@ai-matrx/content-ir 0.19.7 → 0.19.9

package/dist/source.d.cts

@@ -632,9 +632,12 @@ declare function pairDisplayMath(text: string): Map<number, number>;
  * - A cell's `\|` shows as `|` — inside code spans too (spec example 200).
  * - A table starts at a header holding an unescaped pipe whose NEXT line is a
  *   delimiter row (`--- | :-:`, edge pipes optional); a header without a leading
- *   pipe must match the delimiter row's width. It runs to a blank line or a
- *   line that starts another block (`- x`, `> y`, `# z`, `1.`, a fence, a rule);
- *   every other line is a row, with or without a pipe (spec example 202).
+ *   pipe must match the delimiter row's width. It runs to a blank line, a
+ *   line that starts another block (`- x`, `> y`, `# z`, `1.`, a fence, a rule,
+ *   an HTML block — CommonMark 4.6's seven start conditions), or a line indented
+ *   four or more columns past the table's container (indented code); every other
+ *   line is a row, with or without a pipe (spec example 202), a lone `|` included
+ *   (an empty row).
  *
  * Pure TS, no dependencies.
  */
@@ -644,6 +647,18 @@ declare function splitRowSegments(line: string): string[];
 declare function rowCells(line: string): string[];
 /** GFM's table-level unescape: a cell's `\|` is a pipe, in code spans too. */
 declare function unescapeCellPipes(cell: string): string;
+/** The trimmed line starts an HTML block (any of CommonMark 4.6's seven conditions). */
+declare function startsHtmlBlock(trimmed: string): boolean;
+/** The column a line's content starts at: spaces count 1, a tab runs to the next multiple of 4. */
+declare function lineIndent(line: string): number;
+/**
+ * The content column of the container a table at `start` sits in: the list item
+ * whose content the header line belongs to (the nearest list-marker line above,
+ * past blank and continuation lines, whose content column is at or left of the
+ * header's), or 0 at top level. Indentation past it by 4+ columns is indented
+ * code, which ends the table.
+ */
+declare function tableContainerIndent(lines: readonly string[], start: number): number;
 /** A GFM delimiter row: every cell `---`, `:--`, `--:` or `:-:`, with at least one pipe. */
 declare function isGfmDelimiterRow(line: string): boolean;
 /** A pipe-led row: `|` first, and another `|` after it. */
@@ -652,11 +667,13 @@ declare function isPipeLedRow(line: string): boolean;
 declare function startsPipelessTable(header: string, delimiter: string | undefined): boolean;
 /**
  * A line that continues an open table — GFM's rule (spec example 202): the table
- * runs to the first blank line or the first line that starts another block; any
+ * runs to the first blank line, the first line that starts another block (an
+ * HTML block included), or the first line indented 4+ columns past the table's
+ * container (`containerIndent`, from `tableContainerIndent`; indented code). Any
  * other line is a row, even one without a pipe (`bar` → a one-cell row) and a
- * lone `|` (an empty row). verify-RC-B4 R7-2.
+ * lone `|` (an empty row). verify-RC-B4 R7-2 / round 9.
  */
-declare function continuesTable(line: string): boolean;
+declare function continuesTable(line: string, containerIndent?: number): boolean;
 /** Line `index` opens a table: a pipe-led row, or a pipe-less GFM header over its delimiter row. */
 declare function opensTable(lines: readonly string[], index: number): boolean;
 /** A WHOLE table header at `index`: a table opens here AND its next line is a GFM delimiter row. */
@@ -666,4 +683,4 @@ declare function findTableStart(lines: readonly string[], from?: number): number
 /** Index just past the table whose header is at `start` (header, delimiter, then continuation rows). */
 declare function findTableEnd(lines: readonly string[], start: number): number;
 
-export { AFTER_WINDOW, BacktickRuns, type BlockIslandType, CALLOUT_NAMES, type CodePointIndex, type CodeRange, type DisturbedIsland, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, type FenceEnd, type FenceLineKind, type FenceLines, type FenceOpener, FenceReader, type InlineIslandType, type InnerFenceLine, type IslandMeta, type IslandType, type MapOptions, type MappedPosition, type MappedRange, NESTING_FENCE_LANGUAGES, type ReadXmlTag, type SourceBlock, type SourceBlockKind, type SourceChange, type SourceEdit, type SourceIsland, type SourceSpan, SourceSpliceError, type SourceSpliceErrorCode, type SpliceIntegrity, type SpliceOptions, type SpliceResult, type TokenizeOptions, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsPipelessTable, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };
+export { AFTER_WINDOW, BacktickRuns, type BlockIslandType, CALLOUT_NAMES, type CodePointIndex, type CodeRange, type DisturbedIsland, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, type FenceEnd, type FenceLineKind, type FenceLines, type FenceOpener, FenceReader, type InlineIslandType, type InnerFenceLine, type IslandMeta, type IslandType, type MapOptions, type MappedPosition, type MappedRange, NESTING_FENCE_LANGUAGES, type ReadXmlTag, type SourceBlock, type SourceBlockKind, type SourceChange, type SourceEdit, type SourceIsland, type SourceSpan, SourceSpliceError, type SourceSpliceErrorCode, type SpliceIntegrity, type SpliceOptions, type SpliceResult, type TokenizeOptions, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, lineIndent, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsHtmlBlock, startsPipelessTable, tableContainerIndent, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };

package/dist/source.d.ts

@@ -632,9 +632,12 @@ declare function pairDisplayMath(text: string): Map<number, number>;
  * - A cell's `\|` shows as `|` — inside code spans too (spec example 200).
  * - A table starts at a header holding an unescaped pipe whose NEXT line is a
  *   delimiter row (`--- | :-:`, edge pipes optional); a header without a leading
- *   pipe must match the delimiter row's width. It runs to a blank line or a
- *   line that starts another block (`- x`, `> y`, `# z`, `1.`, a fence, a rule);
- *   every other line is a row, with or without a pipe (spec example 202).
+ *   pipe must match the delimiter row's width. It runs to a blank line, a
+ *   line that starts another block (`- x`, `> y`, `# z`, `1.`, a fence, a rule,
+ *   an HTML block — CommonMark 4.6's seven start conditions), or a line indented
+ *   four or more columns past the table's container (indented code); every other
+ *   line is a row, with or without a pipe (spec example 202), a lone `|` included
+ *   (an empty row).
  *
  * Pure TS, no dependencies.
  */
@@ -644,6 +647,18 @@ declare function splitRowSegments(line: string): string[];
 declare function rowCells(line: string): string[];
 /** GFM's table-level unescape: a cell's `\|` is a pipe, in code spans too. */
 declare function unescapeCellPipes(cell: string): string;
+/** The trimmed line starts an HTML block (any of CommonMark 4.6's seven conditions). */
+declare function startsHtmlBlock(trimmed: string): boolean;
+/** The column a line's content starts at: spaces count 1, a tab runs to the next multiple of 4. */
+declare function lineIndent(line: string): number;
+/**
+ * The content column of the container a table at `start` sits in: the list item
+ * whose content the header line belongs to (the nearest list-marker line above,
+ * past blank and continuation lines, whose content column is at or left of the
+ * header's), or 0 at top level. Indentation past it by 4+ columns is indented
+ * code, which ends the table.
+ */
+declare function tableContainerIndent(lines: readonly string[], start: number): number;
 /** A GFM delimiter row: every cell `---`, `:--`, `--:` or `:-:`, with at least one pipe. */
 declare function isGfmDelimiterRow(line: string): boolean;
 /** A pipe-led row: `|` first, and another `|` after it. */
@@ -652,11 +667,13 @@ declare function isPipeLedRow(line: string): boolean;
 declare function startsPipelessTable(header: string, delimiter: string | undefined): boolean;
 /**
  * A line that continues an open table — GFM's rule (spec example 202): the table
- * runs to the first blank line or the first line that starts another block; any
+ * runs to the first blank line, the first line that starts another block (an
+ * HTML block included), or the first line indented 4+ columns past the table's
+ * container (`containerIndent`, from `tableContainerIndent`; indented code). Any
  * other line is a row, even one without a pipe (`bar` → a one-cell row) and a
- * lone `|` (an empty row). verify-RC-B4 R7-2.
+ * lone `|` (an empty row). verify-RC-B4 R7-2 / round 9.
  */
-declare function continuesTable(line: string): boolean;
+declare function continuesTable(line: string, containerIndent?: number): boolean;
 /** Line `index` opens a table: a pipe-led row, or a pipe-less GFM header over its delimiter row. */
 declare function opensTable(lines: readonly string[], index: number): boolean;
 /** A WHOLE table header at `index`: a table opens here AND its next line is a GFM delimiter row. */
@@ -666,4 +683,4 @@ declare function findTableStart(lines: readonly string[], from?: number): number
 /** Index just past the table whose header is at `start` (header, delimiter, then continuation rows). */
 declare function findTableEnd(lines: readonly string[], start: number): number;
 
-export { AFTER_WINDOW, BacktickRuns, type BlockIslandType, CALLOUT_NAMES, type CodePointIndex, type CodeRange, type DisturbedIsland, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, type FenceEnd, type FenceLineKind, type FenceLines, type FenceOpener, FenceReader, type InlineIslandType, type InnerFenceLine, type IslandMeta, type IslandType, type MapOptions, type MappedPosition, type MappedRange, NESTING_FENCE_LANGUAGES, type ReadXmlTag, type SourceBlock, type SourceBlockKind, type SourceChange, type SourceEdit, type SourceIsland, type SourceSpan, SourceSpliceError, type SourceSpliceErrorCode, type SpliceIntegrity, type SpliceOptions, type SpliceResult, type TokenizeOptions, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsPipelessTable, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };
+export { AFTER_WINDOW, BacktickRuns, type BlockIslandType, CALLOUT_NAMES, type CodePointIndex, type CodeRange, type DisturbedIsland, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, type FenceEnd, type FenceLineKind, type FenceLines, type FenceOpener, FenceReader, type InlineIslandType, type InnerFenceLine, type IslandMeta, type IslandType, type MapOptions, type MappedPosition, type MappedRange, NESTING_FENCE_LANGUAGES, type ReadXmlTag, type SourceBlock, type SourceBlockKind, type SourceChange, type SourceEdit, type SourceIsland, type SourceSpan, SourceSpliceError, type SourceSpliceErrorCode, type SpliceIntegrity, type SpliceOptions, type SpliceResult, type TokenizeOptions, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, lineIndent, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsHtmlBlock, startsPipelessTable, tableContainerIndent, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };

package/dist/source.js

@@ -2625,14 +2625,61 @@ function rowCells(line) {
   const cells = splitRowSegments(line).map((cell) => cell.trim());
   if (cells.length > 1 && cells[0] === "") cells.shift();
   if (cells.length > 1 && cells[cells.length - 1] === "") cells.pop();
-  else if (cells.length === 1 && cells[0] === "") cells.pop();
+  else if (cells.length === 1 && cells[0] === "" && !line.includes("|")) cells.pop();
   return cells;
 }
 function unescapeCellPipes(cell) {
   return cell.replace(/\\\|/g, "|");
 }
 var BLOCK_START = /^(?:[-+*](?:\s|$)|\d{1,9}[.)](?:\s|$)|>|#{1,6}(?:\s|$)|(?:-\s*){3,}$|(?:\*\s*){3,}$|(?:_\s*){3,}$)/;
-var startsBlock = (trimmed) => BLOCK_START.test(trimmed) || opensStrictFence(trimmed);
+var startsBlock = (trimmed) => BLOCK_START.test(trimmed) || opensStrictFence(trimmed) || startsHtmlBlock(trimmed);
+var HTML_RAW = /^<(?:script|pre|style|textarea)(?:[\s>]|$)/i;
+var HTML_COMMENT_ETC = /^(?:<!--|<\?|<![A-Za-z]|<!\[CDATA\[)/;
+var HTML_BLOCK_NAMES = "address|article|aside|base|basefont|blockquote|body|caption|center|col|colgroup|dd|details|dialog|dir|div|dl|dt|fieldset|figcaption|figure|footer|form|frame|frameset|h[1-6]|head|header|hr|html|iframe|legend|li|link|main|menu|menuitem|nav|noframes|ol|optgroup|option|p|param|search|section|summary|table|tbody|td|tfoot|th|thead|title|tr|track|ul";
+var HTML_BLOCK_TAG = new RegExp(`^</?(?:${HTML_BLOCK_NAMES})(?:[\\s>]|/>|$)`, "i");
+var ATTRIBUTE = `\\s+[A-Za-z_:][A-Za-z0-9_.:-]*(?:\\s*=\\s*(?:[^\\s"'=<>\`]+|'[^']*'|"[^"]*"))?`;
+var HTML_COMPLETE_TAG = new RegExp(`^(?:<[A-Za-z][A-Za-z0-9-]*(?:${ATTRIBUTE})*\\s*/?>|</[A-Za-z][A-Za-z0-9-]*\\s*>)\\s*$`);
+function startsHtmlBlock(trimmed) {
+  if (!trimmed.startsWith("<")) return false;
+  return HTML_RAW.test(trimmed) || HTML_COMMENT_ETC.test(trimmed) || HTML_BLOCK_TAG.test(trimmed) || HTML_COMPLETE_TAG.test(trimmed);
+}
+function lineIndent(line) {
+  let column = 0;
+  for (const ch of line) {
+    if (ch === " ") column += 1;
+    else if (ch === "	") column += 4 - column % 4;
+    else break;
+  }
+  return column;
+}
+var LIST_MARKER = /^([ \t]*)([-+*]|\d{1,9}[.)])([ \t]+|$)/;
+function markerContentColumn(marker) {
+  const markerEnd = lineIndent(marker[1] ?? "") + (marker[2] ?? "").length;
+  const gap = marker[3] ? lineIndent(" ".repeat(markerEnd) + marker[3]) - markerEnd : 1;
+  return markerEnd + (gap >= 1 && gap <= 4 ? gap : 1);
+}
+function tableContainerIndent(lines, start) {
+  return tableContainerIndentOf(lines, start, false);
+}
+function tableContainerIndentOf(lines, start, ownItem) {
+  const own = ownItem ? LIST_MARKER.exec(lines[start] ?? "") : null;
+  if (own) return markerContentColumn(own);
+  const header = lineIndent(lines[start] ?? "");
+  if (header === 0) return 0;
+  for (let k = start - 1; k >= 0; k -= 1) {
+    const line = lines[k] ?? "";
+    if (!line.trim()) continue;
+    const indent = lineIndent(line);
+    const marker = LIST_MARKER.exec(line);
+    if (marker && indent < header) {
+      const content = markerContentColumn(marker);
+      if (content <= header) return content;
+      continue;
+    }
+    if (indent === 0) return 0;
+  }
+  return 0;
+}
 var DELIMITER_CELL = /^:?-+:?$/;
 function isGfmDelimiterRow(line) {
   const trimmed = line.trim();
@@ -2651,14 +2698,42 @@ function startsPipelessTable(header, delimiter) {
   if (delimiter === void 0 || !isGfmDelimiterRow(delimiter)) return false;
   return rowCells(head).length === rowCells(delimiter.trim()).length;
 }
-function continuesTable(line) {
+function continuesTable(line, containerIndent = 0) {
   const trimmed = line.trim();
   if (!trimmed) return false;
+  if (lineIndent(line) - containerIndent >= 4) return false;
   return trimmed.startsWith("|") || !startsBlock(trimmed);
 }
 function opensTable(lines, index) {
   const line = lines[index] ?? "";
-  return isPipeLedRow(line) || startsPipelessTable(line, lines[index + 1]);
+  if (!isPipeLedRow(line) && !startsPipelessTable(line, lines[index + 1])) return false;
+  return !isLazyContinuation(lines, index);
+}
+var QUOTE_LINE = /^ {0,3}>/;
+var THEMATIC_BREAK = /^(?:(?:-[ \t]*){3,}|(?:\*[ \t]*){3,}|(?:_[ \t]*){3,})$/;
+var breaksParagraph = (trimmed) => /^#{1,6}(?:\s|$)/.test(trimmed) || THEMATIC_BREAK.test(trimmed) || opensStrictFence(trimmed) || startsHtmlBlock(trimmed);
+function isLazyContinuation(lines, index) {
+  const header = lines[index] ?? "";
+  let first = -1;
+  for (let k = index - 1; k >= 0; k -= 1) {
+    const line = lines[k] ?? "";
+    const trimmed = line.trim();
+    if (!trimmed) break;
+    if (breaksParagraph(trimmed)) break;
+    const marker = LIST_MARKER.exec(line);
+    if (marker) {
+      const rest = line.slice(marker[0].length).trim();
+      if (!rest || breaksParagraph(rest)) break;
+      first = k;
+      break;
+    }
+    first = k;
+    if (QUOTE_LINE.test(line)) break;
+  }
+  if (first < 0) return false;
+  const start = lines[first] ?? "";
+  if (QUOTE_LINE.test(start)) return !QUOTE_LINE.test(header);
+  return lineIndent(header) < tableContainerIndentOf(lines, first, true);
 }
 function tableStartsAt(lines, index) {
   return opensTable(lines, index) && isGfmDelimiterRow(lines[index + 1] ?? "");
@@ -2668,11 +2743,12 @@ function findTableStart(lines, from = 0) {
   return -1;
 }
 function findTableEnd(lines, start) {
+  const container = tableContainerIndent(lines, start);
   let end = Math.min(lines.length, start + 2);
-  while (end < lines.length && continuesTable(lines[end] ?? "")) end += 1;
+  while (end < lines.length && continuesTable(lines[end] ?? "", container)) end += 1;
   return end;
 }
 
-export { AFTER_WINDOW, BacktickRuns, CALLOUT_NAMES, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, FenceReader, NESTING_FENCE_LANGUAGES, SourceSpliceError, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsPipelessTable, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };
+export { AFTER_WINDOW, BacktickRuns, CALLOUT_NAMES, FENCE_OPENER_WHITESPACE, FENCE_WHITESPACE, FenceReader, NESTING_FENCE_LANGUAGES, SourceSpliceError, XmlContainerTracker, blockAt, blockEdit, blockRangeEdit, buildCodePointIndex, classifyInnerFenceLine, closeFence, codeSpanText, continuesTable, fenceLineKinds, fenceNestsInnerFences, fenceOpenerOf, fenceParts, findCodeRanges, findInlineCodeSpans, findTableEnd, findTableStart, isEscapedBracketMath, isGfmDelimiterRow, isMathPlaceholder, isPipeLedRow, isSingleDollarMath, islandEdit, joinSource, lineIndent, listIslands, looksLikeDisplayMath, looksLikeOpenInlineMath, mapCodeRanges, mapInlineCodeSpans, mapPosition, mapRange, opensTable, pairDisplayMath, parseFenceOpener, pendingDisplayMathOpener, readXmlTag, rowCells, singleDollarAfter, singleDollarMathEnd, spliceSave, splitRowSegments, splitsSurrogatePair, startsHtmlBlock, startsPipelessTable, tableContainerIndent, tableStartsAt, toCodePointOffset, toUtf16Offset, tokenizeSource, trimFenceLine, unescapeCellPipes };
 //# sourceMappingURL=source.js.map
 //# sourceMappingURL=source.js.map