CG-12194: Add Web Browsing Tool similar to Linear ViewIssue Tool

src/codegen/extensions/langchain/tools.py

@@ -25,6 +25,12 @@
 from codegen.extensions.tools.search import search
 from codegen.extensions.tools.semantic_edit import semantic_edit
 from codegen.extensions.tools.semantic_search import semantic_search
+from codegen.extensions.web.web import (
+    web_browse_page_tool,
+    web_extract_images_tool,
+    web_search_tool,
+)
+from codegen.extensions.web.web_client import WebClient
 from codegen.sdk.core.codebase import Codebase
 
 from ..tools import (
@@ -43,6 +49,7 @@
 )
 from ..tools.relace_edit_prompts import RELACE_EDIT_PROMPT
 from ..tools.semantic_edit_prompts import FILE_EDIT_PROMPT
+from .web_tools import WebBrowsePageTool, WebExtractImagesTool, WebSearchTool
 
 
 class ViewFileInput(BaseModel):
@@ -823,6 +830,59 @@ def _run(self, content: str) -> str:
         return "✅ Message sent successfully"
 
 
+########################################################################################################################
+# WEB
+########################################################################################################################
+
+
+class WebBrowsePageTool(BaseTool):
+    """Tool for browsing a web page."""
+
+    name: ClassVar[str] = "web_browse_page"
+    description: ClassVar[str] = "Browse a web page and extract relevant information"
+    args_schema: ClassVar[type[BaseModel]] = WebBrowsePageTool.args_schema
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, url: str) -> str:
+        result = web_browse_page_tool(self.client, url)
+        return result.render()
+
+
+class WebSearchTool(BaseTool):
+    """Tool for searching the web."""
+
+    name: ClassVar[str] = "web_search"
+    description: ClassVar[str] = "Search the web for information"
+    args_schema: ClassVar[type[BaseModel]] = WebSearchTool.args_schema
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, query: str) -> str:
+        result = web_search_tool(self.client, query)
+        return result.render()
+
+
+class WebExtractImagesTool(BaseTool):
+    """Tool for extracting images from a web page."""
+
+    name: ClassVar[str] = "web_extract_images"
+    description: ClassVar[str] = "Extract images from a web page"
+    args_schema: ClassVar[type[BaseModel]] = WebExtractImagesTool.args_schema
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, url: str) -> str:
+        result = web_extract_images_tool(self.client, url)
+        return result.render()
+
+
 ########################################################################################################################
 # EXPORT
 ########################################################################################################################
@@ -868,6 +928,10 @@ def get_workspace_tools(codebase: Codebase) -> list["BaseTool"]:
         LinearSearchIssuesTool(codebase),
         LinearCreateIssueTool(codebase),
         LinearGetTeamsTool(codebase),
+        # Web
+        WebBrowsePageTool(WebClient()),
+        WebSearchTool(WebClient()),
+        WebExtractImagesTool(WebClient()),
     ]
 
 

src/codegen/extensions/langchain/web_tools.py

@@ -0,0 +1,81 @@
+"""LangChain tools for web browsing."""
+
+from typing import ClassVar
+
+from langchain_core.tools.base import BaseTool
+from pydantic import BaseModel, Field
+
+from codegen.extensions.web.web import (
+    web_browse_page_tool,
+    web_extract_images_tool,
+    web_search_tool,
+)
+from codegen.extensions.web.web_client import WebClient
+
+
+class WebBrowsePageInput(BaseModel):
+    """Input for browsing a web page."""
+
+    url: str = Field(..., description="URL of the web page to browse")
+
+
+class WebBrowsePageTool(BaseTool):
+    """Tool for browsing web pages."""
+
+    name: ClassVar[str] = "web_browse_page"
+    description: ClassVar[str] = "Browse a web page and extract its content"
+    args_schema: ClassVar[type[BaseModel]] = WebBrowsePageInput
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, url: str) -> str:
+        result = web_browse_page_tool(self.client, url)
+        return result.render()
+
+
+class WebSearchInput(BaseModel):
+    """Input for web search."""
+
+    query: str = Field(..., description="Search query string")
+    num_results: int = Field(default=10, description="Maximum number of results to return")
+
+
+class WebSearchTool(BaseTool):
+    """Tool for searching the web."""
+
+    name: ClassVar[str] = "web_search"
+    description: ClassVar[str] = "Search the web using a search engine"
+    args_schema: ClassVar[type[BaseModel]] = WebSearchInput
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, query: str, num_results: int = 10) -> str:
+        result = web_search_tool(self.client, query, num_results)
+        return result.render()
+
+
+class WebExtractImagesInput(BaseModel):
+    """Input for extracting images from a web page."""
+
+    url: str = Field(..., description="URL of the web page")
+    max_images: int = Field(default=20, description="Maximum number of images to extract")
+
+
+class WebExtractImagesTool(BaseTool):
+    """Tool for extracting images from web pages."""
+
+    name: ClassVar[str] = "web_extract_images"
+    description: ClassVar[str] = "Extract images from a web page"
+    args_schema: ClassVar[type[BaseModel]] = WebExtractImagesInput
+    client: WebClient = Field(exclude=True)
+
+    def __init__(self, client: WebClient) -> None:
+        super().__init__(client=client)
+
+    def _run(self, url: str, max_images: int = 20) -> str:
+        result = web_extract_images_tool(self.client, url, max_images)
+        return result.render()

src/codegen/extensions/web/__init__.py

@@ -0,0 +1,15 @@
+"""Web browsing functionality."""
+
+from codegen.extensions.web.web import (
+    web_browse_page_tool,
+    web_extract_images_tool,
+    web_search_tool,
+)
+from codegen.extensions.web.web_client import WebClient
+
+__all__ = [
+    "WebClient",
+    "web_browse_page_tool",
+    "web_extract_images_tool",
+    "web_search_tool",
+]

src/codegen/extensions/web/types.py

@@ -0,0 +1,29 @@
+"""Types for web browsing functionality."""
+
+from pydantic import BaseModel, Field
+
+
+class WebPage(BaseModel):
+    """Represents a web page."""
+
+    url: str = Field(description="URL of the web page")
+    title: str = Field(description="Title of the web page")
+    content: str = Field(description="Main content of the web page")
+    status_code: int = Field(description="HTTP status code of the response")
+
+
+class WebSearchResult(BaseModel):
+    """Represents a single search result."""
+
+    title: str = Field(description="Title of the search result")
+    url: str = Field(description="URL of the search result")
+    snippet: str = Field(description="Snippet or description of the search result")
+
+
+class WebImage(BaseModel):
+    """Represents an image from a web page."""
+
+    url: str = Field(description="URL of the image")
+    alt_text: str | None = Field(None, description="Alternative text for the image")
+    width: int | None = Field(None, description="Width of the image in pixels")
+    height: int | None = Field(None, description="Height of the image in pixels")