langroid 0.39.5__py3-none-any.whl → 0.41.0__py3-none-any.whl

langroid/agent/tools/exa_search_tool.py

@@ -0,0 +1,68 @@
+"""
+A tool to trigger a Exa search for a given query,
+(https://docs.exa.ai/reference/getting-started)
+and return the top results with their titles, links, summaries.
+Since the tool is stateless (i.e. does not need
+access to agent state), it can be enabled for any agent, without having to define a
+special method inside the agent: `agent.enable_message(ExaSearchTool)`
+
+NOTE: To use this tool, you need to:
+
+* set the EXA_API_KEY environment variables in
+your `.env` file, e.g. `EXA_API_KEY=your_api_key_here`
+(Note as of 28 Jan 2023, Metaphor renamed to Exa, so you can also use
+`EXA_API_KEY=your_api_key_here`)
+
+* install langroid with the `exa-py` extra, e.g.
+`pip install langroid[exa]` or `uv pip install langroid[exa]`
+or `poetry add langroid[exa]`  or `uv add langroid[exa]`
+(it installs the `exa_py` package from pypi).
+
+For more information, please refer to the official docs:
+https://exa.ai/
+"""
+
+from typing import List, Tuple
+
+from langroid.agent.tool_message import ToolMessage
+from langroid.parsing.web_search import exa_search
+
+
+class ExaSearchTool(ToolMessage):
+    request: str = "exa_search"
+    purpose: str = """
+            To search the web and return up to <num_results> 
+            links relevant to the given <query>. When using this tool,
+            ONLY show the required JSON, DO NOT SAY ANYTHING ELSE.
+            Wait for the results of the web search, and then use them to
+            compose your response.
+            """
+    query: str
+    num_results: int
+
+    def handle(self) -> str:
+        """
+        Conducts a search using the exa API based on the provided query
+        and number of results by triggering a exa_search.
+
+        Returns:
+            str: A formatted string containing the titles, links, and
+                summaries of each search result, separated by two newlines.
+        """
+
+        search_results = exa_search(self.query, self.num_results)
+        # return Title, Link, Summary of each result, separated by two newlines
+        results_str = "\n\n".join(str(result) for result in search_results)
+        return f"""
+        BELOW ARE THE RESULTS FROM THE WEB SEARCH. USE THESE TO COMPOSE YOUR RESPONSE:
+        {results_str}
+        """
+
+    @classmethod
+    def examples(cls) -> List["ToolMessage" | Tuple[str, "ToolMessage"]]:
+        return [
+            cls(
+                query="When was the Llama2 Large Language Model (LLM) released?",
+                num_results=3,
+            ),
+        ]

langroid/agent/tools/tavily_search_tool.py

@@ -0,0 +1,50 @@
+"""
+A tool to trigger a Tavily search for a given query, and return the top results with
+their titles, links, summaries. Since the tool is stateless (i.e. does not need
+access to agent state), it can be enabled for any agent, without having to define a
+special method inside the agent: `agent.enable_message(TavilySearchTool)`
+"""
+
+from typing import List, Tuple
+
+from langroid.agent.tool_message import ToolMessage
+from langroid.parsing.web_search import tavily_search
+
+
+class TavilySearchTool(ToolMessage):
+    request: str = "tavily_search"
+    purpose: str = """
+            To search the web and return up to <num_results> 
+            links relevant to the given <query>. When using this tool,
+            ONLY show the required JSON, DO NOT SAY ANYTHING ELSE.
+            Wait for the results of the web search, and then use them to
+            compose your response.
+            """
+    query: str
+    num_results: int
+
+    def handle(self) -> str:
+        """
+        Conducts a search using Tavily based on the provided query
+        and number of results by triggering a tavily_search.
+
+        Returns:
+            str: A formatted string containing the titles, links, and
+                summaries of each search result, separated by two newlines.
+        """
+        search_results = tavily_search(self.query, self.num_results)
+        # return Title, Link, Summary of each result, separated by two newlines
+        results_str = "\n\n".join(str(result) for result in search_results)
+        return f"""
+        BELOW ARE THE RESULTS FROM THE WEB SEARCH. USE THESE TO COMPOSE YOUR RESPONSE:
+        {results_str}
+        """
+
+    @classmethod
+    def examples(cls) -> List["ToolMessage" | Tuple[str, "ToolMessage"]]:
+        return [
+            cls(
+                query="When was the Llama2 Large Language Model (LLM) released?",
+                num_results=3,
+            ),
+        ]

langroid/parsing/search.py

@@ -118,7 +118,7 @@ def preprocess_text(text: str) -> str:
         str: The preprocessed text.
     """
     # Ensure the NLTK resources are available
-    for resource in ["punkt", "wordnet", "stopwords"]:
+    for resource in ["tokenizers/punkt", "corpora/wordnet", "corpora/stopwords"]:
         download_nltk_resource(resource)
 
     # Lowercase the text

langroid/parsing/utils.py

@@ -28,12 +28,13 @@ def download_nltk_resource(resource: str) -> None:
     try:
         nltk.data.find(resource)
     except LookupError:
-        nltk.download(resource, quiet=True)
+        model = resource.split("/")[-1]
+        nltk.download(model, quiet=True)
 
 
 # Download punkt_tab resource at module import
-download_nltk_resource("punkt_tab")
-download_nltk_resource("gutenberg")
+download_nltk_resource("tokenizers/punkt_tab")
+download_nltk_resource("corpora/gutenberg")
 
 T = TypeVar("T")
 

langroid/parsing/web_search.py

@@ -16,6 +16,8 @@ from duckduckgo_search import DDGS
 from googleapiclient.discovery import Resource, build
 from requests.models import Response
 
+from langroid.exceptions import LangroidImportError
+
 
 class WebSearchResult:
     """
@@ -109,13 +111,7 @@ def metaphor_search(query: str, num_results: int = 5) -> List[WebSearchResult]:
     try:
         from metaphor_python import Metaphor
     except ImportError:
-        raise ImportError(
-            "You are attempting to use the `metaphor_python` library;"
-            "To use it, please install langroid with the `metaphor` extra, e.g. "
-            "`pip install langroid[metaphor]` or `poetry add langroid[metaphor]` "
-            "or `uv add langroid[metaphor]`"
-            "(it installs the `metaphor_python` package from pypi)."
-        )
+        raise LangroidImportError("metaphor-python", "metaphor")
 
     client = Metaphor(api_key=api_key)
 
@@ -130,6 +126,53 @@ def metaphor_search(query: str, num_results: int = 5) -> List[WebSearchResult]:
     ]
 
 
+def exa_search(query: str, num_results: int = 5) -> List[WebSearchResult]:
+    """
+    Method that makes an API call by Exa client that queries
+    the top num_results links that matches the query. Returns a list
+    of WebSearchResult objects.
+
+    Args:
+        query (str): The query body that users wants to make.
+        num_results (int): Number of top matching results that we want
+            to grab
+    """
+
+    load_dotenv()
+
+    api_key = os.getenv("EXA_API_KEY")
+    if not api_key:
+        raise ValueError(
+            """
+            EXA_API_KEY environment variables are not set. 
+            Please set one of them to your API key, and try again.
+            """
+        )
+
+    try:
+        from exa_py import Exa
+    except ImportError:
+        raise LangroidImportError("exa-py", "exa")
+
+    client = Exa(api_key=api_key)
+
+    response = client.search(
+        query=query,
+        num_results=num_results,
+    )
+    raw_results = response.results
+
+    return [
+        WebSearchResult(
+            title=result.title or "",
+            link=result.url,
+            max_content_length=3500,
+            max_summary_length=300,
+        )
+        for result in raw_results
+    ]
+
+
 def duckduckgo_search(query: str, num_results: int = 5) -> List[WebSearchResult]:
     """
     Method that makes an API call by DuckDuckGo client that queries
@@ -154,3 +197,44 @@ def duckduckgo_search(query: str, num_results: int = 5) -> List[WebSearchResult]
         )
         for result in search_results
     ]
+
+
+def tavily_search(query: str, num_results: int = 5) -> List[WebSearchResult]:
+    """
+    Method that makes an API call to Tavily API that queries
+    the top `num_results` links that match the query. Returns a list
+    of WebSearchResult objects.
+
+    Args:
+        query (str): The query body that users wants to make.
+        num_results (int): Number of top matching results that we want
+            to grab
+    """
+
+    load_dotenv()
+
+    api_key = os.getenv("TAVILY_API_KEY")
+    if not api_key:
+        raise ValueError(
+            "TAVILY_API_KEY environment variable is not set. "
+            "Please set it to your API key and try again."
+        )
+
+    try:
+        from tavily import TavilyClient
+    except ImportError:
+        raise LangroidImportError("tavily-python", "tavily")
+
+    client = TavilyClient(api_key=api_key)
+    response = client.search(query=query, max_results=num_results)
+    search_results = response["results"]
+
+    return [
+        WebSearchResult(
+            title=result["title"],
+            link=result["url"],
+            max_content_length=3500,
+            max_summary_length=300,
+        )
+        for result in search_results
+    ]

langroid/vector_store/__init__.py

@@ -23,11 +23,7 @@ try:
     MeiliSearch
     MeiliSearchConfig
     __all__.extend(["meilisearch", "MeiliSearch", "MeiliSearchConfig"])
-except ImportError:
-    pass
 
-
-try:
     from . import lancedb
     from .lancedb import LanceDB, LanceDBConfig
 
@@ -35,10 +31,6 @@ try:
     LanceDB
     LanceDBConfig
     __all__.extend(["lancedb", "LanceDB", "LanceDBConfig"])
-except ImportError:
-    pass
-
-try:
     from . import chromadb
     from .chromadb import ChromaDBConfig, ChromaDB
 
@@ -46,10 +38,15 @@ try:
     ChromaDB
     ChromaDBConfig
     __all__.extend(["chromadb", "ChromaDBConfig", "ChromaDB"])
-except ImportError:
-    pass
 
-try:
+    from . import postgres
+    from .postgres import PostgresDB, PostgresDBConfig
+
+    postgres  # silence linters
+    PostgresDB
+    PostgresDBConfig
+    __all__.extend(["postgres", "PostgresDB", "PostgresDBConfig"])
+
     from . import weaviatedb
     from .weaviatedb import WeaviateDBConfig, WeaviateDB
 
@@ -57,5 +54,13 @@ try:
     WeaviateDB
     WeaviateDBConfig
     __all__.extend(["weaviatedb", "WeaviateDB", "WeaviateDBConfig"])
+
+    from . import pineconedb
+    from .pineconedb import PineconeDB, PineconeDBConfig
+
+    pineconedb
+    PineconeDB
+    PineconeDBConfig
+    __all__.extend(["pineconedb", "PineconeDB", "PineconeDBConfig"])
 except ImportError:
     pass

langroid/vector_store/base.py

@@ -59,6 +59,8 @@ class VectorStore(ABC):
         from langroid.vector_store.lancedb import LanceDB, LanceDBConfig
         from langroid.vector_store.meilisearch import MeiliSearch, MeiliSearchConfig
         from langroid.vector_store.momento import MomentoVI, MomentoVIConfig
+        from langroid.vector_store.pineconedb import PineconeDB, PineconeDBConfig
+        from langroid.vector_store.postgres import PostgresDB, PostgresDBConfig
         from langroid.vector_store.qdrantdb import QdrantDB, QdrantDBConfig
         from langroid.vector_store.weaviatedb import WeaviateDB, WeaviateDBConfig
 
@@ -72,8 +74,12 @@ class VectorStore(ABC):
             return LanceDB(config)
         elif isinstance(config, MeiliSearchConfig):
             return MeiliSearch(config)
+        elif isinstance(config, PostgresDBConfig):
+            return PostgresDB(config)
         elif isinstance(config, WeaviateDBConfig):
             return WeaviateDB(config)
+        elif isinstance(config, PineconeDBConfig):
+            return PineconeDB(config)
 
         else:
             logger.warning(