Skip to content

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

fix(pdf_to_md): keep numeric and labelled paragraphs between same-header tables

Claude4eb7b3d61f66
diff --git a/skills/ppt-master/scripts/source_to_md/pdf_to_md.py b/skills/ppt-master/scripts/source_to_md/pdf_to_md.py
index 666fbb36..ecb6dd77 100644
--- a/skills/ppt-master/scripts/source_to_md/pdf_to_md.py
+++ b/skills/ppt-master/scripts/source_to_md/pdf_to_md.py
@@ -478,6 +478,9 @@ NUMBER_RE = re.compile(r'[-+]?\d+(?:\.\d+)?')
MODEL_COLUMN_RE = re.compile(r'^[((]\d+[))]$')
TABLE_NOTE_PREFIX = '\u6ce8'
TABLE_CONTINUATION_Y_RATIO = 0.88
+# A printed page number left between split table parts; trusted only next to a
+# ``<!-- Page N -->`` marker, never on its own.
+PRINTED_PAGE_NUMBER_RE = re.compile(r'\d{1,4}')
TABLE_SCAN_BOTTOM_RATIO = 0.92
@@ -914,14 +917,6 @@ def _looks_like_outcome_row(row: list[str]) -> bool:
return len(short_values) == len(values)
-def _regression_group_labels(row: list[str], data_cols: int) -> list[str]:
- """Infer repeated group labels for common regression-table headings."""
- compact = "".join(row)
- if "总样本" in compact and "国有" in compact and "非国有" in compact and data_cols == 7:
- return ["总样本"] * 3 + ["国有企业"] * 2 + ["非国有企业"] * 2
- return [""] * data_cols
-
-
def _flatten_regression_header(rows: list[list[str]]) -> list[list[str]]:
"""Flatten multi-line regression headings into one Markdown header row."""
if len(rows) < 3:
@@ -942,10 +937,8 @@ def _flatten_regression_header(rows: list[list[str]]) -> list[list[str]]:
return [header] + rows[2:]
header_offset = 0
- groups = [""] * (len(rows[0]) - 1)
if not _looks_like_model_row(rows[0]) and _looks_like_model_row(rows[1]):
header_offset = 1
- groups = _regression_group_labels(rows[0], len(rows[1]) - 1)
if not _looks_like_model_row(rows[header_offset]):
return rows
@@ -957,8 +950,6 @@ def _flatten_regression_header(rows: list[list[str]]) -> list[list[str]]:
header = ["变量"]
for idx, model in enumerate(model_row[1:]):
pieces = []
- if idx < len(groups) and groups[idx]:
- pieces.append(groups[idx])
if model:
pieces.append(model)
if idx + 1 < len(outcome_row) and outcome_row[idx + 1]:
@@ -1439,24 +1430,24 @@ def _compatible_table_headers(first: list[str], second: list[str]) -> bool:
return first[0] == second[0] and _markdown_col_count(first[0]) > 2
-def _is_table_continuation_noise(line: str) -> bool:
- """Allow only page/header noise between split table parts."""
- text = line.strip()
- if not text:
- return True
- if text.startswith("<!-- Page ") and text.endswith("-->"):
- return True
- if re.fullmatch(r'\d+', text):
- return True
- if "重庆大学硕士学位论文" in text:
- return True
- continuation_labels = [
- "营改增",
- "深化增值税改革",
- "国有企业",
- "非国有企业",
- ]
- return any(label in text for label in continuation_labels)
+def _is_page_marker(text: str) -> bool:
+ """Return whether a stripped line is the converter's page-break marker."""
+ return text.startswith("<!-- Page ") and text.endswith("-->")
+
+
+def _is_table_continuation_gap(lines: list[str]) -> bool:
+ """Return whether only page-break furniture separates two table parts.
+
+ A bare number is a printed page number only when the gap crosses a
+ ``<!-- Page N -->`` marker; anywhere else it is content (a quantity, a year).
+ """
+ texts = [line.strip() for line in lines if line.strip()]
+ if not any(_is_page_marker(text) for text in texts):
+ return not texts
+ return all(
+ _is_page_marker(text) or PRINTED_PAGE_NUMBER_RE.fullmatch(text)
+ for text in texts
+ )
def _read_markdown_table_block(lines: list[str], start: int) -> tuple[list[str], int]:
@@ -1479,32 +1470,22 @@ def merge_markdown_continuation_tables(markdown: str) -> str:
index += 1
continue
- table, table_end = _read_markdown_table_block(lines, index)
- search = table_end
+ table, index = _read_markdown_table_block(lines, index)
while True:
- between_start = search
- while search < len(lines) and not _is_markdown_table_line(lines[search]):
- if not _is_table_continuation_noise(lines[search]):
- break
- search += 1
-
- if search >= len(lines) or not _is_markdown_table_line(lines[search]):
- break
- if any(
- not _is_table_continuation_noise(line)
- for line in lines[between_start:search]
- ):
+ next_start = index
+ while next_start < len(lines) and not _is_markdown_table_line(lines[next_start]):
+ next_start += 1
+ if next_start >= len(lines) or not _is_table_continuation_gap(lines[index:next_start]):
break
- next_table, next_end = _read_markdown_table_block(lines, search)
+ next_table, next_end = _read_markdown_table_block(lines, next_start)
if not _compatible_table_headers(table, next_table):
break
table.extend(next_table[2:])
- search = next_end
+ index = next_end
result.extend(table)
- index = search if search != table_end else table_end
return "\n".join(result)
diff --git a/skills/ppt-master/scripts/tests/test_pdf_table_continuation.py b/skills/ppt-master/scripts/tests/test_pdf_table_continuation.py
new file mode 100644
index 00000000..4a5d9c5f
--- /dev/null
+++ b/skills/ppt-master/scripts/tests/test_pdf_table_continuation.py
@@ -0,0 +1,56 @@
+#!/usr/bin/env python3
+"""Regression tests for merging PDF tables split across a page break.
+
+Only page-break furniture may separate two halves of one table; a numeric or
+labelled paragraph between same-header tables is content and must survive.
+"""
+
+from __future__ import annotations
+
+import sys
+import unittest
+from pathlib import Path
+
+SCRIPTS_DIR = Path(__file__).resolve().parents[1]
+BACKEND_DIR = SCRIPTS_DIR / "source_to_md"
+if str(BACKEND_DIR) not in sys.path:
+ sys.path.insert(0, str(BACKEND_DIR))
+
+import pdf_to_md # noqa: E402
+
+FIRST = "| A | B | C |\n|---|---|---|\n| one | two | three |"
+SECOND = "| A | B | C |\n|---|---|---|\n| four | five | six |"
+MERGED = "| A | B | C |\n|---|---|---|\n| one | two | three |\n| four | five | six |"
+
+
+def _merge(markdown: str) -> str:
+ return pdf_to_md.merge_markdown_continuation_tables(markdown)
+
+
+class ContinuationMergeTests(unittest.TestCase):
+ def test_page_break_with_printed_page_number_merges(self) -> None:
+ source = f"{FIRST}\n\n12\n\n<!-- Page 3 -->\n\n{SECOND}"
+ self.assertEqual(_merge(source), MERGED)
+
+ def test_blank_gap_merges(self) -> None:
+ self.assertEqual(_merge(f"{FIRST}\n\n{SECOND}"), MERGED)
+
+ def test_numeric_paragraph_without_page_break_is_kept(self) -> None:
+ for number in ("3000000", "2026", "42"):
+ with self.subTest(number=number):
+ source = f"{FIRST}\n{number}\n{SECOND}"
+ self.assertEqual(_merge(source), source)
+
+ def test_long_number_at_page_break_is_kept(self) -> None:
+ source = f"{FIRST}\n\n3000000\n\n<!-- Page 3 -->\n\n{SECOND}"
+ self.assertEqual(_merge(source), source)
+
+ def test_label_paragraph_between_tables_is_kept(self) -> None:
+ source = f"{FIRST}\n\n国有企业\n\n<!-- Page 3 -->\n\n{SECOND}"
+ self.assertEqual(_merge(source), source)
+
+ def test_lines_after_a_lone_table_are_kept(self) -> None:
+ for tail in ("\n\nSome text", "\n\n2026\n\nSome text", "\n\n<!-- Page 2 -->\n\n7"):
+ with self.subTest(tail=tail):
+ source = FIRST + tail
+ self.assertEqual(_merge(source), source)