goyimnose/Medios-Macina
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

jkj

Browse Source
This commit is contained in:
nose
2025-12-16 23:23:43 -08:00
parent 9873280f0e
commit 86918f2ae2
46 changed files with 2277 additions and 1347 deletions
Download Patch File Download Diff File Expand all files Collapse all files

256
cmdlet/screen_shot.py
View File

@@ -20,7 +20,16 @@ from urllib.parse import urlsplit, quote, urljoin
from SYS.logger import log, debug
from API.HTTP import HTTPClient
from SYS.utils import ensure_directory, unique_path, unique_preserve_order
from ._shared import Cmdlet, CmdletArg, SharedArgs, create_pipe_object_result, normalize_result_input, should_show_help, get_field
from . import _shared as sh
Cmdlet = sh.Cmdlet
CmdletArg = sh.CmdletArg
SharedArgs = sh.SharedArgs
create_pipe_object_result = sh.create_pipe_object_result
normalize_result_input = sh.normalize_result_input
should_show_help = sh.should_show_help
get_field = sh.get_field
parse_cmdlet_args = sh.parse_cmdlet_args
import pipeline as pipeline_context
# ============================================================================
@@ -33,20 +42,7 @@ import pipeline as pipeline_context
# Playwright & Screenshot Dependencies
# ============================================================================
try:
from playwright.sync_api import (
TimeoutError as PlaywrightTimeoutError,
sync_playwright,
)
HAS_PLAYWRIGHT = True
except Exception:
HAS_PLAYWRIGHT = False
PlaywrightTimeoutError = TimeoutError # type: ignore
def sync_playwright(*_args: Any, **_kwargs: Any) -> Any: # type: ignore
raise RuntimeError(
"playwright is required for screenshot capture; install with: pip install playwright; then: playwright install"
)
from tool.playwright import HAS_PLAYWRIGHT, PlaywrightTimeoutError, PlaywrightTool
try:
from config import resolve_output_dir
@@ -128,6 +124,7 @@ class ScreenshotOptions:
prefer_platform_target: bool = False
target_selectors: Optional[Sequence[str]] = None
selector_timeout_ms: int = 10_000
playwright_tool: Optional[PlaywrightTool] = None
@dataclass(slots=True)
@@ -324,142 +321,119 @@ def _prepare_output_path(options: ScreenshotOptions) -> Path:
def _capture(options: ScreenshotOptions, destination: Path, warnings: List[str]) -> None:
"""Capture screenshot using Playwright."""
debug(f"[_capture] Starting capture for {options.url} -> {destination}")
playwright = None
browser = None
context = None
try:
debug("Starting Playwright...", flush=True)
playwright = sync_playwright().start()
log("Launching Chromium browser...", flush=True)
tool = options.playwright_tool or PlaywrightTool({})
tool.debug_dump()
log("Launching browser...", flush=True)
format_name = _normalise_format(options.output_format)
headless = options.headless or format_name == "pdf"
debug(f"[_capture] Format: {format_name}, Headless: {headless}")
if format_name == "pdf" and not options.headless:
warnings.append("pdf output requires headless Chromium; overriding headless mode")
browser = playwright.chromium.launch(
headless=headless,
args=["--disable-blink-features=AutomationControlled"],
)
log("Creating browser context...", flush=True)
context = browser.new_context(
user_agent=USER_AGENT,
viewport=DEFAULT_VIEWPORT,
ignore_https_errors=True,
)
page = context.new_page()
log(f"Navigating to {options.url}...", flush=True)
try:
page.goto(options.url, timeout=90_000, wait_until="domcontentloaded")
log("Page loaded successfully", flush=True)
except PlaywrightTimeoutError:
warnings.append("navigation timeout; capturing current page state")
log("Navigation timeout; proceeding with current state", flush=True)
# Skip article lookup by default (wait_for_article defaults to False)
if options.wait_for_article:
with tool.open_page(headless=headless) as page:
log(f"Navigating to {options.url}...", flush=True)
try:
log("Waiting for article element...", flush=True)
page.wait_for_selector("article", timeout=10_000)
log("Article element found", flush=True)
tool.goto(page, options.url)
log("Page loaded successfully", flush=True)
except PlaywrightTimeoutError:
warnings.append("<article> selector not found; capturing fallback")
log("Article element not found; using fallback", flush=True)
if options.wait_after_load > 0:
log(f"Waiting {options.wait_after_load}s for page stabilization...", flush=True)
time.sleep(min(10.0, max(0.0, options.wait_after_load)))
if options.replace_video_posters:
log("Replacing video elements with posters...", flush=True)
page.evaluate(
"""
document.querySelectorAll('video').forEach(v => {
if (v.poster) {
const img = document.createElement('img');
img.src = v.poster;
img.style.maxWidth = '100%';
img.style.borderRadius = '12px';
v.replaceWith(img);
}
});
"""
)
# Attempt platform-specific target capture if requested (and not PDF)
element_captured = False
if options.prefer_platform_target and format_name != "pdf":
log("Attempting platform-specific content capture...", flush=True)
try:
_platform_preprocess(options.url, page, warnings)
except Exception as e:
debug(f"[_capture] Platform preprocess failed: {e}")
pass
selectors = list(options.target_selectors or [])
if not selectors:
selectors = _selectors_for_url(options.url)
warnings.append("navigation timeout; capturing current page state")
log("Navigation timeout; proceeding with current state", flush=True)
debug(f"[_capture] Trying selectors: {selectors}")
for sel in selectors:
# Skip article lookup by default (wait_for_article defaults to False)
if options.wait_for_article:
try:
log(f"Trying selector: {sel}", flush=True)
el = page.wait_for_selector(sel, timeout=max(0, int(options.selector_timeout_ms)))
log("Waiting for article element...", flush=True)
page.wait_for_selector("article", timeout=10_000)
log("Article element found", flush=True)
except PlaywrightTimeoutError:
log(f"Selector not found: {sel}", flush=True)
continue
warnings.append("<article> selector not found; capturing fallback")
log("Article element not found; using fallback", flush=True)
if options.wait_after_load > 0:
log(f"Waiting {options.wait_after_load}s for page stabilization...", flush=True)
time.sleep(min(10.0, max(0.0, options.wait_after_load)))
if options.replace_video_posters:
log("Replacing video elements with posters...", flush=True)
page.evaluate(
"""
document.querySelectorAll('video').forEach(v => {
if (v.poster) {
const img = document.createElement('img');
img.src = v.poster;
img.style.maxWidth = '100%';
img.style.borderRadius = '12px';
v.replaceWith(img);
}
});
"""
)
# Attempt platform-specific target capture if requested (and not PDF)
element_captured = False
if options.prefer_platform_target and format_name != "pdf":
log("Attempting platform-specific content capture...", flush=True)
try:
if el is not None:
log(f"Found element with selector: {sel}", flush=True)
try:
el.scroll_into_view_if_needed(timeout=1000)
except Exception:
pass
log(f"Capturing element to {destination}...", flush=True)
el.screenshot(path=str(destination), type=("jpeg" if format_name == "jpeg" else None))
element_captured = True
log("Element captured successfully", flush=True)
break
except Exception as exc:
warnings.append(f"element capture failed for '{sel}': {exc}")
log(f"Failed to capture element: {exc}", flush=True)
# Fallback to default capture paths
if element_captured:
pass
elif format_name == "pdf":
log("Generating PDF...", flush=True)
page.emulate_media(media="print")
page.pdf(path=str(destination), print_background=True)
log(f"PDF saved to {destination}", flush=True)
else:
log(f"Capturing full page to {destination}...", flush=True)
screenshot_kwargs: Dict[str, Any] = {"path": str(destination)}
if format_name == "jpeg":
screenshot_kwargs["type"] = "jpeg"
screenshot_kwargs["quality"] = 90
if options.full_page:
page.screenshot(full_page=True, **screenshot_kwargs)
_platform_preprocess(options.url, page, warnings)
except Exception as e:
debug(f"[_capture] Platform preprocess failed: {e}")
pass
selectors = list(options.target_selectors or [])
if not selectors:
selectors = _selectors_for_url(options.url)
debug(f"[_capture] Trying selectors: {selectors}")
for sel in selectors:
try:
log(f"Trying selector: {sel}", flush=True)
el = page.wait_for_selector(sel, timeout=max(0, int(options.selector_timeout_ms)))
except PlaywrightTimeoutError:
log(f"Selector not found: {sel}", flush=True)
continue
try:
if el is not None:
log(f"Found element with selector: {sel}", flush=True)
try:
el.scroll_into_view_if_needed(timeout=1000)
except Exception:
pass
log(f"Capturing element to {destination}...", flush=True)
el.screenshot(path=str(destination), type=("jpeg" if format_name == "jpeg" else None))
element_captured = True
log("Element captured successfully", flush=True)
break
except Exception as exc:
warnings.append(f"element capture failed for '{sel}': {exc}")
log(f"Failed to capture element: {exc}", flush=True)
# Fallback to default capture paths
if element_captured:
pass
elif format_name == "pdf":
log("Generating PDF...", flush=True)
page.emulate_media(media="print")
page.pdf(path=str(destination), print_background=True)
log(f"PDF saved to {destination}", flush=True)
else:
article = page.query_selector("article")
if article is not None:
article_kwargs = dict(screenshot_kwargs)
article_kwargs.pop("full_page", None)
article.screenshot(**article_kwargs)
log(f"Capturing full page to {destination}...", flush=True)
screenshot_kwargs: Dict[str, Any] = {"path": str(destination)}
if format_name == "jpeg":
screenshot_kwargs["type"] = "jpeg"
screenshot_kwargs["quality"] = 90
if options.full_page:
page.screenshot(full_page=True, **screenshot_kwargs)
else:
page.screenshot(**screenshot_kwargs)
log(f"Screenshot saved to {destination}", flush=True)
article = page.query_selector("article")
if article is not None:
article_kwargs = dict(screenshot_kwargs)
article_kwargs.pop("full_page", None)
article.screenshot(**article_kwargs)
else:
page.screenshot(**screenshot_kwargs)
log(f"Screenshot saved to {destination}", flush=True)
except Exception as exc:
debug(f"[_capture] Exception: {exc}")
raise ScreenshotError(f"Failed to capture screenshot: {exc}") from exc
finally:
log("Cleaning up browser resources...", flush=True)
with contextlib.suppress(Exception):
if context is not None:
context.close()
with contextlib.suppress(Exception):
if browser is not None:
browser.close()
with contextlib.suppress(Exception):
if playwright is not None:
playwright.stop()
log("Cleanup complete", flush=True)
def _capture_screenshot(options: ScreenshotOptions) -> ScreenshotResult:
@@ -511,8 +485,6 @@ def _run(result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
Screenshots are created using Playwright and marked as temporary
so they can be cleaned up later with the cleanup cmdlet.
"""
from ._shared import parse_cmdlet_args
debug(f"[_run] screen-shot invoked with args: {args}")
# Help check
@@ -534,6 +506,19 @@ def _run(result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
parsed = parse_cmdlet_args(args, CMDLET)
format_value = parsed.get("format")
if not format_value:
# Default format can be set via config.conf tool block:
# [tool=playwright]
# format="pdf"
try:
tool_cfg = config.get("tool", {}) if isinstance(config, dict) else {}
pw_cfg = tool_cfg.get("playwright") if isinstance(tool_cfg, dict) else None
if isinstance(pw_cfg, dict):
format_value = pw_cfg.get("format")
except Exception:
pass
if not format_value:
format_value = "png"
storage_value = parsed.get("storage")
selector_arg = parsed.get("selector")
selectors = [selector_arg] if selector_arg else []
@@ -669,6 +654,7 @@ def _run(result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
prefer_platform_target=False,
wait_for_article=False,
full_page=True,
playwright_tool=PlaywrightTool(config),
)
screenshot_result = _capture_screenshot(options)