docvortex 0.4.8__tar.gz → 0.4.10__tar.gz

{docvortex-0.4.8/src/docvortex.egg-info → docvortex-0.4.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docvortex
-Version: 0.4.8
+Version: 0.4.10
 Summary: A fast, multi-format document parsing and conversion engine
 License-Expression: MIT
 Project-URL: Repository, https://github.com/myhloli/DocVortex
@@ -77,13 +77,12 @@ DocVortex is a standalone Python engine for parsing and converting documents.
 It reads native text and document structure into a unified representation,
 then exports the result in the formats your workflow needs.
 
-- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD and CSV.
+- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD, CSV and TSV.
 - **Parse once, export many times** — reuse the same result for Markdown, HTML, LaTeX, DOCX, EPUB, PDF and structured JSON.
 - **Portable results** — save document structure and image assets in a Bundle, then export again without the source file.
 - **Composable APIs** — use the complete pipeline or integrate analysis, postprocessing and rendering separately.
 
-Native parsing works without an OCR or VLM inference service. Use DocVortex
-directly through its CLI or Python SDK, independently of MinerU.
+Native parsing works without an OCR or VLM inference service. Use DocVortex directly through its CLI or Python SDK.
 
 ![DocVortex pipeline: native documents become a unified representation, then Markdown, HTML, LaTeX, DOCX, EPUB, PDF or structured JSON.](https://gcore.jsdelivr.net/gh/myhloli/DocVortex@main/docs/images/docvortex-overview.jpg)
 
@@ -135,7 +134,7 @@ use `overwrite=True` in Python or `--overwrite` in the CLI to replace them.
 | PDF with native text | PDF |
 | Word & rich text | DOC, DOCX, RTF |
 | Presentations | PPT, PPTX |
-| Spreadsheets | XLS, XLSX, CSV |
+| Spreadsheets | XLS, XLSX, CSV, TSV |
 | OpenDocument | ODT, ODS, ODP |
 | E-books & web documents | EPUB, HTML |
 | Open Fixed-layout Document | OFD |

{docvortex-0.4.8 → docvortex-0.4.10}/README.md

@@ -23,13 +23,12 @@ DocVortex is a standalone Python engine for parsing and converting documents.
 It reads native text and document structure into a unified representation,
 then exports the result in the formats your workflow needs.
 
-- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD and CSV.
+- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD, CSV and TSV.
 - **Parse once, export many times** — reuse the same result for Markdown, HTML, LaTeX, DOCX, EPUB, PDF and structured JSON.
 - **Portable results** — save document structure and image assets in a Bundle, then export again without the source file.
 - **Composable APIs** — use the complete pipeline or integrate analysis, postprocessing and rendering separately.
 
-Native parsing works without an OCR or VLM inference service. Use DocVortex
-directly through its CLI or Python SDK, independently of MinerU.
+Native parsing works without an OCR or VLM inference service. Use DocVortex directly through its CLI or Python SDK.
 
 ![DocVortex pipeline: native documents become a unified representation, then Markdown, HTML, LaTeX, DOCX, EPUB, PDF or structured JSON.](https://gcore.jsdelivr.net/gh/myhloli/DocVortex@main/docs/images/docvortex-overview.jpg)
 
@@ -81,7 +80,7 @@ use `overwrite=True` in Python or `--overwrite` in the CLI to replace them.
 | PDF with native text | PDF |
 | Word & rich text | DOC, DOCX, RTF |
 | Presentations | PPT, PPTX |
-| Spreadsheets | XLS, XLSX, CSV |
+| Spreadsheets | XLS, XLSX, CSV, TSV |
 | OpenDocument | ODT, ODS, ODP |
 | E-books & web documents | EPUB, HTML |
 | Open Fixed-layout Document | OFD |

{docvortex-0.4.8 → docvortex-0.4.10}/docs/METADATA.md

@@ -43,7 +43,7 @@ FILETIME values are UTC. Missing or invalid dates are not replaced with filesyst
 | EPUB | OPF Dublin Core and recognized meta properties | `spine` |
 | OFD | DocInfo; first nonempty scalar across DocBody entries | `physical`, summed across documents |
 | HTML | title, standard meta, Dublin Core, Open Graph, document language | `logical`, 1 |
-| CSV | no embedded descriptive property convention | `logical`, 1 |
+| CSV / TSV | no embedded descriptive property convention | `logical`, 1 |
 
 A declared count may be stale and is not a measured layout. Counts are never inferred from the
 number of pages selected for parsing. Source author strings are not authors inferred from body text.

{docvortex-0.4.8 → docvortex-0.4.10}/schemas/middle-2.0.json

@@ -762,6 +762,7 @@
             "xlsx",
             "rtf",
             "csv",
+            "tsv",
             "epub",
             "html",
             "ofd",

{docvortex-0.4.8 → docvortex-0.4.10}/schemas/model-2.0.json

@@ -15,6 +15,7 @@
             "xlsx",
             "rtf",
             "csv",
+            "tsv",
             "epub",
             "html",
             "ofd",

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/analyzers/native/html/constants.py

@@ -8,6 +8,7 @@ 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
@@ -22,4 +23,5 @@ __all__ = [
     "MAX_HTML_RENDERED_BYTES",
     "MAX_HTML_STYLESHEET_BYTES",
     "MAX_HTML_STYLESHEET_TOTAL_BYTES",
+    "MAX_REMOTE_IMAGE_FETCH_SECONDS",
 ]

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/analyzers/native/html/resources.py

@@ -6,6 +6,7 @@ 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]
 
@@ -19,6 +20,7 @@ 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
 
@@ -65,6 +67,7 @@ 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
@@ -157,14 +160,14 @@ 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、本地安全文件、远程 URL 的顺序解析图片。"""
+        """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。"""
         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 ResolvedMarkupImage(image_url=remote_url, alt=alt)
+            return self._resolve_remote_image(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
@@ -243,6 +246,41 @@ 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-0.4.8 → docvortex-0.4.10}/src/docvortex/analyzers/native/office/odf/styles.py

@@ -260,11 +260,15 @@ class OdfStyles:
             current = definition.parent or ""
         return False
 
-    def list_level(self, style_name: str | None, depth: int) -> ListLevel:
-        """返回指定列表深度的定义，不存在时使用无序默认值。"""
+    def list_level(self, style_name: str | None, depth: int) -> ListLevel | None:
+        """返回指定列表深度的定义；样式已定义但无任何层级时返回 None（无可见标记）。"""
         if not style_name:
             return ListLevel()
-        levels = self._list_styles.get(style_name, {})
+        levels = self._list_styles.get(style_name)
+        if levels is None:
+            return ListLevel()
+        if not levels:
+            return None
         return levels.get(depth, ListLevel())
 
     def table_is_visible(self, style_name: str | None) -> bool:

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/analyzers/native/office/odf/table.py

@@ -169,7 +169,6 @@ def parse_table_grid(
     duplicated_text_bytes = 0
     row_index = 0
     pending_empty_rows = 0
-    pending_empty_width = 0
 
     def ensure_row(target_row: int, width: int = 0) -> list[GridCell | None]:
         """在兼容独立表格调用的同时应用可选文档级预留。"""
@@ -179,7 +178,6 @@ def parse_table_grid(
 
     for row_element, header in _iter_rows(table):
         row_repeat = _positive_int(row_element.get(qname("table", "number-rows-repeated")))
-        _validate_grid_extent(row_index + row_repeat, grid.width)
         cell_templates: list[tuple[bool, int, GridCell | None]] = []
         for cell in row_element:
             if not isinstance(cell.tag, str):
@@ -219,18 +217,14 @@ def parse_table_grid(
         )
         if not row_has_content:
             pending_empty_rows += row_repeat
-            pending_empty_width = max(
-                pending_empty_width,
-                sum(repeat * (template.col_span if template is not None else 1) for _, repeat, template in cell_templates),
-            )
             continue
         if pending_empty_rows:
-            _validate_grid_extent(row_index + pending_empty_rows, max(grid.width, pending_empty_width))
+            _validate_grid_extent(row_index + pending_empty_rows, grid.width)
             for _ in range(pending_empty_rows):
-                ensure_row(row_index, pending_empty_width)
+                ensure_row(row_index, grid.width)
                 row_index += 1
             pending_empty_rows = 0
-            pending_empty_width = 0
+        _validate_grid_extent(row_index + row_repeat, grid.width)
         for _ in range(row_repeat):
             row = ensure_row(row_index)
             col_index = 0

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/analyzers/native/office/odf/text.py

@@ -453,7 +453,7 @@ class OdfBlockParser:
         style_name = element.get(qname("text", "style-name")) or inherited_style
         level = self.styles.list_level(style_name, depth)
         key = (style_name or "", depth)
-        start = level.start
+        start = level.start if level is not None else 1
         continue_list = element.get(qname("text", "continue-list"))
         if continue_list and continue_list in self._list_ids:
             start = self._list_ids[continue_list]
@@ -466,17 +466,20 @@ class OdfBlockParser:
         active_master: str | None = None
 
         def flush_content(fragment_start: int) -> None:
-            """把当前合法子块冻结为一个 LIST 分片。"""
+            """把当前合法子块冻结为一个 LIST 分片；无标记列表按普通段落序列输出。"""
             if content:
-                block: dict[str, Any] = {
-                    "type": BlockType.LIST,
-                    "attribute": "ordered" if level.ordered else "unordered",
-                    "ilevel": depth,
-                    "content": list(content),
-                }
-                if level.ordered:
-                    block["start"] = fragment_start
-                results.append(block)
+                if level is None:
+                    results.extend(content)
+                else:
+                    block: dict[str, Any] = {
+                        "type": BlockType.LIST,
+                        "attribute": "ordered" if level.ordered else "unordered",
+                        "ilevel": depth,
+                        "content": list(content),
+                    }
+                    if level.ordered:
+                        block["start"] = fragment_start
+                    results.append(block)
                 content.clear()
             if fragment_notes:
                 results.extend(fragment_notes)
@@ -875,6 +878,20 @@ class OdfBlockParser:
         """把单元格行内流中的 note marker 排入当前逻辑页队列。"""
         self.notes.extend(atom.content for atom in atoms if isinstance(atom, InlineNote))
 
+    def _render_cell_paragraph(self, child: etree._Element, parts: list[str]) -> None:
+        """把单元格里的一个裸段落渲染为 <p>，并回收行内块组的单元格子块。"""
+        atoms = self.parse_inline_atoms(child)
+        self._queue_inline_notes(atoms)
+        rendered_atoms = [atom for atom in atoms if not isinstance(atom, InlineImage)] if self._collect_cell_visuals else atoms
+        parts.append(f"<p>{render_atoms_to_html(rendered_atoms)}</p>")
+        for atom in atoms:
+            if isinstance(atom, InlineBlockGroup):
+                self._append_cell_blocks(
+                    parts,
+                    atom.blocks,
+                    inline_image_rendered=atom.inline_image_rendered and not self._collect_cell_visuals,
+                )
+
     def render_cell_html(self, cell: etree._Element) -> str:
         """把表格单元格中的段落、列表、嵌套表和 frame 转为 HTML。"""
         parts: list[str] = []
@@ -882,19 +899,7 @@ class OdfBlockParser:
             if not isinstance(child.tag, str):
                 continue
             if child.tag in {qname("text", "p"), qname("text", "h")}:
-                atoms = self.parse_inline_atoms(child)
-                self._queue_inline_notes(atoms)
-                rendered_atoms = (
-                    [atom for atom in atoms if not isinstance(atom, InlineImage)] if self._collect_cell_visuals else atoms
-                )
-                parts.append(f"<p>{render_atoms_to_html(rendered_atoms)}</p>")
-                for atom in atoms:
-                    if isinstance(atom, InlineBlockGroup):
-                        self._append_cell_blocks(
-                            parts,
-                            atom.blocks,
-                            inline_image_rendered=atom.inline_image_rendered and not self._collect_cell_visuals,
-                        )
+                self._render_cell_paragraph(child, parts)
             elif child.tag == qname("text", "list"):
                 parts.append(self._render_list_html(child))
             elif child.tag == qname("table", "table"):
@@ -915,6 +920,8 @@ class OdfBlockParser:
         """把单元格内 ODF 列表递归渲染为 ol/ul HTML。"""
         style_name = element.get(qname("text", "style-name")) or inherited_style
         level = self.styles.list_level(style_name, depth)
+        if level is None:
+            return self._render_unmarked_list_html(element, depth=depth, inherited_style=style_name)
         tag = "ol" if level.ordered else "ul"
         start = f' start="{level.start}"' if level.ordered and level.start != 1 else ""
         parts = [f"<{tag}{start}>"]
@@ -940,6 +947,30 @@ class OdfBlockParser:
         parts.append(f"</{tag}>")
         return "".join(parts)
 
+    def _render_unmarked_list_html(
+        self,
+        element: etree._Element,
+        *,
+        depth: int,
+        inherited_style: str | None,
+    ) -> str:
+        """把引用空列表样式（无可见标记）的单元格列表按裸段落序列渲染。"""
+        parts: list[str] = []
+        for item in element:
+            if not isinstance(item.tag, str) or item.tag not in {
+                qname("text", "list-item"),
+                qname("text", "list-header"),
+            }:
+                continue
+            for child in item:
+                if not isinstance(child.tag, str):
+                    continue
+                if child.tag in {qname("text", "p"), qname("text", "h")}:
+                    self._render_cell_paragraph(child, parts)
+                elif child.tag == qname("text", "list"):
+                    parts.append(self._render_list_html(child, depth=depth + 1, inherited_style=inherited_style))
+        return "".join(part for part in parts if part)
+
     def drain_notes(self) -> list[str]:
         """取出当前累计脚注并清空共享队列。"""
         values = list(self.notes)

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/api.py

@@ -84,6 +84,7 @@ def analyze(
         else:
             model_types: dict[FileSuffix, type[NativeBinaryAnalyzer]] = {
                 "csv": models.CsvModel,
+                "tsv": models.CsvModel,
                 "epub": models.EpubModel,
                 "doc": models.DocModel,
                 "docx": models.DocxModel,

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/document/contracts.py

@@ -8,11 +8,16 @@ from pathlib import Path
 
 @dataclass(frozen=True, slots=True)
 class HtmlSourceContext:
-    """保存相对链接解析及 HTML 解码所需的来源上下文。"""
+    """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。
+    """
 
     source_uri: str | None = None
     local_resource_root: Path | None = None
     transport_encoding: str | None = None
+    fetch_remote_images: bool = False
 
 
 __all__ = ["HtmlSourceContext"]

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/document/detection.py

@@ -279,13 +279,13 @@ def _has_rtf_signature_by_path(file_path: Path) -> bool:
 
 
 def _resolve_signatureless_csv_suffix(detected_suffix: str, file_path: str | Path | None) -> str:
-    """以 .csv/.tsv 扩展名兜底无签名分隔文本，并保留强内容类型的优先级。"""
+    """以 .csv/.tsv 扩展名兜底无签名分隔文本：按扩展名返回独立后缀，并保留强内容类型的优先级。"""
     extension = Path(file_path).suffix.lower().lstrip(".") if file_path else ""
     if extension in CSV_EXTENSIONS:
         if detected_suffix in _STRONG_CONTENT_SUFFIXES:
             return detected_suffix
-        return "csv"
-    if detected_suffix == "csv":
+        return extension
+    if detected_suffix in ("csv", "tsv"):
         if extension in ODF_MIMETYPE_SUFFIXES.values():
             return "txt"
         return extension or "txt"

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/document/filetypes.py

@@ -65,7 +65,8 @@ FILE_TYPE_BY_EXTENSION: dict[str, str] = {
     **dict.fromkeys(IMAGE_EXTENSIONS, "image"),
     **{ext: ext for ext in OFFICE_EXTENSIONS},
     **dict.fromkeys(HTML_EXTENSIONS, "html"),
-    **dict.fromkeys(CSV_EXTENSIONS, "csv"),
+    "csv": "csv",
+    "tsv": "tsv",
     **dict.fromkeys(EPUB_EXTENSIONS, "epub"),
     **dict.fromkeys(OFD_EXTENSIONS, "ofd"),
     "txt": "text",

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/schema.py

@@ -56,6 +56,7 @@ FileSuffix: TypeAlias = Literal[
     "xlsx",
     "rtf",
     "csv",
+    "tsv",
     "epub",
     "html",
     "ofd",

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex/version.py

@@ -1,3 +1,3 @@
-__version__ = "0.4.8"
+__version__ = "0.4.10"
 
 __all__ = ["__version__"]

{docvortex-0.4.8 → docvortex-0.4.10/src/docvortex.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docvortex
-Version: 0.4.8
+Version: 0.4.10
 Summary: A fast, multi-format document parsing and conversion engine
 License-Expression: MIT
 Project-URL: Repository, https://github.com/myhloli/DocVortex
@@ -77,13 +77,12 @@ DocVortex is a standalone Python engine for parsing and converting documents.
 It reads native text and document structure into a unified representation,
 then exports the result in the formats your workflow needs.
 
-- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD and CSV.
+- **Multi-format input** — read text PDFs, Office files, OpenDocument files, EPUB, HTML, OFD, CSV and TSV.
 - **Parse once, export many times** — reuse the same result for Markdown, HTML, LaTeX, DOCX, EPUB, PDF and structured JSON.
 - **Portable results** — save document structure and image assets in a Bundle, then export again without the source file.
 - **Composable APIs** — use the complete pipeline or integrate analysis, postprocessing and rendering separately.
 
-Native parsing works without an OCR or VLM inference service. Use DocVortex
-directly through its CLI or Python SDK, independently of MinerU.
+Native parsing works without an OCR or VLM inference service. Use DocVortex directly through its CLI or Python SDK.
 
 ![DocVortex pipeline: native documents become a unified representation, then Markdown, HTML, LaTeX, DOCX, EPUB, PDF or structured JSON.](https://gcore.jsdelivr.net/gh/myhloli/DocVortex@main/docs/images/docvortex-overview.jpg)
 
@@ -135,7 +134,7 @@ use `overwrite=True` in Python or `--overwrite` in the CLI to replace them.
 | PDF with native text | PDF |
 | Word & rich text | DOC, DOCX, RTF |
 | Presentations | PPT, PPTX |
-| Spreadsheets | XLS, XLSX, CSV |
+| Spreadsheets | XLS, XLSX, CSV, TSV |
 | OpenDocument | ODT, ODS, ODP |
 | E-books & web documents | EPUB, HTML |
 | Open Fixed-layout Document | OFD |

{docvortex-0.4.8 → docvortex-0.4.10}/src/docvortex.egg-info/SOURCES.txt

@@ -430,7 +430,6 @@ src/docvortex/render/_internal/pdf/title_layout.py
 src/docvortex/render/_internal/structured_content/__init__.py
 src/docvortex/render/_internal/structured_content/renderer.py
 src/docvortex/resources/epub/docvortex.css
-src/docvortex/resources/fasttext-langdetect/lid.176.ftz
 src/docvortex/resources/fonts/DroidSansFallbackFull.ttf
 src/docvortex/resources/fonts/NOTICE
 src/docvortex/resources/fonts/manifest.json

{docvortex-0.4.8 → docvortex-0.4.10}/tests/test_cli_logging.py

@@ -20,7 +20,7 @@ def isolated_logger() -> Iterator[None]:
     logger.remove()
     yield
     logger.remove()
-    logger.add(sys.stderr)
+    logger.add(sys.stderr, level="INFO")
 
 
 class _DummyPDFDocument:

{docvortex-0.4.8 → docvortex-0.4.10}/tests/test_format_matrix.py

@@ -1,4 +1,4 @@
-"""覆盖全部十五种原生文档输入与七种输出的独立端到端矩阵。"""
+"""覆盖全部十六种原生文档输入与七种输出的独立端到端矩阵。"""
 
 from __future__ import annotations
 
@@ -27,6 +27,8 @@ def source_payload(suffix: str) -> bytes:
         return builders[suffix]()
     if suffix == "csv":
         return b"name,value\nalpha,1\nbeta,2\n"
+    if suffix == "tsv":
+        return b"name\tvalue\nalpha\t1\nbeta\t2\n"
     if suffix == "html":
         return b"<html><h1>Title</h1><p>Native conversion</p></html>"
     if suffix == "pdf":
@@ -35,7 +37,7 @@ def source_payload(suffix: str) -> bytes:
 
 
 @pytest.mark.parametrize(
-    "suffix", ["pdf", "doc", "docx", "ppt", "pptx", "xls", "xlsx", "rtf", "csv", "html", "epub", "ofd", "odt", "ods", "odp"]
+    "suffix", ["pdf", "doc", "docx", "ppt", "pptx", "xls", "xlsx", "rtf", "csv", "tsv", "html", "epub", "ofd", "odt", "ods", "odp"]
 )
 def test_all_native_formats_render_all_targets(suffix: str, monkeypatch: pytest.MonkeyPatch) -> None:
     """每种原生格式完成分析、后处理和七种目标编码，不导入宿主。"""

{docvortex-0.4.8 → docvortex-0.4.10}/tests/unittest/test_flash_csv.py

@@ -2,6 +2,7 @@ from __future__ import annotations
 
 import codecs
 from io import BytesIO
+from pathlib import Path
 
 import pytest
 from _native_test_utils import analyze_native_test_document
@@ -9,6 +10,7 @@ from bs4 import BeautifulSoup
 
 from docvortex.analyzers.native import CsvModel
 from docvortex.analyzers.native import csv as csv_module
+from docvortex.document.detection import guess_suffix_by_bytes, guess_suffix_by_path
 from docvortex.render.html import render_html
 from docvortex.render.markdown import render_markdown
 from docvortex.schema import BlockType
@@ -169,3 +171,45 @@ def test_csv_default_grid_budget_rejects_wide_dom_before_rendering() -> None:
 
     with pytest.raises(ValueError, match="max_grid_slots"):
         CsvModel().predict(BytesIO(payload))
+
+
+def test_delimited_extension_resolves_to_independent_suffix() -> None:
+    """以 .csv/.tsv 路径兜底时按扩展名返回独立后缀，不再把 tsv 折叠为 csv。"""
+    payload = b"name,city\nAlice,London\n"
+
+    assert guess_suffix_by_bytes(payload, "demo.csv") == "csv"
+    assert guess_suffix_by_bytes(payload, "demo.tsv") == "tsv"
+
+
+def test_tsv_signatureless_bytes_without_path_require_explicit_suffix() -> None:
+    """无路径字节流不自动进入 tsv 解析，必须显式 file_suffix（与 csv 对称）。"""
+    payload = b"name\tcity\nAlice\tLondon\nBob\tParis\n" * 5
+
+    assert guess_suffix_by_bytes(payload) == "txt"
+
+
+def test_strong_content_signature_overrides_tsv_extension_fallback() -> None:
+    """RTF 强内容签名必须覆盖 .tsv 的无签名扩展名兜底。"""
+    payload = b"\xef\xbb\xbf \r\n{\\RTF1\\ANSI body}"
+
+    assert guess_suffix_by_bytes(payload, "disguised.tsv") == "rtf"
+
+
+def test_tsv_path_detection_returns_independent_suffix(tmp_path: Path) -> None:
+    """真实 .tsv 路径按独立后缀识别。"""
+    tsv_path = tmp_path / "demo.tsv"
+    tsv_path.write_bytes("姓名\t年龄\n张三\t30\n李四\t40\n".encode())
+
+    assert guess_suffix_by_path(tsv_path) == "tsv"
+
+
+def test_tsv_uses_shared_csv_engine_and_records_independent_file_suffix() -> None:
+    """tsv 复用 CSV 引擎与自动分隔符嗅探，并如实记录独立 file_suffix。"""
+    payload = "姓名\t年龄\n张三\t30\n李四\t40\n".encode()
+    middle, _ = analyze_native_test_document(payload, file_suffix="tsv")
+
+    assert middle.metadata.file_suffix == "tsv"
+    assert len(middle.pages) == 1
+    markdown = render_markdown(middle)
+    assert "张三" in markdown
+    assert "30" in markdown