docvortex 0.4.11__tar.gz → 0.4.12__tar.gz

{docvortex-0.4.11/src/docvortex.egg-info → docvortex-0.4.12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docvortex
-Version: 0.4.11
+Version: 0.4.12
 Summary: A fast, multi-format document parsing and conversion engine
 License-Expression: MIT
 Project-URL: Repository, https://github.com/myhloli/DocVortex

{docvortex-0.4.11 → docvortex-0.4.12}/src/docvortex/analyzers/native/html/constants.py

@@ -8,7 +8,6 @@ MAX_HTML_NODES: Final = 2_000_000
 MAX_HTML_DEPTH: Final = 256
 MAX_HTML_IMAGE_BYTES: Final = 20 * 1024 * 1024
 MAX_HTML_IMAGE_TOTAL_BYTES: Final = 128 * 1024 * 1024
-MAX_REMOTE_IMAGE_FETCH_SECONDS: Final = 10
 MAX_HTML_STYLESHEET_BYTES: Final = 5 * 1024 * 1024
 MAX_HTML_STYLESHEET_TOTAL_BYTES: Final = 20 * 1024 * 1024
 MAX_HTML_RENDERED_BYTES: Final = 256 * 1024 * 1024
@@ -23,5 +22,4 @@ __all__ = [
     "MAX_HTML_RENDERED_BYTES",
     "MAX_HTML_STYLESHEET_BYTES",
     "MAX_HTML_STYLESHEET_TOTAL_BYTES",
-    "MAX_REMOTE_IMAGE_FETCH_SECONDS",
 ]

{docvortex-0.4.11 → docvortex-0.4.12}/src/docvortex/analyzers/native/html/resources.py

@@ -6,7 +6,6 @@ import base64
 from pathlib import Path
 from typing import Protocol
 from urllib.parse import SplitResult, unquote, urljoin, urlsplit, urlunsplit
-from urllib.request import urlopen
 
 from lxml import etree  # type: ignore[reportMissingImports]
 
@@ -21,7 +20,6 @@ from .constants import (
     MAX_HTML_IMAGE_TOTAL_BYTES,
     MAX_HTML_STYLESHEET_BYTES,
     MAX_HTML_STYLESHEET_TOTAL_BYTES,
-    MAX_REMOTE_IMAGE_FETCH_SECONDS,
 )
 from .errors import HtmlResourceLimitError
 
@@ -68,7 +66,6 @@ class HtmlResourceContext:
         self._image_bytes = 0
         self._stylesheet_bytes = 0
         self._image_cache: dict[Path, str] = {}
-        self._remote_image_cache: dict[str, str] = {}
         self._data_image_cache: dict[str, ResolvedMarkupImage] = {}
         self._stylesheet_cache: dict[Path, str] = {}
         self._local_root = source_context.local_resource_root.resolve() if source_context.local_resource_root else None
@@ -161,14 +158,14 @@ class HtmlResourceContext:
         return sanitize_hyperlink_target(normalized, allow_relative=True, allow_fragment=True)
 
     def resolve_image(self, source: str, *, alt: str = "") -> ResolvedMarkupImage | None:
-        """按 data URI、远程 URL、本地安全文件的顺序解析图片，远程图片仅在来源显式开启下载时内嵌。"""
+        """按 data URI、远程 URL、本地安全文件的顺序解析图片，远程图片仅保留受限外链。"""
         normalized = (source or "").strip()
         if not normalized:
             return ResolvedMarkupImage(alt=alt) if alt else None
         if normalized.casefold().startswith("data:"):
             return self._resolve_data_image(normalized, alt=alt)
         if remote_url := self._resolve_remote_image_url(normalized):
-            return self._resolve_remote_image(remote_url, alt=alt)
+            return ResolvedMarkupImage(image_url=remote_url, alt=alt)
         local_path = self._resolve_local_path(normalized)
         if local_path is None or not local_path.is_file():
             return ResolvedMarkupImage(alt=alt) if alt else None
@@ -258,41 +255,6 @@ class HtmlResourceContext:
         self._data_image_cache[data_uri] = resolved
         return resolved
 
-    def _resolve_remote_image(self, remote_url: str, *, alt: str) -> ResolvedMarkupImage:
-        """远程图片默认保持外链，来源显式开启下载时按缓存下载内嵌。"""
-        if not self.source_context.fetch_remote_images:
-            return ResolvedMarkupImage(image_url=remote_url, alt=alt)
-        if cached := self._remote_image_cache.get(remote_url):
-            return ResolvedMarkupImage(image_base64=cached, alt=alt)
-        data_uri = self._fetch_remote_image(remote_url)
-        if data_uri is None:
-            return ResolvedMarkupImage(image_url=remote_url, alt=alt)
-        self._remote_image_cache[remote_url] = data_uri
-        return ResolvedMarkupImage(image_base64=data_uri, alt=alt)
-
-    def _fetch_remote_image(self, remote_url: str) -> str | None:
-        """在单图上限与超时约束内下载远程图片并复核签名，网络失败返回 None 退回外链。"""
-        try:
-            with urlopen(remote_url, timeout=MAX_REMOTE_IMAGE_FETCH_SECONDS) as response:
-                chunks: list[bytes] = []
-                remaining = MAX_HTML_IMAGE_BYTES + 1
-                while remaining > 0:
-                    chunk = response.read(remaining)
-                    if not chunk:
-                        break
-                    chunks.append(chunk)
-                    remaining -= len(chunk)
-        except (OSError, ValueError):
-            return None
-        payload = b"".join(chunks)
-        if len(payload) > MAX_HTML_IMAGE_BYTES:
-            raise HtmlResourceLimitError(f"HTML image exceeds max_html_image_bytes={MAX_HTML_IMAGE_BYTES}")
-        data_uri = _image_data_uri(payload)
-        if data_uri is None:
-            return None
-        self._charge_image_bytes(len(payload))
-        return data_uri
-
     def _charge_image_bytes(self, byte_count: int) -> None:
         """累计实际保留的图片字节，并在超限时终止整份文档。"""
         self._image_bytes += byte_count

{docvortex-0.4.11 → docvortex-0.4.12}/src/docvortex/document/contracts.py

@@ -10,14 +10,12 @@ from pathlib import Path
 class HtmlSourceContext:
     """保存相对链接解析及 HTML 解码所需的来源上下文。
 
-    fetch_remote_images 默认关闭，远程图片仅记录 URL；开启后解析阶段会对受限
-    HTTP(S) 图片发起网络请求并内嵌，不可信输入的 SSRF 风险由调用者承担。
+    远程图片永不下载，仅在结果中保留受限 HTTP(S) 外链，解析过程保持零网络请求。
     """
 
     source_uri: str | None = None
     local_resource_root: Path | None = None
     transport_encoding: str | None = None
-    fetch_remote_images: bool = False
 
 
 __all__ = ["HtmlSourceContext"]

{docvortex-0.4.11 → docvortex-0.4.12}/src/docvortex/export/files.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 from pathlib import Path
+from urllib.parse import urlsplit
 
 from ..assets import AssetStore
 from ..content.tree import iter_child_blocks
@@ -28,7 +29,7 @@ def materialize_middle(middle_json: MiddleJson, assets: AssetStore | None = None
 
 
 def validate_materialized_assets(middle_json: MiddleJson, assets: AssetStore) -> None:
-    """拒绝依赖外部文件或网络的未物化图片，保证结果包能离线重渲染。"""
+    """拒绝缺失的已物化素材，保证结果包内部引用完整；源文档外链原样保留。"""
     from bs4 import BeautifulSoup
 
     pending = [block for page in middle_json.pages for block in page.blocks]
@@ -38,8 +39,6 @@ def validate_materialized_assets(middle_json: MiddleJson, assets: AssetStore) ->
         image_path = getattr(block, "image_path", None)
         if image_path and image_path not in assets:
             raise ValueError(f"Missing materialized asset: {image_path}")
-        if getattr(block, "image_url", None) and not image_path:
-            raise ValueError("An external image must be materialized before saving a bundle")
         content = getattr(block, "content", None)
         if str(block.type) not in {"table_body", "chart_body", "image_body"} or not isinstance(content, str):
             continue
@@ -47,6 +46,8 @@ def validate_materialized_assets(middle_json: MiddleJson, assets: AssetStore) ->
             continue
         for image in BeautifulSoup(content, "html.parser").find_all("img"):
             reference = image.get("src")
+            if reference and _is_external_reference(reference):
+                continue
             if reference and reference not in assets:
                 raise ValueError(f"Missing materialized HTML image: {reference}")
 
@@ -63,4 +64,12 @@ def write_artifact(artifact: RenderArtifact, path: Path, *, overwrite: bool) ->
     return ExportResult(path=path, asset_paths=tuple(path.parent / name for name in artifact.assets))
 
 
+def _is_external_reference(reference: str) -> bool:
+    """判断富文本图片引用是否为保留的受限 HTTP(S) 外链。"""
+    try:
+        return urlsplit(reference).scheme.casefold() in {"http", "https"}
+    except ValueError:
+        return False
+
+
 __all__ = ["materialize_middle", "validate_materialized_assets", "write_artifact"]

{docvortex-0.4.11 → docvortex-0.4.12}/src/docvortex/version.py

@@ -1,3 +1,3 @@
-__version__ = "0.4.11"
+__version__ = "0.4.12"
 
 __all__ = ["__version__"]

{docvortex-0.4.11 → docvortex-0.4.12/src/docvortex.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docvortex
-Version: 0.4.11
+Version: 0.4.12
 Summary: A fast, multi-format document parsing and conversion engine
 License-Expression: MIT
 Project-URL: Repository, https://github.com/myhloli/DocVortex

{docvortex-0.4.11 → docvortex-0.4.12}/tests/unittest/test_flash_html.py

@@ -23,11 +23,11 @@ from docvortex.analyzers.native.html import HtmlResourceLimitError
 from docvortex.analyzers.native.html import document as html_document_module
 from docvortex.analyzers.native.html import resources as html_resources_module
 from docvortex.analyzers.native.html import selector as html_selector_module
-from docvortex.analyzers.native.html.constants import MAX_HTML_IMAGE_BYTES
 from docvortex.analyzers.native.html.resources import HtmlResourceContext
 from docvortex.api import parse
 from docvortex.codecs.html import decode_docvortex_html_wire
 from docvortex.document.contracts import HtmlSourceContext
+from docvortex.export.bundle import load_bundle
 from docvortex.render import RenderMode
 from docvortex.render.html import render_html
 from docvortex.render.markdown import render_markdown
@@ -2083,8 +2083,8 @@ def _remote_chart_middle(url: str) -> MiddleJson:
     )
 
 
-def test_html_remote_images_stay_external_and_offline_by_default(tmp_path: Path) -> None:
-    """默认不下载远程图片：保持外链、零网络请求，save_bundle 仍拒绝未物化图片。"""
+def test_html_remote_images_stay_external_and_offline(tmp_path: Path) -> None:
+    """远程图片永不下载：保持外链、零网络请求，save_bundle 保留外链且可重新加载。"""
     with _local_image_server({"/chart.png": _tiny_png_bytes()}) as (base, hits):
         url = f"{base}/chart.png"
         default = HtmlResourceContext(HtmlSourceContext())
@@ -2096,62 +2096,9 @@ def test_html_remote_images_stay_external_and_offline_by_default(tmp_path: Path)
         body = result.middle_json.pages[0].blocks[0].content[0]
         assert body.image_url == url and not body.image_path
 
-        with pytest.raises(ValueError, match="must be materialized"):
-            result.save_bundle(tmp_path / "bundle")
+        bundle = result.save_bundle(tmp_path / "bundle")
+        assert list(bundle.asset_paths) == []
+        reloaded = load_bundle(tmp_path / "bundle")
+        reloaded_body = reloaded.middle_json.pages[0].blocks[0].content[0]
+        assert reloaded_body.image_url == url and not reloaded_body.image_path
         assert hits == []
-
-
-def test_html_remote_image_fetch_embeds_and_caches_by_url() -> None:
-    """开启下载后有效图片内嵌为 data URI，同一 URL 复用缓存不重复请求。"""
-    png = _tiny_png_bytes()
-    with _local_image_server({"/chart.png": png}) as (base, hits):
-        context = HtmlResourceContext(HtmlSourceContext(fetch_remote_images=True))
-        first = context.resolve_image(f"{base}/chart.png", alt="chart")
-        assert first is not None and first.image_url is None
-        assert first.image_base64 is not None and first.image_base64.startswith("data:image/png;base64,")
-        assert base64.b64decode(first.image_base64.split(",", 1)[1]) == png
-        assert hits == ["/chart.png"]
-
-        second = context.resolve_image(f"{base}/chart.png", alt="chart")
-        assert second is not None and second.image_base64 == first.image_base64
-        assert hits == ["/chart.png"]
-
-
-def test_html_remote_image_fetch_falls_back_to_url_when_unavailable() -> None:
-    """非图片字节或网络不可达时下载退回外链，不阻断解析。"""
-    with _local_image_server({"/text.png": b"not-an-image"}) as (base, hits):
-        context = HtmlResourceContext(HtmlSourceContext(fetch_remote_images=True))
-        resolved = context.resolve_image(f"{base}/text.png", alt="chart")
-        assert resolved is not None and resolved.image_url == f"{base}/text.png" and not resolved.image_base64
-
-        dead = "http://127.0.0.1:9/chart.png"
-        resolved = context.resolve_image(dead)
-        assert resolved is not None and resolved.image_url == dead and not resolved.image_base64
-
-
-def test_html_remote_image_fetch_rejects_oversized_download() -> None:
-    """下载超过单图上限的远程图片时与本地文件一致地终止整份文档。"""
-    oversized = b"\x89PNG\r\n\x1a\n" + b"\x00" * (MAX_HTML_IMAGE_BYTES + 1)
-    with _local_image_server({"/big.png": oversized}) as (base, hits):
-        context = HtmlResourceContext(HtmlSourceContext(fetch_remote_images=True))
-        with pytest.raises(HtmlResourceLimitError, match="max_html_image_bytes"):
-            context.resolve_image(f"{base}/big.png")
-
-
-def test_html_remote_image_fetch_roundtrips_into_saved_bundle(tmp_path: Path) -> None:
-    """开启下载后：远程图片在解析阶段内嵌，save_bundle 不再依赖网络且图片入包。"""
-    png = _tiny_png_bytes()
-    with _local_image_server({"/chart.png": png}) as (base, hits):
-        url = f"{base}/chart.png"
-        rendered = render_html(_remote_chart_middle(url), standalone=False)
-        result = parse(
-            rendered.encode(),
-            file_suffix="html",
-            source_context=HtmlSourceContext(fetch_remote_images=True),
-        )
-        assert hits == ["/chart.png"]
-        body = result.middle_json.pages[0].blocks[0].content[0]
-        assert body.image_path and not body.image_url and not body.image_base64
-
-        result.save_bundle(tmp_path / "bundle")
-        assert [path.read_bytes() for path in sorted((tmp_path / "bundle" / "images").glob("*"))] == [png]