Spaces:
Running
Running
Update app/core/service/playwright/playwright_context.py
Browse files
app/core/service/playwright/playwright_context.py
CHANGED
@@ -14,19 +14,22 @@ from .models import GetContentModel, PageModel, ScreenshotModel # noqa: TCH001
|
|
14 |
|
15 |
|
16 |
class AsyncMixin:
|
17 |
-
"""
|
18 |
|
19 |
-
async def
|
20 |
-
|
21 |
|
22 |
def __await__(self) -> AsyncIterator[Awaitable]:
|
23 |
-
|
|
|
24 |
|
25 |
|
26 |
class PlaywrightInstance(AsyncMixin):
|
27 |
"""This class is designed to keep playwright browser instance open for reusability and scalability handling api requests.""" # noqa: E501
|
28 |
|
29 |
-
HEADERS: str =
|
|
|
|
|
30 |
|
31 |
FIREFOX_USER_PREFS: ClassVar[dict[str, int | str]] = {
|
32 |
"extensions.enabledScopes": 1,
|
@@ -37,11 +40,11 @@ class PlaywrightInstance(AsyncMixin):
|
|
37 |
}
|
38 |
|
39 |
def __init__(self) -> None:
|
40 |
-
|
41 |
-
self.
|
|
|
42 |
|
43 |
-
async def
|
44 |
-
"""starts playwright and browser instance."""
|
45 |
if not self.playwright:
|
46 |
self.playwright = await async_playwright().start()
|
47 |
self.browser = await self.playwright.firefox.launch(
|
@@ -54,7 +57,7 @@ class PlaywrightInstance(AsyncMixin):
|
|
54 |
browser: Browser,
|
55 |
screenshot_model: GetContentModel,
|
56 |
page_model: PageModel,
|
57 |
-
) -> tuple[BrowserContext, Page]:
|
58 |
"""create a brwoser or new browser context page.
|
59 |
|
60 |
Parameters:
|
@@ -73,11 +76,11 @@ class PlaywrightInstance(AsyncMixin):
|
|
73 |
"java_script_enabled": page_model.java_script_enabled,
|
74 |
"no_viewport": page_model.no_viewport,
|
75 |
"proxy": page_model.proxy.model_dump() if page_model.proxy else None,
|
76 |
-
"viewport": page_model.viewport
|
77 |
}
|
78 |
|
79 |
if not screenshot_model.new_browser:
|
80 |
-
return None, await
|
81 |
|
82 |
new_context = await browser.new_context(**params)
|
83 |
return new_context, await new_context.new_page()
|
@@ -87,24 +90,39 @@ class PlaywrightInstance(AsyncMixin):
|
|
87 |
screenshot_model: ScreenshotModel,
|
88 |
page_model: PageModel,
|
89 |
) -> bytes:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
90 |
context, page = await self.new_context_page(
|
91 |
-
screenshot_model=screenshot_model,
|
|
|
|
|
92 |
)
|
93 |
|
94 |
await page.goto(str(screenshot_model.url))
|
95 |
-
|
96 |
await page.wait_for_timeout(screenshot_model.ms_delay)
|
97 |
-
|
98 |
screenshot_locator = (
|
99 |
page.locator(screenshot_model.query_selector)
|
100 |
if screenshot_model.query_selector
|
101 |
-
else None
|
102 |
)
|
103 |
|
104 |
if screenshot_locator:
|
105 |
-
screenshot_data: bytes = await screenshot_locator.screenshot(
|
|
|
|
|
106 |
else:
|
107 |
-
screenshot_data: bytes = await page.screenshot(
|
|
|
|
|
108 |
|
109 |
await page.close()
|
110 |
|
@@ -114,7 +132,7 @@ class PlaywrightInstance(AsyncMixin):
|
|
114 |
return screenshot_data
|
115 |
|
116 |
async def close_instance(self) -> None:
|
117 |
-
"""
|
118 |
if self.playwright:
|
119 |
await self.browser.close()
|
120 |
await self.playwright.stop()
|
|
|
14 |
|
15 |
|
16 |
class AsyncMixin:
|
17 |
+
"""Experimental: making awaitable class."""
|
18 |
|
19 |
+
async def __ainit__(self) -> None:
|
20 |
+
"""Initialize the class."""
|
21 |
|
22 |
def __await__(self) -> AsyncIterator[Awaitable]:
|
23 |
+
"""Make the class awaitable."""
|
24 |
+
return self.__ainit__().__await__()
|
25 |
|
26 |
|
27 |
class PlaywrightInstance(AsyncMixin):
|
28 |
"""This class is designed to keep playwright browser instance open for reusability and scalability handling api requests.""" # noqa: E501
|
29 |
|
30 |
+
HEADERS: ClassVar[str] = (
|
31 |
+
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36" # noqa: E501
|
32 |
+
)
|
33 |
|
34 |
FIREFOX_USER_PREFS: ClassVar[dict[str, int | str]] = {
|
35 |
"extensions.enabledScopes": 1,
|
|
|
40 |
}
|
41 |
|
42 |
def __init__(self) -> None:
|
43 |
+
"""Initialize the class."""
|
44 |
+
self.playwright: async_playwright | None = None
|
45 |
+
self.browser: Browser | None = None
|
46 |
|
47 |
+
async def __ainit__(self) -> PlaywrightInstance:
|
|
|
48 |
if not self.playwright:
|
49 |
self.playwright = await async_playwright().start()
|
50 |
self.browser = await self.playwright.firefox.launch(
|
|
|
57 |
browser: Browser,
|
58 |
screenshot_model: GetContentModel,
|
59 |
page_model: PageModel,
|
60 |
+
) -> tuple[BrowserContext | None, Page]:
|
61 |
"""create a brwoser or new browser context page.
|
62 |
|
63 |
Parameters:
|
|
|
76 |
"java_script_enabled": page_model.java_script_enabled,
|
77 |
"no_viewport": page_model.no_viewport,
|
78 |
"proxy": page_model.proxy.model_dump() if page_model.proxy else None,
|
79 |
+
"viewport": page_model.viewport,
|
80 |
}
|
81 |
|
82 |
if not screenshot_model.new_browser:
|
83 |
+
return None, await browser.new_page(**params)
|
84 |
|
85 |
new_context = await browser.new_context(**params)
|
86 |
return new_context, await new_context.new_page()
|
|
|
90 |
screenshot_model: ScreenshotModel,
|
91 |
page_model: PageModel,
|
92 |
) -> bytes:
|
93 |
+
"""Take a screenshot of a webpage url.
|
94 |
+
|
95 |
+
Parameters:
|
96 |
+
screenshot_model (ScreenshotModel):
|
97 |
+
A pydantic BaseModel instance containing the configuration for the screenshot.
|
98 |
+
page_model (PageModel):
|
99 |
+
A pydantic BaseModel instance containing the configuration for the page.
|
100 |
+
|
101 |
+
Returns:
|
102 |
+
bytes: The screenshot data in bytes.
|
103 |
+
"""
|
104 |
context, page = await self.new_context_page(
|
105 |
+
screenshot_model=screenshot_model,
|
106 |
+
browser=self.browser,
|
107 |
+
page_model=page_model,
|
108 |
)
|
109 |
|
110 |
await page.goto(str(screenshot_model.url))
|
|
|
111 |
await page.wait_for_timeout(screenshot_model.ms_delay)
|
|
|
112 |
screenshot_locator = (
|
113 |
page.locator(screenshot_model.query_selector)
|
114 |
if screenshot_model.query_selector
|
115 |
+
else None,
|
116 |
)
|
117 |
|
118 |
if screenshot_locator:
|
119 |
+
screenshot_data: bytes = await screenshot_locator.screenshot(
|
120 |
+
type=screenshot_model.image_type
|
121 |
+
)
|
122 |
else:
|
123 |
+
screenshot_data: bytes = await page.screenshot(
|
124 |
+
full_page=screenshot_model.full_page, type=screenshot_model.image_type
|
125 |
+
)
|
126 |
|
127 |
await page.close()
|
128 |
|
|
|
132 |
return screenshot_data
|
133 |
|
134 |
async def close_instance(self) -> None:
|
135 |
+
"""For manual closing of playwright if needed."""
|
136 |
if self.playwright:
|
137 |
await self.browser.close()
|
138 |
await self.playwright.stop()
|