Merge pull request #1294 from lark-parser/issue1293

propagate_positions: Bugfix + Better documentation

lark/lark.py

@@ -87,7 +87,9 @@ class LarkOptions(Serialize):
  transformer
  Applies the transformer to every parse tree (equivalent to applying it after the parse, but faster)
  propagate_positions
- Propagates (line, column, end_line, end_column) attributes into all tree branches.
+ Propagates positional attributes into the 'meta' attribute of all tree branches.
+ Sets attributes: (line, column, end_line, end_column, start_pos, end_pos,
+ container_line, container_column, container_end_line, container_end_column)
  Accepts ``False``, ``True``, or a callable, which will filter which nodes to ignore when propagating.
  maybe_placeholders
  When ``True``, the ``[]`` operator returns ``None`` when not matched.

lark/parse_tree_builder.py

@@ -50,6 +50,7 @@ def __call__(self, children):
 
  res_meta.container_line = getattr(first_meta, 'container_line', first_meta.line)
  res_meta.container_column = getattr(first_meta, 'container_column', first_meta.column)
+ res_meta.container_start_pos = getattr(first_meta, 'container_start_pos', first_meta.start_pos)
 
  last_meta = self._pp_get_meta(reversed(children))
  if last_meta is not None:
@@ -61,6 +62,7 @@ def __call__(self, children):
 
  res_meta.container_end_line = getattr(last_meta, 'container_end_line', last_meta.end_line)
  res_meta.container_end_column = getattr(last_meta, 'container_end_column', last_meta.end_column)
+ res_meta.container_end_pos = getattr(first_meta, 'container_end_pos', last_meta.end_pos)
 
  return res
 

lark/tree.py

@@ -47,7 +47,12 @@ class Tree(Generic[_Leaf_T]):
  data: The name of the rule or alias
  children: List of matched sub-rules and terminals
  meta: Line & Column numbers (if ``propagate_positions`` is enabled).
- meta attributes: line, column, start_pos, end_line, end_column, end_pos
+ meta attributes: (line, column, end_line, end_column, start_pos, end_pos,
+ container_line, container_column, container_end_line, container_end_column)
+ container_* attributes consider all symbols, including those that have been inlined in the tree.
+ For example, in the rule 'a: _A B _C', the regular attributes will mark the start and end of B,
+ but the container_* attributes will also include _A and _C in the range. However, rules that
+ contain 'a' will consider it in full, including _A and _C for all attributes.
  """
 
  data: str