html-reader-llm 0.0.1__py3-none-any.whl

html_reader_llm/__init__.py

@@ -0,0 +1,39 @@
+"""html-reader-llm — HTML simplification and intelligent extraction for LLM.
+
+Quick start (pip install):
+    pip install selectolax
+    from html_reader_llm import simplify_html
+    result = simplify_html(html_string)
+
+Copy usage (no pip):
+    Copy settings.py and simplify.py into your project,
+    change `from html_reader_llm import settings` to `from . import settings` (or
+    adjust the import path), then use:
+    from simplify import simplify_html
+    result = simplify_html(html_string)
+"""
+
+from html_reader_llm.browser import (
+    BrowserNotFoundError,
+    detect_browser_paths,
+    get_browser_path,
+    validate_browser_path,
+)
+from html_reader_llm.llm import AnalysisResult, analyze_page
+from html_reader_llm.render_detect import RenderDetectResult, detect_render
+from html_reader_llm.selector_validator import SelectorValidator, validate_url_selectors
+from html_reader_llm.simplify import simplify_html
+
+__all__ = [
+    "AnalysisResult",
+    "BrowserNotFoundError",
+    "RenderDetectResult",
+    "SelectorValidator",
+    "analyze_page",
+    "detect_browser_paths",
+    "detect_render",
+    "get_browser_path",
+    "simplify_html",
+    "validate_browser_path",
+    "validate_url_selectors",
+]

html_reader_llm/browser.py

@@ -0,0 +1,161 @@
+"""Chrome/Edge browser path detection — platform-aware discovery and validation.
+
+Three usage modes:
+- detect_browser_paths() → list all found paths
+- get_browser_path() → auto-select first (or None), checks ENV first
+- validate_browser_path(path) → validate user path, raise if invalid
+
+ENV override: HTMLREADER_CHROME_PATH
+"""
+
+from __future__ import annotations
+
+import logging
+import os
+import platform
+import shutil
+import subprocess
+
+logger = logging.getLogger(__name__)
+
+# --- Platform-specific known paths ---
+
+WIN32_PATHS: list[str] = [
+    "C:/Program Files (x86)/Google/Chrome/Application/chrome.exe",
+    "C:/Program Files/Google/Chrome/Application/chrome.exe",
+    "C:/Program Files (x86)/Microsoft/Edge/Application/msedge.exe",
+    "C:/Program Files/Microsoft/Edge/Application/msedge.exe",
+]
+
+# Add %USERPROFILE% based paths at runtime
+_WIN32_USERPROFILE_PATHS: list[str] = [
+    "AppData/Local/Google/Chrome/Application/chrome.exe",
+    "AppData/Local/Microsoft/Edge/Application/msedge.exe",
+]
+
+LINUX_PATHS: list[str] = [
+    "google-chrome",
+    "google-chrome-stable",
+    "google-chrome-beta",
+    "google-chrome-dev",
+    "microsoft-edge-stable",
+]
+
+DARWIN_PATHS: list[str] = [
+    "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome",
+    "/Applications/Microsoft Edge.app/Contents/MacOS/Microsoft Edge",
+]
+
+_VERSION_PREFIXES = (b"Google Chrome ", b"Microsoft Edge")
+
+
+class BrowserNotFoundError(Exception):
+    """Raised when no valid Chrome/Edge executable is found."""
+
+
+def _iter_browser_paths() -> list[str]:
+    """Detect installed Chrome/Edge paths on the current platform.
+
+    Returns a list of absolute paths, sorted by mtime descending (newest first).
+    """
+    current = platform.system()
+    found: set[str] = set()
+
+    if current == "Windows":
+        # Fixed known paths
+        for p in WIN32_PATHS:
+            if os.path.isfile(p):
+                found.add(os.path.abspath(p))
+        # USERPROFILE-based paths
+        userprofile = os.environ.get("USERPROFILE", "")
+        if userprofile:
+            for rel in _WIN32_USERPROFILE_PATHS:
+                full = os.path.join(userprofile, rel)
+                if os.path.isfile(full):
+                    found.add(os.path.abspath(full))
+
+    elif current == "Linux":
+        for cmd in LINUX_PATHS:
+            resolved = shutil.which(cmd)
+            if resolved and _verify_browser(resolved):
+                found.add(os.path.abspath(resolved))
+
+    elif current == "Darwin":
+        for p in DARWIN_PATHS:
+            if os.path.isfile(p):
+                found.add(os.path.abspath(p))
+
+    else:
+        logger.warning("Unsupported platform: %s", current)
+
+    result = sorted(found, key=_mtime_key, reverse=True)
+    return result
+
+
+def _verify_browser(path: str) -> bool:
+    """Check if a binary is actually Chrome/Edge by running --version."""
+    try:
+        out = subprocess.check_output(
+            [path, "--version"], timeout=2, stderr=subprocess.DEVNULL
+        )
+        return any(out.startswith(prefix) for prefix in _VERSION_PREFIXES)
+    except (FileNotFoundError, subprocess.TimeoutExpired, OSError):
+        return False
+
+
+def _mtime_key(path: str) -> float:
+    """Get mtime for sorting, 0 if file doesn't exist."""
+    try:
+        return os.path.getmtime(path)
+    except OSError:
+        return 0.0
+
+
+def detect_browser_paths() -> list[str]:
+    """List all detected Chrome/Edge browser paths on this platform.
+
+    Returns absolute paths sorted by modification time (newest first).
+    """
+    return _iter_browser_paths()
+
+
+def get_browser_path() -> str | None:
+    """Auto-select the best browser path.
+
+    Priority:
+    1. HTMLREADER_CHROME_PATH env var (if set and valid)
+    2. First detected path (newest by mtime)
+
+    Returns None if no browser is found.
+    """
+    # ENV override
+    env_path = os.environ.get("HTMLREADER_CHROME_PATH", "").strip()
+    if env_path:
+        try:
+            return validate_browser_path(env_path)
+        except BrowserNotFoundError:
+            logger.warning(
+                "HTMLREADER_CHROME_PATH=%s is invalid, falling back to auto-detect",
+                env_path,
+            )
+
+    paths = detect_browser_paths()
+    return paths[0] if paths else None
+
+
+def validate_browser_path(path: str) -> str:
+    """Validate a user-provided browser path.
+
+    Args:
+        path: Path to Chrome/Edge executable.
+
+    Returns:
+        Absolute resolved path.
+
+    Raises:
+        BrowserNotFoundError: If the path does not exist or is not a valid file.
+    """
+    resolved = os.path.abspath(path)
+    if not os.path.isfile(resolved):
+        raise BrowserNotFoundError(f"Browser executable not found: {resolved}")
+    return resolved