@fitlab-ai/agent-infra 0.8.5 → 0.9.0

package/lib/task/sections.ts

@@ -1,7 +1,469 @@
+import type { FrontmatterScalar } from './frontmatter.ts';
+
 function escapeRegExp(value: string): string {
   return value.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
 }
 
+type SectionMutationInput = {
+  aliases: readonly string[];
+  heading: string;
+  body: string;
+};
+
+type TableRowBase = {
+  kind: 'table-row';
+  sectionAliases: readonly string[];
+  columns: readonly string[];
+  keyColumn: string;
+  key: string;
+};
+
+type TableRowUpsertMutation = TableRowBase & {
+  action: 'upsert';
+  values: Readonly<Record<string, FrontmatterScalar>>;
+};
+
+type TableRowDeleteMutation = TableRowBase & { action: 'delete' };
+type TableRowMutation = TableRowUpsertMutation | TableRowDeleteMutation;
+
+type SectionMutationResult = {
+  content: string;
+  heading: string;
+  operation: 'create' | 'update';
+};
+
+type TableMutationResult = {
+  content: string;
+  section: string;
+  operation: 'insert' | 'update' | 'delete';
+};
+
+type ParsedTableRow = {
+  values: Readonly<Record<string, string>>;
+  sourceLine: number;
+};
+
+type ParsedTable = {
+  heading: string;
+  rows: readonly ParsedTableRow[];
+};
+
+type DocumentMutationErrorCode =
+  | 'TASK_DOCUMENT_INVALID'
+  | 'MUTATION_INVALID'
+  | 'TABLE_NOT_FOUND'
+  | 'TABLE_AMBIGUOUS'
+  | 'TABLE_DUPLICATE_KEY'
+  | 'TABLE_UNKNOWN_COLUMN'
+  | 'TABLE_KEY_COLUMN_IN_VALUES'
+  | 'TABLE_KEY_CONFLICT'
+  | 'TABLE_MISSING_COLUMN'
+  | 'TABLE_DELETE_VALUES_FORBIDDEN'
+  | 'TABLE_CELL_INVALID';
+
+class DocumentMutationError extends Error {
+  code: DocumentMutationErrorCode;
+
+  constructor(code: DocumentMutationErrorCode, message: string) {
+    super(message);
+    this.name = 'DocumentMutationError';
+    this.code = code;
+  }
+}
+
+function parseTable(
+  content: string,
+  input: { sectionAliases: readonly string[]; columns: readonly string[] }
+): ParsedTable | null {
+  validateAliases(input.sectionAliases, 'table section aliases');
+  if (
+    !Array.isArray(input.columns) || input.columns.length === 0 ||
+    input.columns.some((column) => !column || /[\r\n|]/.test(column)) ||
+    new Set(input.columns).size !== input.columns.length
+  ) {
+    throw new DocumentMutationError('MUTATION_INVALID', 'table columns are invalid');
+  }
+  const sections = matchingSections(content, input.sectionAliases);
+  if (sections.length === 0) return null;
+  if (sections.length > 1) {
+    throw new DocumentMutationError('TASK_DOCUMENT_INVALID', 'table section is ambiguous');
+  }
+  const section = sections[0]!;
+  const sectionLines = linesOf(content).filter(
+    (line) => line.start >= section.bodyStart && line.start < section.end
+  );
+  const firstContent = sectionLines.findIndex((line) => line.text.trim() !== '');
+  const tables: number[] = [];
+  for (let index = 0; index + 1 < sectionLines.length; index += 1) {
+    const header = splitTableCells(sectionLines[index]!.text);
+    const separator = splitTableCells(sectionLines[index + 1]!.text);
+    if (!header || !separator || header.cells.length !== input.columns.length) continue;
+    const names = header.cells.map((cell) => decodeCell(cell).trim());
+    if (!names.every((name, column) => name === input.columns[column])) continue;
+    if (
+      separator.cells.length === input.columns.length &&
+      separator.cells.every((cell) => /^\s*:?-{3,}:?\s*$/.test(cell))
+    ) tables.push(index);
+  }
+  if (tables.length === 0) {
+    throw new DocumentMutationError('TABLE_NOT_FOUND', 'matching table not found');
+  }
+  if (tables.length > 1) {
+    throw new DocumentMutationError('TABLE_AMBIGUOUS', 'matching table appears more than once');
+  }
+  const rows: ParsedTableRow[] = [];
+  const seen = new Set<string>();
+  for (let index = tables[0]! + 2; index < sectionLines.length; index += 1) {
+    const line = sectionLines[index]!;
+    if (!line.text.trim()) break;
+    const parsed = splitTableCells(line.text);
+    if (!parsed) break;
+    if (parsed.cells.length !== input.columns.length) {
+      throw new DocumentMutationError('TASK_DOCUMENT_INVALID', 'table row has the wrong column count');
+    }
+    const values = Object.fromEntries(input.columns.map((column, columnIndex) => [
+      column,
+      decodeCell(parsed.cells[columnIndex]!).trim()
+    ]));
+    const key = values[input.columns[0]!]!;
+    if (seen.has(key)) {
+      throw new DocumentMutationError('TABLE_DUPLICATE_KEY', `duplicate table key '${key}'`);
+    }
+    seen.add(key);
+    rows.push({ values, sourceLine: index - Math.max(firstContent, 0) });
+  }
+  return { heading: section.heading, rows };
+}
+
+type Line = {
+  start: number;
+  contentEnd: number;
+  end: number;
+  text: string;
+  eol: string;
+};
+
+function linesOf(content: string): Line[] {
+  const lines: Line[] = [];
+  let start = 0;
+  while (start < content.length) {
+    const newline = content.indexOf('\n', start);
+    const end = newline === -1 ? content.length : newline + 1;
+    const hasCr = newline !== -1 && content[newline - 1] === '\r';
+    const contentEnd = newline === -1 ? content.length : newline - (hasCr ? 1 : 0);
+    lines.push({
+      start,
+      contentEnd,
+      end,
+      text: content.slice(start, contentEnd),
+      eol: newline === -1 ? '' : hasCr ? '\r\n' : '\n'
+    });
+    start = end;
+  }
+  return lines;
+}
+
+function documentEol(content: string): '\n' | '\r\n' {
+  return content.includes('\r\n') ? '\r\n' : '\n';
+}
+
+function matchingSections(content: string, aliases: readonly string[]) {
+  const allowed = new Set(aliases);
+  return linesOf(content)
+    .map((line, index, lines) => {
+      const match = /^##\s+(.+?)\s*$/.exec(line.text);
+      if (!match?.[1] || !allowed.has(match[1])) return null;
+      let end = content.length;
+      for (let i = index + 1; i < lines.length; i += 1) {
+        if (/^##\s+/.test(lines[i]!.text)) {
+          end = lines[i]!.start;
+          break;
+        }
+      }
+      return { heading: match[1], line, bodyStart: line.end, end };
+    })
+    .filter((match): match is NonNullable<typeof match> => match !== null);
+}
+
+function validateAliases(aliases: readonly string[], name: string): void {
+  if (!Array.isArray(aliases) || aliases.length === 0 || aliases.some((alias) => !alias || /[\r\n]/.test(alias))) {
+    throw new DocumentMutationError('MUTATION_INVALID', `${name} must contain headings`);
+  }
+  if (new Set(aliases).size !== aliases.length) {
+    throw new DocumentMutationError('MUTATION_INVALID', `${name} must be unique`);
+  }
+}
+
+function upsertSection(content: string, input: SectionMutationInput): SectionMutationResult {
+  validateAliases(input.aliases, 'section aliases');
+  if (!input.heading || /[\r\n]/.test(input.heading) || typeof input.body !== 'string') {
+    throw new DocumentMutationError('MUTATION_INVALID', 'section heading and body are invalid');
+  }
+  const matches = matchingSections(content, input.aliases);
+  if (matches.length > 1) {
+    throw new DocumentMutationError('TASK_DOCUMENT_INVALID', 'section aliases match more than once');
+  }
+  const eol = documentEol(content);
+  const normalizedBody = input.body.replace(/\r?\n/g, eol).replace(/(?:\r?\n)+$/, '');
+  if (matches.length === 0) {
+    const separator = content.length === 0
+      ? ''
+      : content.endsWith(`${eol}${eol}`)
+        ? ''
+        : content.endsWith(eol)
+          ? eol
+          : `${eol}${eol}`;
+    const body = normalizedBody ? `${eol}${eol}${normalizedBody}` : '';
+    return {
+      content: `${content}${separator}## ${input.heading}${body}${eol}`,
+      heading: input.heading,
+      operation: 'create'
+    };
+  }
+  const match = matches[0]!;
+  const existingBody = content
+    .slice(match.bodyStart, match.end)
+    .replace(/^(?:\r?\n)+/, '')
+    .replace(/(?:\r?\n)+$/, '')
+    .replace(/\r?\n/g, eol);
+  if (existingBody === normalizedBody) {
+    return { content, heading: match.heading, operation: 'update' };
+  }
+  const body = normalizedBody ? `${eol}${normalizedBody}${eol}${eol}` : eol;
+  return {
+    content: content.slice(0, match.bodyStart) + body + content.slice(match.end),
+    heading: match.heading,
+    operation: 'update'
+  };
+}
+
+function splitTableCells(line: string): { prefix: string; suffix: string; cells: string[] } | null {
+  const first = line.indexOf('|');
+  const last = line.lastIndexOf('|');
+  if (first === -1 || first === last || line.slice(0, first).trim() || line.slice(last + 1).trim()) {
+    return null;
+  }
+  const inner = line.slice(first + 1, last);
+  const cells: string[] = [];
+  let start = 0;
+  let backslashes = 0;
+  for (let i = 0; i < inner.length; i += 1) {
+    const char = inner[i]!;
+    if (char === '\\') {
+      backslashes += 1;
+      continue;
+    }
+    if (char === '|' && backslashes % 2 === 0) {
+      cells.push(inner.slice(start, i));
+      start = i + 1;
+    }
+    backslashes = 0;
+  }
+  if (backslashes % 2 === 1) {
+    throw new DocumentMutationError('TASK_DOCUMENT_INVALID', 'table cell has an unclosed escape');
+  }
+  cells.push(inner.slice(start));
+  return { prefix: line.slice(0, first), suffix: line.slice(last + 1), cells };
+}
+
+function decodeCell(raw: string): string {
+  let result = '';
+  for (let i = 0; i < raw.length; i += 1) {
+    if (raw[i] === '\\' && (raw[i + 1] === '\\' || raw[i + 1] === '|')) {
+      result += raw[i + 1];
+      i += 1;
+    } else {
+      result += raw[i];
+    }
+  }
+  return result;
+}
+
+function scalarCell(value: FrontmatterScalar): string {
+  if (typeof value === 'number' && !Number.isFinite(value)) {
+    throw new DocumentMutationError('TABLE_CELL_INVALID', 'table numbers must be finite');
+  }
+  if (
+    value !== null &&
+    typeof value !== 'string' &&
+    typeof value !== 'number' &&
+    typeof value !== 'boolean'
+  ) {
+    throw new DocumentMutationError('TABLE_CELL_INVALID', 'table cells must be scalar');
+  }
+  const raw = value === null ? '' : String(value);
+  if (/[\r\n]/.test(raw)) {
+    throw new DocumentMutationError('TABLE_CELL_INVALID', 'table cells cannot contain newlines');
+  }
+  return raw.replace(/\\/g, '\\\\').replace(/\|/g, '\\|');
+}
+
+function replaceCellToken(raw: string, value: string): string {
+  const firstContent = raw.search(/\S/);
+  if (firstContent === -1) {
+    const boundary = Math.floor(raw.length / 2);
+    return `${raw.slice(0, boundary)}${value}${raw.slice(boundary)}`;
+  }
+  let lastContent = raw.length - 1;
+  while (/\s/.test(raw[lastContent]!)) lastContent -= 1;
+  return `${raw.slice(0, firstContent)}${value}${raw.slice(lastContent + 1)}`;
+}
+
+function mutateTableRow(content: string, mutation: TableRowMutation): TableMutationResult {
+  validateAliases(mutation.sectionAliases, 'table section aliases');
+  if (
+    !Array.isArray(mutation.columns) ||
+    mutation.columns.length === 0 ||
+    mutation.columns.some((column) => !column || /[\r\n|]/.test(column)) ||
+    new Set(mutation.columns).size !== mutation.columns.length ||
+    !mutation.columns.includes(mutation.keyColumn) ||
+    typeof mutation.key !== 'string' ||
+    !mutation.key.trim() ||
+    /[\r\n]/.test(mutation.key)
+  ) {
+    throw new DocumentMutationError('MUTATION_INVALID', 'table mutation shape is invalid');
+  }
+  const normalizedKey = mutation.key.trim();
+  if (mutation.action !== 'upsert' && mutation.action !== 'delete') {
+    throw new DocumentMutationError('MUTATION_INVALID', 'table action is invalid');
+  }
+  if (mutation.action === 'delete' && 'values' in mutation) {
+    throw new DocumentMutationError('TABLE_DELETE_VALUES_FORBIDDEN', 'delete cannot include values');
+  }
+  if (
+    mutation.action === 'upsert' &&
+    (!mutation.values || typeof mutation.values !== 'object' || Array.isArray(mutation.values))
+  ) {
+    throw new DocumentMutationError('MUTATION_INVALID', 'upsert values must be an object');
+  }
+
+  const sections = matchingSections(content, mutation.sectionAliases);
+  if (sections.length !== 1) {
+    throw new DocumentMutationError(
+      sections.length === 0 ? 'TABLE_NOT_FOUND' : 'TASK_DOCUMENT_INVALID',
+      sections.length === 0 ? 'table section not found' : 'table section is ambiguous'
+    );
+  }
+  const section = sections[0]!;
+  const sectionLines = linesOf(content).filter(
+    (line) => line.start >= section.bodyStart && line.start < section.end
+  );
+  const tables: { headerIndex: number; cells: string[] }[] = [];
+  for (let i = 0; i + 1 < sectionLines.length; i += 1) {
+    const header = splitTableCells(sectionLines[i]!.text);
+    const separator = splitTableCells(sectionLines[i + 1]!.text);
+    if (!header || !separator || header.cells.length !== mutation.columns.length) continue;
+    const names = header.cells.map((cell) => decodeCell(cell).trim());
+    if (!names.every((name, index) => name === mutation.columns[index])) continue;
+    if (
+      separator.cells.length !== mutation.columns.length ||
+      !separator.cells.every((cell) => /^\s*:?-{3,}:?\s*$/.test(cell))
+    ) continue;
+    tables.push({ headerIndex: i, cells: header.cells });
+  }
+  if (tables.length === 0) {
+    throw new DocumentMutationError('TABLE_NOT_FOUND', 'matching table not found');
+  }
+  if (tables.length > 1) {
+    throw new DocumentMutationError('TABLE_AMBIGUOUS', 'matching table appears more than once');
+  }
+
+  const table = tables[0]!;
+  const keyIndex = mutation.columns.indexOf(mutation.keyColumn);
+  let rowIndex = table.headerIndex + 2;
+  const rows: { line: Line; parsed: NonNullable<ReturnType<typeof splitTableCells>> }[] = [];
+  while (rowIndex < sectionLines.length) {
+    const line = sectionLines[rowIndex]!;
+    const parsed = splitTableCells(line.text);
+    if (!parsed) break;
+    if (parsed.cells.length !== mutation.columns.length) {
+      throw new DocumentMutationError('TASK_DOCUMENT_INVALID', 'table row has the wrong column count');
+    }
+    rows.push({ line, parsed });
+    rowIndex += 1;
+  }
+  const matches = rows.filter(
+    (row) => decodeCell(row.parsed.cells[keyIndex]!).trim() === normalizedKey
+  );
+  if (matches.length > 1) {
+    throw new DocumentMutationError('TABLE_DUPLICATE_KEY', `duplicate table key '${normalizedKey}'`);
+  }
+
+  if (mutation.action === 'delete') {
+    if (matches.length === 0) return { content, section: section.heading, operation: 'delete' };
+    const line = matches[0]!.line;
+    return {
+      content: content.slice(0, line.start) + content.slice(line.end),
+      section: section.heading,
+      operation: 'delete'
+    };
+  }
+
+  const values = mutation.values;
+  for (const [column, value] of Object.entries(values)) {
+    if (!mutation.columns.includes(column)) {
+      throw new DocumentMutationError('TABLE_UNKNOWN_COLUMN', `unknown table column '${column}'`);
+    }
+    if (column === mutation.keyColumn) {
+      const serialized = scalarCell(value);
+      throw new DocumentMutationError(
+        decodeCell(serialized).trim() === normalizedKey
+          ? 'TABLE_KEY_COLUMN_IN_VALUES'
+          : 'TABLE_KEY_CONFLICT',
+        `key column '${column}' must not appear in values`
+      );
+    }
+    scalarCell(value);
+  }
+
+  if (matches.length === 1) {
+    const row = matches[0]!;
+    const cells = [...row.parsed.cells];
+    for (const [column, value] of Object.entries(values)) {
+      const index = mutation.columns.indexOf(column);
+      cells[index] = replaceCellToken(cells[index]!, scalarCell(value));
+    }
+    const rebuilt = `${row.parsed.prefix}|${cells.join('|')}|${row.parsed.suffix}`;
+    return {
+      content:
+        content.slice(0, row.line.start) +
+        rebuilt +
+        row.line.eol +
+        content.slice(row.line.end),
+      section: section.heading,
+      operation: 'update'
+    };
+  }
+
+  const missing = mutation.columns.filter(
+    (column) => column !== mutation.keyColumn && !Object.hasOwn(values, column)
+  );
+  if (missing.length > 0) {
+    throw new DocumentMutationError(
+      'TABLE_MISSING_COLUMN',
+      `new table row is missing columns: ${missing.join(', ')}`
+    );
+  }
+  const newCells = mutation.columns.map((column) =>
+    column === mutation.keyColumn ? scalarCell(normalizedKey) : scalarCell(values[column]!)
+  );
+  const newLine = `| ${newCells.join(' | ')} |`;
+  const separatorLine = sectionLines[table.headerIndex + 1]!;
+  const lastLine = rows.at(-1)?.line ?? separatorLine;
+  const eol = lastLine.eol || documentEol(content);
+  const needsLeadingEol = lastLine.eol === '';
+  return {
+    content:
+      content.slice(0, lastLine.end) +
+      (needsLeadingEol ? eol : '') +
+      newLine +
+      eol +
+      content.slice(lastLine.end),
+    section: section.heading,
+    operation: 'insert'
+  };
+}
+
 /**
  * Return the body of the first `## {alias}` section (any alias matches), from
  * the heading line to the next `## ` heading or EOF. Lines are preserved
@@ -44,9 +506,9 @@ function findSectionHeading(content: string, aliases: string[]): string {
 /**
  * Return the body of the first `### {headingPrefix}` sub-section, from the
  * heading line (inclusive) to the next `### ` / `## ` heading or EOF. Used to
- * pull a single `### HD-N` human-decision detail block out of an artifact. The
- * prefix must be followed by a word boundary so `HD-1` does not match `HD-10`
- * (e.g. `### HD-1`, `### HD-1：标题`, `### HD-1 [needs-human-decision]`). Leading
+ * pull a single `### {ledger-id}` detail block out of an artifact. The prefix
+ * must be followed by a word boundary so `PL-1` does not match `PL-10`
+ * (e.g. `### PL-1`, `### PL-1：标题`, `### PL-1 [needs-human-decision]`). Leading
  * and trailing blank lines are trimmed. Returns '' when no match is present.
  */
 function extractSubSection(content: string, headingPrefix: string): string {
@@ -70,4 +532,24 @@ function extractSubSection(content: string, headingPrefix: string): string {
   return lines.slice(start, end).join('\n').replace(/^\n+/, '').replace(/\n+$/, '');
 }
 
-export { extractSection, findSectionHeading, extractSubSection };
+export {
+  extractSection,
+  findSectionHeading,
+  extractSubSection,
+  parseTable,
+  upsertSection,
+  mutateTableRow,
+  DocumentMutationError
+};
+export type {
+  SectionMutationInput,
+  SectionMutationResult,
+  TableRowBase,
+  TableRowUpsertMutation,
+  TableRowDeleteMutation,
+  TableRowMutation,
+  TableMutationResult,
+  DocumentMutationErrorCode,
+  ParsedTableRow,
+  ParsedTable
+};