mirror of
https://github.com/browser-use/browser-use
synced 2026-05-06 17:52:15 +02:00
539274a7d4
Replaces pixel-based scrolling with page unit determination, allowing for more intuitive scroll actions by specifying the number of pages. Adjusts related documentation, examples, and tests to reflect this change for improved code consistency and user experience.
102 lines
2.7 KiB
Python
102 lines
2.7 KiB
Python
from typing import Generic, TypeVar
|
|
|
|
from pydantic import BaseModel, ConfigDict, Field
|
|
|
|
|
|
# Action Input Models
|
|
class SearchGoogleAction(BaseModel):
|
|
query: str
|
|
|
|
|
|
class GoToUrlAction(BaseModel):
|
|
url: str
|
|
new_tab: bool # True to open in new tab, False to navigate in current tab
|
|
|
|
|
|
class ClickElementAction(BaseModel):
|
|
index: int
|
|
|
|
|
|
class InputTextAction(BaseModel):
|
|
index: int
|
|
text: str
|
|
|
|
|
|
class DoneAction(BaseModel):
|
|
text: str
|
|
success: bool
|
|
files_to_display: list[str] | None = []
|
|
|
|
|
|
T = TypeVar('T', bound=BaseModel)
|
|
|
|
|
|
class StructuredOutputAction(BaseModel, Generic[T]):
|
|
success: bool = True
|
|
data: T
|
|
|
|
|
|
class SwitchTabAction(BaseModel):
|
|
page_id: int
|
|
|
|
|
|
class CloseTabAction(BaseModel):
|
|
page_id: int
|
|
|
|
|
|
class ScrollAction(BaseModel):
|
|
down: bool # True to scroll down, False to scroll up
|
|
num_pages: float | None = (
|
|
None # Number of pages to scroll (0.5 = half page, 1.0 = one page, etc.). If None, defaults to one page
|
|
)
|
|
|
|
|
|
class SendKeysAction(BaseModel):
|
|
keys: str
|
|
|
|
|
|
class UploadFileAction(BaseModel):
|
|
index: int
|
|
path: str
|
|
|
|
|
|
class ExtractPageContentAction(BaseModel):
|
|
value: str
|
|
|
|
|
|
class NoParamsAction(BaseModel):
|
|
"""
|
|
Accepts absolutely anything in the incoming data
|
|
and discards it, so the final parsed model is empty.
|
|
"""
|
|
|
|
model_config = ConfigDict(extra='ignore')
|
|
# No fields defined - all inputs are ignored automatically
|
|
|
|
|
|
class Position(BaseModel):
|
|
x: int
|
|
y: int
|
|
|
|
|
|
class DragDropAction(BaseModel):
|
|
# Element-based approach
|
|
element_source: str | None = Field(None, description='CSS selector or XPath of the element to drag from')
|
|
element_target: str | None = Field(None, description='CSS selector or XPath of the element to drop onto')
|
|
element_source_offset: Position | None = Field(
|
|
None, description='Precise position within the source element to start drag (in pixels from top-left corner)'
|
|
)
|
|
element_target_offset: Position | None = Field(
|
|
None, description='Precise position within the target element to drop (in pixels from top-left corner)'
|
|
)
|
|
|
|
# Coordinate-based approach (used if selectors not provided)
|
|
coord_source_x: int | None = Field(None, description='Absolute X coordinate on page to start drag from (in pixels)')
|
|
coord_source_y: int | None = Field(None, description='Absolute Y coordinate on page to start drag from (in pixels)')
|
|
coord_target_x: int | None = Field(None, description='Absolute X coordinate on page to drop at (in pixels)')
|
|
coord_target_y: int | None = Field(None, description='Absolute Y coordinate on page to drop at (in pixels)')
|
|
|
|
# Common options
|
|
steps: int | None = Field(10, description='Number of intermediate points for smoother movement (5-20 recommended)')
|
|
delay_ms: int | None = Field(5, description='Delay in milliseconds between steps (0 for fastest, 10-20 for more natural)')
|