ws-bom-robot-app 0.0.33__tar.gz → 0.0.34__tar.gz

{ws_bom_robot_app-0.0.33/ws_bom_robot_app.egg-info → ws_bom_robot_app-0.0.34}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: ws_bom_robot_app
-Version: 0.0.33
+Version: 0.0.34
 Summary: A FastAPI application serving ws bom/robot/llm platform ai.
 Home-page: https://github.com/websolutespa/bom
 Author: Websolute Spa
@@ -13,17 +13,25 @@ Description-Content-Type: text/markdown
 Requires-Dist: standardwebhooks==1.0.0
 Requires-Dist: apscheduler==3.11.0
 Requires-Dist: aiofiles==24.1.0
-Requires-Dist: pydantic==2.10.5
+Requires-Dist: pydantic==2.10.6
 Requires-Dist: pydantic-settings==2.7.1
-Requires-Dist: fastapi[standard]==0.115.6
-Requires-Dist: langchain==0.3.14
-Requires-Dist: langchain-openai==0.3.0
-Requires-Dist: langchain-community==0.3.14
-Requires-Dist: langchain-core==0.3.29
+Requires-Dist: fastapi[standard]==0.115.8
+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: langchain-anthropic==0.3.6
+Requires-Dist: langchain-google-genai==2.0.7
+Requires-Dist: langchain-google-vertexai==2.0.13
 Requires-Dist: faiss-cpu==1.9.0
-Requires-Dist: unstructured==0.16.13
+Requires-Dist: chromadb==0.6.3
+Requires-Dist: langchain_chroma==0.2.1
+Requires-Dist: fastembed==0.5.1
+Requires-Dist: langchain-qdrant==0.2.0
+Requires-Dist: lark==1.2.2
+Requires-Dist: unstructured==0.16.17
 Requires-Dist: unstructured[image]
-Requires-Dist: unstructured-ingest==0.3.14
+Requires-Dist: unstructured-ingest==0.4.6
 Requires-Dist: unstructured-ingest[azure]
 Requires-Dist: unstructured-ingest[confluence]
 Requires-Dist: unstructured-ingest[dropbox]
@@ -185,6 +193,7 @@ py -m pip install --upgrade setuptools build twine streamlit
 ### 🪛 build
 
 ```pwsh
+if (Test-Path ./dist) {rm ./dist -r -force}; `
 py -m build && twine check dist/*
 ```
 
@@ -217,14 +226,18 @@ prospector ./ws_bom_robot_app -t pyroma
 lauch pytest
 
 ```pwsh
-!py -m pip install -U pytest pytest-asyncio pytest-mock pytest-cov
+!py -m pip install -U pytest pytest-asyncio pytest-mock pytest-cov pyclean
+# clean cache if needed
+# pyclean --verbose .
 pytest --cov=ws_bom_robot_app --log-cli-level=info
+# directory
+# pytest --cov=ws_bom_robot_app --log-cli-level=info ./tests/app/llm/vector_store/db
 ```
 
 launch debugger
 
 ```pwsh
-streamlit run debugger.py --server.port 6002
+streamlit run debugger.py --server.port 6011
 ```
 
 dockerize base image
@@ -242,7 +255,7 @@ dockerize app from src
 
 ```pwsh
 docker build -f Dockerfile-src -t ws-bom-robot-app:src .
-docker run --name ws-bom-robot-app-src -d -v "$(pwd)/ws_bom_robot_app:/app/ws_bom_robot_app" -v "$(pwd)/.data:/app/.data" -v "$(pwd)/tests:/app/tests" -v "$(pwd)/tmp:/tmp"  -p 6001:6001 ws-bom-robot-app:src
+docker run --name ws-bom-robot-app-src -d -v "$(pwd)/ws_bom_robot_app:/app/ws_bom_robot_app" -v "$(pwd)/.data:/app/.data" -v "$(pwd)/tests:/app/tests" -v "$(pwd)/tmp:/tmp"  -p 6002:6001 ws-bom-robot-app:src
 ```
 
 ### ✈️ publish

{ws_bom_robot_app-0.0.33 → ws_bom_robot_app-0.0.34}/README.md

@@ -135,6 +135,7 @@ py -m pip install --upgrade setuptools build twine streamlit
 ### 🪛 build
 
 ```pwsh
+if (Test-Path ./dist) {rm ./dist -r -force}; `
 py -m build && twine check dist/*
 ```
 
@@ -167,14 +168,18 @@ prospector ./ws_bom_robot_app -t pyroma
 lauch pytest
 
 ```pwsh
-!py -m pip install -U pytest pytest-asyncio pytest-mock pytest-cov
+!py -m pip install -U pytest pytest-asyncio pytest-mock pytest-cov pyclean
+# clean cache if needed
+# pyclean --verbose .
 pytest --cov=ws_bom_robot_app --log-cli-level=info
+# directory
+# pytest --cov=ws_bom_robot_app --log-cli-level=info ./tests/app/llm/vector_store/db
 ```
 
 launch debugger
 
 ```pwsh
-streamlit run debugger.py --server.port 6002
+streamlit run debugger.py --server.port 6011
 ```
 
 dockerize base image
@@ -192,7 +197,7 @@ dockerize app from src
 
 ```pwsh
 docker build -f Dockerfile-src -t ws-bom-robot-app:src .
-docker run --name ws-bom-robot-app-src -d -v "$(pwd)/ws_bom_robot_app:/app/ws_bom_robot_app" -v "$(pwd)/.data:/app/.data" -v "$(pwd)/tests:/app/tests" -v "$(pwd)/tmp:/tmp"  -p 6001:6001 ws-bom-robot-app:src
+docker run --name ws-bom-robot-app-src -d -v "$(pwd)/ws_bom_robot_app:/app/ws_bom_robot_app" -v "$(pwd)/.data:/app/.data" -v "$(pwd)/tests:/app/tests" -v "$(pwd)/tmp:/tmp"  -p 6002:6001 ws-bom-robot-app:src
 ```
 
 ### ✈️ publish

ws_bom_robot_app-0.0.34/requirements.txt

@@ -0,0 +1,48 @@
+#app
+standardwebhooks==1.0.0
+apscheduler==3.11.0
+aiofiles==24.1.0
+pydantic==2.10.6
+pydantic-settings==2.7.1
+fastapi[standard]==0.115.8
+
+#framework
+langchain==0.3.18
+langchain-community==0.3.17
+langchain-core==0.3.34
+langchain-openai==0.3.5
+langchain-anthropic==0.3.6 #issue get_models() from 0.3.7
+langchain-google-genai==2.0.7 #waiting for new release: https://github.com/langchain-ai/langchain-google/issues/711
+langchain-google-vertexai==2.0.13
+
+#vector DB
+faiss-cpu==1.9.0
+chromadb==0.6.3
+langchain_chroma==0.2.1
+fastembed==0.5.1 #qdrant sparse embedding
+langchain-qdrant==0.2.0
+lark==1.2.2 #self-query retriever
+
+
+#loaders
+unstructured==0.16.17
+unstructured[image]
+unstructured-ingest==0.4.6
+unstructured-ingest[azure]
+unstructured-ingest[confluence]
+unstructured-ingest[dropbox]
+unstructured-ingest[gcs]
+unstructured-ingest[github]
+unstructured-ingest[google_drive]
+unstructured-ingest[jira]
+unstructured-ingest[s3]
+unstructured-ingest[sftp]
+unstructured-ingest[sharepoint]
+unstructured-ingest[slack]
+html5lib==1.1 #beautifulsoup4 parser
+
+#integrations
+markdownify==0.14.1 #sitemap
+
+#telemetry
+nebuly==0.3.36

{ws_bom_robot_app-0.0.33 → ws_bom_robot_app-0.0.34}/setup.py

@@ -4,7 +4,7 @@ _requirements = [line.split('#')[0].strip() for line in open("requirements.txt")
 
 setup(
     name="ws_bom_robot_app",
-    version="0.0.33",
+    version="0.0.34",
     description="A FastAPI application serving ws bom/robot/llm platform ai.",
     long_description=open("README.md", encoding='utf-8').read(),
     long_description_content_type="text/markdown",

{ws_bom_robot_app-0.0.33 → ws_bom_robot_app-0.0.34}/ws_bom_robot_app/config.py

@@ -20,12 +20,21 @@ class Settings(BaseSettings):
     robot_cms_auth: str = ''
     robot_cms_db_folder: str = 'llmVectorDb'
     robot_cms_kb_folder: str ='llmKbFile'
-    robot_debugger_openai_key: str = ''
+    robot_debugger_llm_provider: str = 'openai'
+    robot_debugger_llm_model: str = 'gpt-4o'
+    robot_debugger_llm_key: str = ''
+    robot_debugger_embedding_key: str = ''
+    OPENAI_API_KEY: str = '' # used for dall-e api
+    GOOGLE_APPLICATION_CREDENTIALS: str = './.data/credentials/google.json' # path to google credentials iam file
     model_config = ConfigDict(
         env_file='./.env',
         extra='ignore',
         case_sensitive=False
     )
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        os.environ["OPENAI_API_KEY"] = self.OPENAI_API_KEY
+        os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = self.GOOGLE_APPLICATION_CREDENTIALS
 
     class RuntimeOptions(BaseModel):
         @staticmethod

{ws_bom_robot_app-0.0.33 → ws_bom_robot_app-0.0.34}/ws_bom_robot_app/llm/agent_description.py

@@ -1,124 +1,123 @@
-import json, requests, re
-from typing import Any
-from abc import ABC, abstractmethod
-from langchain_openai import ChatOpenAI
-from langchain_core.prompts import ChatPromptTemplate
-from langchain_core.messages import AIMessage
-from langchain_core.runnables import RunnableSerializable
-from langchain_core.runnables import RunnableLambda
-from bs4 import BeautifulSoup
-from ws_bom_robot_app.llm.models.api import LlmRules
-from ws_bom_robot_app.llm.utils.agent_utils import get_rules
-
-# SafeDict helper class
-class SafeDict(dict):
-    def __missing__(self, key):
-        return ''
-
-# Strategy Interface
-class AgentDescriptorStrategy(ABC):
-    @abstractmethod
-    def enrich_prompt(self, prompt: str, input: dict) -> str:
-        pass
-
-    @abstractmethod
-    def rule_input(self, input: dict) -> str:
-        pass
-
-# Concrete Strategy for Default Agent
-class DefaultAgentDescriptor(AgentDescriptorStrategy):
-    def enrich_prompt(self, prompt: str, input: dict) -> str:
-        # Default enrichment logic (could be minimal or no-op)
-        return prompt.format_map(SafeDict(input))
-
-    def rule_input(self, input: dict) -> str:
-        return input.get('content', "")
-
-# Concrete Strategy for URL2Text Agent
-class URL2TextAgentDescriptor(AgentDescriptorStrategy):
-    def enrich_prompt(self, prompt: str, input: dict) -> str:
-        input["context"] = self._get_page_text(input)
-        return prompt.format_map(SafeDict(input))
-
-    def rule_input(self, input: dict) -> str:
-        return input.get('context', "")
-
-    def _get_page_text(self, input: dict) -> str:
-        url = input.get("content", "")
-        exclusions = input.get("exclude", {})
-        response = requests.get(url)
-        response.raise_for_status()
-        soup = BeautifulSoup(response.content, 'html5lib')
-        classes_to_exclude = exclusions.get("classes", [])
-        ids_to_exclude = exclusions.get("ids", [])
-        for class_name in classes_to_exclude:
-            for element in soup.find_all(class_=class_name):
-                element.extract()
-        for id_name in ids_to_exclude:
-            for element in soup.find_all(id=id_name):
-                element.extract()
-        for script in soup(["script", "noscript", "style", "head", "footer", "iframe"]):
-            script.extract()
-        return re.sub(' +', ' ', soup.get_text())
-
-
-class AgentDescriptor:
-    # Dictionary to hold all agent strategies
-    _list: dict[str,AgentDescriptorStrategy] = {
-        "default": DefaultAgentDescriptor(),
-        "url2text": URL2TextAgentDescriptor(),
-    }
-
-    # Functions to manage strategies
-    @staticmethod
-    def add_strategy(name: str, strategy: AgentDescriptorStrategy):
-        """_summary_
-            add a new strategy to the dictionary
-        Args:
-            name (str): name of the strategy, in lowercase
-            strategy (AgentDescriptorStrategy): class implementing the strategy
-        Examples:
-            AgentDescriptor.add_strategy("custom_agent_descriptor", CustomAgentDescriptor())
-        """
-        AgentDescriptor._list[name.lower()] = strategy
-
-    @staticmethod
-    def get_strategy(name: str) -> AgentDescriptorStrategy:
-        return AgentDescriptor._list.get(name.lower(), DefaultAgentDescriptor())
-
-    def __init__(self, api_key: str, prompt: str, mode: str, rules: LlmRules = None):
-        self.__prompt = prompt
-        self.__llm = ChatOpenAI(model="gpt-4o", temperature=0, api_key=api_key)  # type: ignore
-        self.api_key = api_key
-        self.rules= rules
-        self.strategy = self.get_strategy(mode)  # Selects the strategy from the dictionary
-
-    async def __create_prompt(self, input_dict: dict):
-        input_data = json.loads(input_dict.get("input", {}))
-        system = self.strategy.enrich_prompt(self.__prompt, input_data)
-        if self.rules:
-            rule_input = self.strategy.rule_input(input_data)
-            rules_prompt = await get_rules(self.rules,self.api_key, rule_input)
-            system += rules_prompt
-        return ChatPromptTemplate.from_messages(
-            [
-                ("system", system),
-                ("user", input_data.get("content", ""))
-            ]
-        )
-
-    def __create_agent_descriptor(self, content) -> RunnableSerializable[Any, Any]:
-        content = json.loads(content)
-        agent = (
-            {
-                "input": lambda x: x["input"],
-            }
-            | RunnableLambda(self.__create_prompt)
-            | self.__llm
-        )
-        return agent
-
-    async def run_agent(self, content) -> Any:
-        agent_descriptor = self.__create_agent_descriptor(content)
-        response: AIMessage = await agent_descriptor.ainvoke({"input": content})
-        return response
+import json, requests, re
+from typing import Any
+from abc import ABC, abstractmethod
+from langchain_core.prompts import ChatPromptTemplate
+from langchain_core.messages import AIMessage
+from langchain_core.runnables import RunnableSerializable
+from langchain_core.runnables import RunnableLambda
+from bs4 import BeautifulSoup
+from ws_bom_robot_app.llm.models.api import LlmRules
+from ws_bom_robot_app.llm.providers.llm_manager import LlmInterface
+from ws_bom_robot_app.llm.utils.agent_utils import get_rules
+
+# SafeDict helper class
+class SafeDict(dict):
+    def __missing__(self, key):
+        return ''
+
+# Strategy Interface
+class AgentDescriptorStrategy(ABC):
+    @abstractmethod
+    def enrich_prompt(self, prompt: str, input: dict) -> str:
+        pass
+
+    @abstractmethod
+    def rule_input(self, input: dict) -> str:
+        pass
+
+# Concrete Strategy for Default Agent
+class DefaultAgentDescriptor(AgentDescriptorStrategy):
+    def enrich_prompt(self, prompt: str, input: dict) -> str:
+        # Default enrichment logic (could be minimal or no-op)
+        return prompt.format_map(SafeDict(input))
+
+    def rule_input(self, input: dict) -> str:
+        return input.get('content', "")
+
+# Concrete Strategy for URL2Text Agent
+class URL2TextAgentDescriptor(AgentDescriptorStrategy):
+    def enrich_prompt(self, prompt: str, input: dict) -> str:
+        input["context"] = self._get_page_text(input)
+        return prompt.format_map(SafeDict(input))
+
+    def rule_input(self, input: dict) -> str:
+        return input.get('context', "")
+
+    def _get_page_text(self, input: dict) -> str:
+        url = input.get("content", "")
+        exclusions = input.get("exclude", {})
+        response = requests.get(url)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.content, 'html5lib')
+        classes_to_exclude = exclusions.get("classes", [])
+        ids_to_exclude = exclusions.get("ids", [])
+        for class_name in classes_to_exclude:
+            for element in soup.find_all(class_=class_name):
+                element.extract()
+        for id_name in ids_to_exclude:
+            for element in soup.find_all(id=id_name):
+                element.extract()
+        for script in soup(["script", "noscript", "style", "head", "footer", "iframe"]):
+            script.extract()
+        return re.sub(' +', ' ', soup.get_text())
+
+
+class AgentDescriptor:
+    # Dictionary to hold all agent strategies
+    _list: dict[str,AgentDescriptorStrategy] = {
+        "default": DefaultAgentDescriptor(),
+        "url2text": URL2TextAgentDescriptor(),
+    }
+
+    # Functions to manage strategies
+    @staticmethod
+    def add_strategy(name: str, strategy: AgentDescriptorStrategy):
+        """_summary_
+            add a new strategy to the dictionary
+        Args:
+            name (str): name of the strategy, in lowercase
+            strategy (AgentDescriptorStrategy): class implementing the strategy
+        Examples:
+            AgentDescriptor.add_strategy("custom_agent_descriptor", CustomAgentDescriptor())
+        """
+        AgentDescriptor._list[name.lower()] = strategy
+
+    @staticmethod
+    def get_strategy(name: str) -> AgentDescriptorStrategy:
+        return AgentDescriptor._list.get(name.lower(), DefaultAgentDescriptor())
+
+    def __init__(self, llm: LlmInterface, prompt: str, mode: str, rules: LlmRules = None):
+        self.__prompt = prompt
+        self.__llm = llm,
+        self.rules= rules
+        self.strategy = self.get_strategy(mode)  # Selects the strategy from the dictionary
+
+    async def __create_prompt(self, input_dict: dict):
+        input_data = json.loads(input_dict.get("input", {}))
+        system = self.strategy.enrich_prompt(self.__prompt, input_data)
+        if self.rules:
+            rule_input = self.strategy.rule_input(input_data)
+            rules_prompt = await get_rules(self.__llm.get_embeddings(), self.rules, rule_input)
+            system += rules_prompt
+        return ChatPromptTemplate.from_messages(
+            [
+                ("system", system),
+                ("user", input_data.get("content", ""))
+            ]
+        )
+
+    def __create_agent_descriptor(self, content) -> RunnableSerializable[Any, Any]:
+        content = json.loads(content)
+        agent = (
+            {
+                "input": lambda x: x["input"],
+            }
+            | RunnableLambda(self.__create_prompt)
+            | self.__llm.get_llm()
+        )
+        return agent
+
+    async def run_agent(self, content) -> Any:
+        agent_descriptor = self.__create_agent_descriptor(content)
+        response: AIMessage = await agent_descriptor.ainvoke({"input": content})
+        return response