _list: dict[str,ToolConfig] = {
         "document_retriever": ToolConfig(function=document_retriever, model=DocumentRetrieverInput),
         "image_generator": ToolConfig(function=image_generator, model=ImageGeneratorInput),
+        "llm_chain": ToolConfig(function=llm_chain, model=LlmChainInput),
+        "search_online": ToolConfig(function=search_online, model=SearchOnlineInput),
     }
 
     #instance methods

ws_bom_robot_app/llm/tools/utils.py

@@ -23,3 +23,19 @@ async def translate_text(llm: LlmInterface, language, text: str, callbacks: list
   prompt = PromptTemplate.from_template(sys_message)
   chain = prompt | llm.get_llm()
   await chain.ainvoke({"language":language, "testo_da_tradurre": text}, {"callbacks": callbacks})
+
+async def fetch_page(session, url):
+    try:
+        async with session.get(url, timeout=10, ssl=False) as response:
+            if response.status == 200:
+                text = await response.text()
+                return {"url": url, "html": text}
+            else:
+                return {"url": url, "html": None}
+    except Exception as e:
+        return {"url": url, "html": None}
+
+async def extract_content_with_trafilatura(html):
+    """Estrae solo il testo principale usando trafilatura"""
+    import trafilatura
+    return trafilatura.extract(html)

{ws_bom_robot_app-0.0.57.dist-info → ws_bom_robot_app-0.0.59.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ws_bom_robot_app
-Version: 0.0.57
+Version: 0.0.59
 Summary: A FastAPI application serving ws bom/robot/llm platform ai.
 Home-page: https://github.com/websolutespa/bom
 Author: Websolute Spa
@@ -17,15 +17,16 @@ Requires-Dist: pydantic==2.10.6
 Requires-Dist: pydantic-settings==2.7.1
 Requires-Dist: fastapi[standard]==0.115.8
 Requires-Dist: chevron==0.14.0
-Requires-Dist: langchain==0.3.18
-Requires-Dist: langchain-community==0.3.17
-Requires-Dist: langchain-core==0.3.34
-Requires-Dist: langchain-openai==0.3.5
+Requires-Dist: trafilatura==2.0.0
+Requires-Dist: langchain==0.3.25
+Requires-Dist: langchain-community==0.3.24
+Requires-Dist: langchain-core==0.3.59
+Requires-Dist: langchain-openai==0.3.16
 Requires-Dist: langchain-anthropic==0.3.6
 Requires-Dist: langchain-google-genai==2.0.7
 Requires-Dist: langchain-google-vertexai==2.0.13
-Requires-Dist: langchain-groq==0.2.4
-Requires-Dist: langchain-ollama==0.2.3
+Requires-Dist: langchain-groq==0.3.2
+Requires-Dist: langchain-ollama==0.3.2
 Requires-Dist: faiss-cpu==1.9.0
 Requires-Dist: chromadb==0.6.3
 Requires-Dist: langchain_chroma==0.2.1
@@ -48,6 +49,7 @@ Requires-Dist: unstructured-ingest[sharepoint]
 Requires-Dist: unstructured-ingest[slack]
 Requires-Dist: html5lib==1.1
 Requires-Dist: markdownify==0.14.1
+Requires-Dist: duckduckgo-search==8.0.4
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{ws_bom_robot_app-0.0.57.dist-info → ws_bom_robot_app-0.0.59.dist-info}/RECORD

@@ -1,6 +1,6 @@
 ws_bom_robot_app/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ws_bom_robot_app/auth.py,sha256=84nIbmJsMrNs0sxIQGEHbjsjc2P6ZrZZGSn8dkiL6is,895
-ws_bom_robot_app/config.py,sha256=XMr7SOU5Citev1jwfab_7UDaS6ALcXlUhYDyTONSJro,4057
+ws_bom_robot_app/config.py,sha256=9W3cz92hbogDrtbCcybpCY29zCka7G225CNQEptpx30,4183
 ws_bom_robot_app/cron_manager.py,sha256=0Yt5AMTPGlXZ_M5ck0SKMX8wvzoPsseEezg_s0Q3HKY,9224
 ws_bom_robot_app/main.py,sha256=zO3B-v-v9ESASvw8IaQj9Y9hNvNmOxohFmA0R82EybQ,6518
 ws_bom_robot_app/task_manager.py,sha256=Zedzs2R3O-wNSQOqs4jorgFwPRi-ji_0TN4mGfk-VvE,15958
@@ -8,7 +8,7 @@ ws_bom_robot_app/util.py,sha256=b49ItlZgh2Wzw-6K8k5Wa44eVgjQ0JmWQwJnEaQBVGw,3502
 ws_bom_robot_app/llm/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ws_bom_robot_app/llm/agent_context.py,sha256=uatHJ8wcRly6h0S762BgfzDMpmcwCHwNzwo37aWjeE0,1305
 ws_bom_robot_app/llm/agent_description.py,sha256=5IP0qFSJvaE3zjGS7f0W1DuiegP0RHXRMBoDC5pCofA,4779
-ws_bom_robot_app/llm/agent_handler.py,sha256=_mh2anSdkSnXncXBYnKars7zPd-PVewnk8O8du1s4do,7756
+ws_bom_robot_app/llm/agent_handler.py,sha256=4HYP8wbdtJhRi3bk6PvJ3cRDZyLYWt3Ow5tnHpkEg1o,7738
 ws_bom_robot_app/llm/agent_lcel.py,sha256=8d10b43BXqE4rfXE5uh8YGT67o1bw0q0l7QXFT6wPKA,2320
 ws_bom_robot_app/llm/api.py,sha256=1nzQ7g2n_DlX6Ixo5ecS10UvyyKJ42qZQ6aD8-EI7BE,4709
 ws_bom_robot_app/llm/defaut_prompt.py,sha256=D9dn8yPveu0bVwGM1wQWLYftmBs5O76o0R_caLLll8w,1121
@@ -23,13 +23,13 @@ ws_bom_robot_app/llm/models/base.py,sha256=1TqxuTK3rjJEALn7lvgoen_1ba3R2brAgGx6E
 ws_bom_robot_app/llm/models/feedback.py,sha256=pYNQGxNOBgeAAfdJLI95l7ePLBI5tVdsgnyjp5oMOQU,1722
 ws_bom_robot_app/llm/models/kb.py,sha256=oVSw6_dmNxikAHrPqcfxDXz9M0ezLIYuxpgvzfs_Now,9514
 ws_bom_robot_app/llm/providers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-ws_bom_robot_app/llm/providers/llm_manager.py,sha256=C63GJOaJHcKraDYZgemOdh8F2UPrheH_0E-gdhmpD6s,8336
+ws_bom_robot_app/llm/providers/llm_manager.py,sha256=zIkxgTLYQCcup2Ixf4eWap4mNinuJH2YmkjLjZGDyJM,8371
 ws_bom_robot_app/llm/tools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ws_bom_robot_app/llm/tools/tool_builder.py,sha256=p_Q32_-OSydcxzj69PgPIuiny816zYv5dVsCHSY0ELc,1188
-ws_bom_robot_app/llm/tools/tool_manager.py,sha256=xyGDa4BnPrdZv4-AL_kXIwaUpD7JOVO4FAhLbSTuujg,6001
-ws_bom_robot_app/llm/tools/utils.py,sha256=LEId1UolLvJsMYbl-awm7h-gJ9Up9DrrnH6HTglGqVE,1347
+ws_bom_robot_app/llm/tools/tool_manager.py,sha256=adWvaSIur5Ez2gGsuTFqNprZZlQP6ZZj5WknzJjtQ0c,8355
+ws_bom_robot_app/llm/tools/utils.py,sha256=Ba7ScFZPVJ3ke8KLO8ik1wyR2f_zC99Bikqx0OGnKoI,1924
 ws_bom_robot_app/llm/tools/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-ws_bom_robot_app/llm/tools/models/main.py,sha256=o3Rwbn5nsugKOgLG0FUIuvtPPHYhfVpqG4E3BQB2nWM,388
+ws_bom_robot_app/llm/tools/models/main.py,sha256=pBQNWPd1OZgZ2xkOnUOawNbujQ5oJXLdyuAex1afLWc,579
 ws_bom_robot_app/llm/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ws_bom_robot_app/llm/utils/agent.py,sha256=ISF9faaD5tBi-8sbgQpgfqWT1JIVcgv_lRhyaNAkI2Q,1445
 ws_bom_robot_app/llm/utils/chunker.py,sha256=N7570xBYlObneg-fsvDhPAJ-Pv8C8OaYZOBK6q7LmMI,607
@@ -65,7 +65,7 @@ ws_bom_robot_app/llm/vector_store/loader/__init__.py,sha256=47DEQpj8HBSa-_TImW-5
 ws_bom_robot_app/llm/vector_store/loader/base.py,sha256=L_ugekNuAq0N9O-24wtlHSNHkqSeD-KsJrfGt_FX9Oc,5340
 ws_bom_robot_app/llm/vector_store/loader/docling.py,sha256=yP0zgXLeFAlByaYuj-6cYariuknckrFds0dxdRcnVz8,3456
 ws_bom_robot_app/llm/vector_store/loader/json_loader.py,sha256=LDppW0ZATo4_1hh-KlsAM3TLawBvwBxva_a7k5Oz1sc,858
-ws_bom_robot_app-0.0.57.dist-info/METADATA,sha256=pNu9PdNL8evyWeVAD8LyHUMG-MPY0nYICCHdS4mC4oA,8329
-ws_bom_robot_app-0.0.57.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-ws_bom_robot_app-0.0.57.dist-info/top_level.txt,sha256=Yl0akyHVbynsBX_N7wx3H3ZTkcMLjYyLJs5zBMDAKcM,17
-ws_bom_robot_app-0.0.57.dist-info/RECORD,,
+ws_bom_robot_app-0.0.59.dist-info/METADATA,sha256=hTSpZWSUDrn0IY9jsYvUCGocnzK9kbrEiRINEwiDVe0,8406
+ws_bom_robot_app-0.0.59.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+ws_bom_robot_app-0.0.59.dist-info/top_level.txt,sha256=Yl0akyHVbynsBX_N7wx3H3ZTkcMLjYyLJs5zBMDAKcM,17
+ws_bom_robot_app-0.0.59.dist-info/RECORD,,