intuned-runtime 1.2.4__tar.gz → 1.3.0__tar.gz

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: intuned-runtime
-Version: 1.2.4
+Version: 1.3.0
 Summary: Runtime commands for Intuned platform Python scrapers
 License: Elastic-2.0
 License-File: LICENSE

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/intuned_cli/types.py

@@ -1,5 +1,4 @@
 from pydantic import BaseModel
-from pydantic import Field
 from pydantic import RootModel
 
 
@@ -20,14 +19,3 @@ class FileNode(BaseModel):
 
 
 FileSystemTree.model_rebuild()
-
-
-class IntunedJson(BaseModel):
-    model_config = {"populate_by_name": True}
-
-    class _AuthSessions(BaseModel):
-        enabled: bool
-
-    auth_sessions: _AuthSessions = Field(alias="authSessions")
-    project_name: str | None = Field(alias="projectName", default=None)
-    workspace_id: str | None = Field(alias="workspaceId", default=None)

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/intuned_cli/utils/api_helpers.py

@@ -3,8 +3,8 @@ from typing import overload
 
 from anyio import Path
 
-from intuned_cli.types import IntunedJson
 from intuned_cli.utils.error import CLIError
+from runtime.types import IntunedJson
 
 
 @overload

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/intuned_cli/utils/backend.py

@@ -1,7 +1,7 @@
 import os
 
-from intuned_cli.types import IntunedJson
 from intuned_cli.utils.error import CLIError
+from runtime.types import IntunedJson
 
 
 def get_base_url():

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "intuned-runtime"
-version = "1.2.4"
+version = "1.3.0"
 description = "Runtime commands for Intuned platform Python scrapers"
 authors = ["Intuned Developers <engineering@intunedhq.com>"]
 readme = "README.md"

intuned_runtime-1.3.0/runtime/browser/extensions/__init__.py

@@ -0,0 +1,3 @@
+from .helpers import build_extensions_list
+
+__all__ = ["build_extensions_list"]

intuned_runtime-1.3.0/runtime/browser/extensions/helpers.py

@@ -0,0 +1,11 @@
+from .intuned_extension import get_intuned_extension_path
+from .intuned_extension import is_intuned_extension_enabled
+
+
+def build_extensions_list() -> list[str]:
+    extensions_list: list[str] = []
+
+    if is_intuned_extension_enabled():
+        intuned_extension_path = get_intuned_extension_path()
+        extensions_list.append(str(intuned_extension_path))
+    return extensions_list

intuned_runtime-1.3.0/runtime/browser/extensions/intuned_extension.py

@@ -0,0 +1,87 @@
+import logging
+import os
+from pathlib import Path
+from typing import Any
+
+from playwright.async_api import BrowserContext
+
+from runtime.context.context import IntunedContext
+from runtime.env import get_functions_domain
+from runtime.env import get_project_id
+from runtime.env import get_workspace_id
+from runtime.types import CaptchaSolverSettings
+from runtime.utils.config_loader import load_intuned_json
+
+logger = logging.getLogger(__name__)
+
+
+def get_intuned_extension_path() -> Path | None:
+    if "INTUNED_EXTENSION_PATH" not in os.environ:
+        return None
+    intuned_extension_path = Path(os.environ["INTUNED_EXTENSION_PATH"])
+    if not intuned_extension_path.exists():
+        return None
+    return intuned_extension_path
+
+
+def is_intuned_extension_enabled() -> bool:
+    intuned_extension_path = get_intuned_extension_path()
+    if intuned_extension_path is None:
+        return False
+    else:
+        return True
+
+
+async def get_intuned_worker(context: BrowserContext):
+    if not is_intuned_extension_enabled():
+        return None
+
+    for attempt in range(5):
+        for service_worker in context.service_workers:
+            if "intunedWorker.js" in service_worker.url:
+                return service_worker
+        try:
+            if attempt < 4:
+                await context.wait_for_event("serviceworker", timeout=3000)
+        except Exception as e:
+            logger.warning(f"Error accessing service workers (attempt {attempt + 1}): {e}")
+
+    logger.warning("Failed to get intuned worker after 5 attempts")
+    return None
+
+
+async def get_intuned_extension_settings() -> dict[str, Any]:
+    intuned_json = await load_intuned_json()
+    captcha_settings: CaptchaSolverSettings = (
+        intuned_json.captcha_solver
+        if intuned_json and intuned_json.captcha_solver is not None
+        else CaptchaSolverSettings()
+    )
+    context = IntunedContext.current()
+    return {
+        **captcha_settings.model_dump(mode="json"),
+        "workspaceId": get_workspace_id(),
+        "projectId": get_project_id(),
+        "token": context.functions_token,
+        "baseUrl": get_functions_domain(),
+    }
+
+
+async def setup_intuned_extension():
+    if not is_intuned_extension_enabled():
+        return
+    intuned_extension_path = get_intuned_extension_path()
+    if intuned_extension_path is None:
+        logger.warning("Intuned extension path not found, intuned extension might not work properly")
+        return
+
+    settings_path = intuned_extension_path / "intunedSettings.json"
+    settings_data = await get_intuned_extension_settings()
+
+    try:
+        with open(settings_path, "w") as f:
+            import json
+
+            json.dump(settings_data, f)
+    except Exception as e:
+        logger.warning(f"Failed to write intuned settings to {settings_path}: {e}")

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/runtime/browser/launch_chromium.py

@@ -8,6 +8,11 @@ from typing import TYPE_CHECKING
 
 import anyio
 
+from runtime.browser.extensions import build_extensions_list
+from runtime.browser.extensions.intuned_extension import get_intuned_worker
+from runtime.browser.extensions.intuned_extension import is_intuned_extension_enabled
+from runtime.browser.extensions.intuned_extension import setup_intuned_extension
+
 from .helpers import get_local_cdp_address
 from .helpers import get_proxy_env
 from .helpers import wait_on_cdp_address
@@ -80,15 +85,36 @@ async def launch_chromium(
             if cdp_port:
                 extra_args.append(f"--remote-debugging-port={cdp_port}")
 
+            args_to_ignore = [
+                "--disable-extensions",
+                "--disable-component-extensions-with-background-pages",
+                "--disable-background-networking",
+                "--disable-backgrounding-occluded-windows",
+                "--disable-background-timer-throttling",
+            ]
+            if is_intuned_extension_enabled():
+                extensions_list = build_extensions_list()
+                extensions = ",".join(extensions_list)
+                extra_args.append("--disable-extensions-except=" + extensions)
+                extra_args.append(f"--load-extension={extensions}")
+                await setup_intuned_extension()
+
+            if headless:
+                args_to_ignore.append("--headless=old")
+                extra_args.append("--headless=new")
             context = await playwright.chromium.launch_persistent_context(
                 os.fspath(user_preferences_dir),
                 headless=headless,
                 viewport=viewport,
                 proxy=proxy,
+                ignore_default_args=args_to_ignore,
                 user_agent=os.environ.get("USER_AGENT", default_user_agent),
                 args=extra_args,
                 **kwargs,
             )
+            if is_intuned_extension_enabled():
+                # wait for intuned extension to be ready
+                await get_intuned_worker(context)
             if cdp_port:
                 await wait_on_cdp_address(get_local_cdp_address(cdp_port))
 

{intuned_runtime-1.2.4 → intuned_runtime-1.3.0}/runtime/types/__init__.py

@@ -4,6 +4,8 @@ from .run_types import RunAutomationErrorResult
 from .run_types import RunAutomationResult
 from .run_types import RunAutomationSuccessResult
 from .run_types import RunBody
+from .settings_types import CaptchaSolverSettings
+from .settings_types import IntunedJson
 
 __all__ = [
     "Payload",
@@ -12,4 +14,6 @@ __all__ = [
     "RunAutomationErrorResult",
     "RunAutomationSuccessResult",
     "PayloadToAppend",
+    "IntunedJson",
+    "CaptchaSolverSettings",
 ]

intuned_runtime-1.3.0/runtime/types/settings_types.py

@@ -0,0 +1,63 @@
+from typing import List
+
+from pydantic import BaseModel
+from pydantic import Field
+
+
+class CaptchaSettings(BaseModel):
+    enabled: bool = Field(default=False)
+
+
+class CustomCaptchaSettings(CaptchaSettings):
+    model_config = {
+        "populate_by_name": True,
+        "serialize_by_alias": True,
+    }
+
+    image_locators: List[str] = Field(alias="imageLocators", default=[])
+    submit_locators: List[str] = Field(alias="submitLocators", default=[])
+    input_locators: List[str] = Field(alias="inputLocators", default=[])
+
+
+class TextCaptchaSettings(CaptchaSettings):
+    model_config = {
+        "populate_by_name": True,
+        "serialize_by_alias": True,
+    }
+    label_locators: List[str] = Field(alias="labelLocators", default=[])
+    submit_locators: List[str] = Field(alias="submitLocators", default=[])
+    input_locators: List[str] = Field(alias="inputLocators", default=[])
+
+
+class CaptchaSolverSettings(BaseModel):
+    model_config = {
+        "populate_by_name": True,
+        "serialize_by_alias": True,
+    }
+
+    enabled: bool = Field(default=False)
+    cloudflare: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    google_recaptcha_v2: CaptchaSettings = Field(alias="googleRecaptchaV2", default_factory=CaptchaSettings)
+    google_recaptcha_v3: CaptchaSettings = Field(alias="googleRecaptchaV3", default_factory=CaptchaSettings)
+    awscaptcha: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    hcaptcha: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    funcaptcha: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    geetest: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    lemin: CaptchaSettings = Field(default_factory=CaptchaSettings)
+    custom_captcha: CustomCaptchaSettings = Field(alias="customCaptcha", default_factory=CustomCaptchaSettings)
+    text: TextCaptchaSettings = Field(default_factory=TextCaptchaSettings)
+    settings: dict[str, int | bool] = Field(
+        default={"autoSolve": True, "solveDelay": 2000, "maxRetries": 3, "timeout": 30000}
+    )
+
+
+class IntunedJson(BaseModel):
+    model_config = {"populate_by_name": True}
+
+    class _AuthSessions(BaseModel):
+        enabled: bool
+
+    auth_sessions: _AuthSessions = Field(alias="authSessions")
+    project_name: str | None = Field(alias="projectName", default=None)
+    workspace_id: str | None = Field(alias="workspaceId", default=None)
+    captcha_solver: CaptchaSolverSettings | None = Field(alias="captchaSolver", default=None)

intuned_runtime-1.3.0/runtime/utils/__init__.py

@@ -0,0 +1,3 @@
+from .config_loader import load_intuned_json
+
+__all__ = ["load_intuned_json"]

intuned_runtime-1.3.0/runtime/utils/config_loader.py

@@ -0,0 +1,17 @@
+from anyio import Path
+
+from runtime.types import IntunedJson
+
+
+async def load_intuned_json() -> IntunedJson | None:
+    """
+    Load the Intuned.json configuration file.
+    Returns None if file doesn't exist or fails to parse.
+    """
+    intuned_json_path = Path("Intuned.json")
+    if not await intuned_json_path.exists():
+        return None
+    try:
+        return IntunedJson.model_validate_json(await intuned_json_path.read_text())
+    except Exception:
+        return None