def generate_test_segments_func(elems):
"""Roughly generate test segments.
This function isn't totally robust, but good enough
for testing. Use with caution.
"""
buff = []
raw_file = "".join(elems)
templated_file = TemplatedFile.from_string(raw_file)
idx = 0
for elem in elems:
if elem == "<indent>":
buff.append(
Indent(pos_marker=PositionMarker.from_point(
idx, idx, templated_file)))
continue
elif elem == "<dedent>":
buff.append(
Dedent(pos_marker=PositionMarker.from_point(
idx, idx, templated_file)))
continue
seg_kwargs = {}
if set(elem) <= {" ", "\t"}:
SegClass = WhitespaceSegment
elif set(elem) <= {"\n"}:
SegClass = NewlineSegment
elif elem == "(":
SegClass = SymbolSegment
seg_kwargs = {"name": "bracket_open"}
elif elem == ")":
SegClass = SymbolSegment
seg_kwargs = {"name": "bracket_close"}
elif elem.startswith("--"):
SegClass = CommentSegment
seg_kwargs = {"name": "inline_comment"}
elif elem.startswith('"'):
SegClass = CodeSegment
seg_kwargs = {"name": "double_quote"}
elif elem.startswith("'"):
SegClass = CodeSegment
seg_kwargs = {"name": "single_quote"}
else:
SegClass = CodeSegment
# Set a none position marker which we'll realign at the end.
buff.append(
SegClass(raw=elem,
pos_marker=PositionMarker(
slice(idx, idx + len(elem)),
slice(idx, idx + len(elem)),
templated_file,
),
**seg_kwargs))
idx += len(elem)
return tuple(buff)
def elements_to_segments(
self, elements: List[TemplateElement],
templated_file: TemplatedFile) -> Tuple[RawSegment, ...]:
"""Convert a tuple of lexed elements into a tuple of segments."""
# Working buffer to build up segments
segment_buffer: List[RawSegment] = []
lexer_logger.info("Elements to Segments.")
# Get the templated slices to re-insert tokens for them
source_only_slices = templated_file.source_only_slices()
lexer_logger.info("Source-only slices: %s", source_only_slices)
# Now work out source slices, and add in template placeholders.
for element in elements:
# Calculate Source Slice
source_slice = templated_file.templated_slice_to_source_slice(
element.template_slice)
# The calculated source slice will include any source only slices.
# We should consider all of them in turn to see whether we can
# insert them.
for source_only_slice in source_only_slices:
# If it's later in the source, stop looking. Any later
# ones *also* won't match.
if source_only_slice.source_idx > source_slice.start:
break
# Is there a templated section within this source slice?
# If there is then for some reason I can't quite explain,
# it will always be at the start of the section. This is
# very convenient beause it means we'll always have the
# start and end of it in a definite position. This makes
# slicing and looping much easier.
elif source_only_slice.source_idx == source_slice.start:
lexer_logger.debug(
"Found templated section! %s, %s, %s",
source_only_slice.source_slice(),
source_only_slice.slice_type,
element.template_slice.start,
)
# Calculate a slice for any placeholders
placeholder_source_slice = slice(
source_slice.start, source_only_slice.end_source_idx())
# Adjust the source slice accordingly.
source_slice = slice(source_only_slice.end_source_idx(),
source_slice.stop)
# TODO: Readjust this to remove .when once ProtoSegment is in.
# Add segments as appropriate.
# If it's a block end, add a dedent.
if source_only_slice.slice_type in ("block_end",
"block_mid"):
segment_buffer.append(
Dedent.when(template_blocks_indent=True)(
pos_marker=PositionMarker.from_point(
placeholder_source_slice.start,
element.template_slice.start,
templated_file,
)))
# Always add a placeholder
segment_buffer.append(
TemplateSegment(
pos_marker=PositionMarker(
placeholder_source_slice,
slice(
element.template_slice.start,
element.template_slice.start,
),
templated_file,
),
source_str=source_only_slice.raw,
block_type=source_only_slice.slice_type,
))
# If it's a block end, add a dedent.
if source_only_slice.slice_type in ("block_start",
"block_mid"):
segment_buffer.append(
Indent.when(template_blocks_indent=True)(
pos_marker=PositionMarker.from_point(
placeholder_source_slice.stop,
element.template_slice.start,
templated_file,
)))
# Add the actual segment
segment_buffer.append(
element.to_segment(pos_marker=PositionMarker(
source_slice,
element.template_slice,
templated_file,
), ))
# Convert to tuple before return
return tuple(segment_buffer)
def elements_to_segments(
self, elements: List[TemplateElement],
templated_file: TemplatedFile) -> Tuple[RawSegment, ...]:
"""Convert a tuple of lexed elements into a tuple of segments."""
# Working buffer to build up segments
segment_buffer: List[RawSegment] = []
lexer_logger.info("Elements to Segments.")
# Get the templated slices to re-insert tokens for them
source_only_slices = templated_file.source_only_slices()
lexer_logger.info("Source-only slices: %s", source_only_slices)
stash_source_slice, last_source_slice = None, None
# Now work out source slices, and add in template placeholders.
for idx, element in enumerate(elements):
# Calculate Source Slice
if idx != 0:
last_source_slice = stash_source_slice
source_slice = templated_file.templated_slice_to_source_slice(
element.template_slice)
stash_source_slice = source_slice
# Output the slice as we lex.
lexer_logger.debug(
" %s, %s, %s, %r",
idx,
element,
source_slice,
templated_file.templated_str[element.template_slice],
)
# The calculated source slice will include any source only slices.
# We should consider all of them in turn to see whether we can
# insert them.
so_slices = []
# Only look for source only slices if we've got a new source slice to
# avoid unnecessary duplication.
if last_source_slice != source_slice:
for source_only_slice in source_only_slices:
# If it's later in the source, stop looking. Any later
# ones *also* won't match.
if source_only_slice.source_idx >= source_slice.stop:
break
elif source_only_slice.source_idx >= source_slice.start:
so_slices.append(source_only_slice)
if so_slices:
lexer_logger.debug(" Collected Source Only Slices")
for so_slice in so_slices:
lexer_logger.debug(" %s", so_slice)
# Calculate some things which will be useful
templ_str = templated_file.templated_str[
element.template_slice]
source_str = templated_file.source_str[source_slice]
# For reasons which aren't entirely clear right now, if there is
# an included literal, it will always be at the end. Let's see if it's
# there.
if source_str.endswith(templ_str):
existing_len = len(templ_str)
else:
existing_len = 0
# Calculate slices
placeholder_slice = slice(source_slice.start,
source_slice.stop - existing_len)
placeholder_str = source_str[:-existing_len]
source_slice = slice(source_slice.stop - existing_len,
source_slice.stop)
# If it doesn't manage to extract a placeholder string from the source
# just concatenate the source only strings. There is almost always
# only one of them.
if not placeholder_str:
placeholder_str = "".join(s.raw for s in so_slices)
lexer_logger.debug(
" Overlap Length: %s. PS: %s, LS: %s, p_str: %r, templ_str: %r",
existing_len,
placeholder_slice,
source_slice,
placeholder_str,
templ_str,
)
# Caluculate potential indent/dedent
block_slices = sum(
s.slice_type.startswith("block_") for s in so_slices)
block_balance = sum(s.slice_type == "block_start"
for s in so_slices) - sum(
s.slice_type == "block_end"
for s in so_slices)
lead_dedent = so_slices[0].slice_type in ("block_end",
"block_mid")
trail_indent = so_slices[-1].slice_type in ("block_start",
"block_mid")
add_indents = self.config.get("template_blocks_indent",
"indentation")
lexer_logger.debug(
" Block Slices: %s. Block Balance: %s. Lead: %s, Trail: %s, Add: %s",
block_slices,
block_balance,
lead_dedent,
trail_indent,
add_indents,
)
# Add a dedent if appropriate.
if lead_dedent and add_indents:
lexer_logger.debug(" DEDENT")
segment_buffer.append(
Dedent(pos_marker=PositionMarker.from_point(
placeholder_slice.start,
element.template_slice.start,
templated_file,
)))
# Always add a placeholder
segment_buffer.append(
TemplateSegment(
pos_marker=PositionMarker(
placeholder_slice,
slice(
element.template_slice.start,
element.template_slice.start,
),
templated_file,
),
source_str=placeholder_str,
block_type=so_slices[0].slice_type
if len(so_slices) == 1 else "compound",
))
lexer_logger.debug(" Placholder: %s, %r",
segment_buffer[-1], placeholder_str)
# Add a dedent if appropriate.
if trail_indent and add_indents:
lexer_logger.debug(" INDENT")
segment_buffer.append(
Indent(pos_marker=PositionMarker.from_point(
placeholder_slice.stop,
element.template_slice.start,
templated_file,
)))
# Add the actual segment
segment_buffer.append(
element.to_segment(pos_marker=PositionMarker(
source_slice,
element.template_slice,
templated_file,
), ))
# Convert to tuple before return
return tuple(segment_buffer)
def generate_test_segments_func(elems):
"""Roughly generate test segments.
This function isn't totally robust, but good enough
for testing. Use with caution.
"""
buff = []
raw_file = "".join(elems)
templated_file = TemplatedFile.from_string(raw_file)
idx = 0
for elem in elems:
if elem == "<indent>":
buff.append(
Indent(pos_marker=PositionMarker.from_point(
idx, idx, templated_file)))
continue
elif elem == "<dedent>":
buff.append(
Dedent(pos_marker=PositionMarker.from_point(
idx, idx, templated_file)))
continue
if set(elem) <= {" ", "\t"}:
cls = RawSegment.make(" ",
name="whitespace",
type="whitespace",
_is_code=False)
elif set(elem) <= {"\n"}:
cls = RawSegment.make("\n",
name="newline",
type="newline",
_is_code=False)
elif elem == "(":
cls = RawSegment.make("(", name="bracket_open")
elif elem == ")":
cls = RawSegment.make(")", name="bracket_close")
elif elem.startswith("--"):
cls = RawSegment.make("--",
name="inline_comment",
_is_code=False)
elif elem.startswith('"'):
cls = RawSegment.make('"', name="double_quote")
elif elem.startswith("'"):
cls = RawSegment.make("'", name="single_quote")
else:
cls = RawSegment.make("")
# Set a none position marker which we'll realign at the end.
buff.append(
cls(
elem,
pos_marker=PositionMarker(
slice(idx, idx + len(elem)),
slice(idx, idx + len(elem)),
templated_file,
),
))
idx += len(elem)
return tuple(buff)
def elements_to_segments(
self, elements: List[TemplateElement],
templated_file: TemplatedFile) -> Tuple[RawSegment, ...]:
"""Convert a tuple of lexed elements into a tuple of segments."""
# Working buffer to build up segments
segment_buffer: List[RawSegment] = []
lexer_logger.info("Elements to Segments.")
# Get the templated slices to re-insert tokens for them
source_only_slices = templated_file.source_only_slices()
lexer_logger.info("Source-only slices: %s", source_only_slices)
stash_source_slice, last_source_slice = None, None
# Now work out source slices, and add in template placeholders.
for idx, element in enumerate(elements):
# Calculate Source Slice
if idx != 0:
last_source_slice = stash_source_slice
source_slice = templated_file.templated_slice_to_source_slice(
element.template_slice)
stash_source_slice = source_slice
# Output the slice as we lex.
lexer_logger.debug(
" %s, %s, %s, %r",
idx,
element,
source_slice,
templated_file.templated_str[element.template_slice],
)
# Detect when we've gone backward in the source.
# NOTE: If it's the _same_ slice then don't insert a marker
# because we're probably just within a single templated
# section.
if (last_source_slice
and last_source_slice.stop > source_slice.start
and last_source_slice != source_slice):
# If we have, insert a loop marker to reflect that.
lexer_logger.debug(
" Backward jump detected. Inserting Loop Marker")
segment_buffer.append(
TemplateLoop(pos_marker=PositionMarker.from_point(
last_source_slice.stop,
element.template_slice.start,
templated_file,
)))
# The calculated source slice will include any source only slices.
# We should consider all of them in turn to see whether we can
# insert them.
so_slices = []
# Only look for source only slices if we've got a new source slice to
# avoid unnecessary duplication.
if last_source_slice != source_slice:
for source_only_slice in source_only_slices:
# If it's later in the source, stop looking. Any later
# ones *also* won't match.
if source_only_slice.source_idx >= source_slice.stop:
break
elif source_only_slice.source_idx >= source_slice.start:
so_slices.append(source_only_slice)
if so_slices:
lexer_logger.debug(" Collected Source Only Slices")
for so_slice in so_slices:
lexer_logger.debug(" %s", so_slice)
# Calculate some things which will be useful
templ_str = templated_file.templated_str[
element.template_slice]
source_str = templated_file.source_str[source_slice]
# For reasons which aren't entirely clear right now, if there is
# an included literal, it will always be at the end. Let's see if it's
# there.
if source_str.endswith(templ_str):
existing_len = len(templ_str)
else:
existing_len = 0
# Calculate slices
placeholder_slice = slice(source_slice.start,
source_slice.stop - existing_len)
placeholder_str = source_str[:-existing_len]
source_slice = slice(source_slice.stop - existing_len,
source_slice.stop)
# If it doesn't manage to extract a placeholder string from the source
# just concatenate the source only strings. There is almost always
# only one of them.
if not placeholder_str:
placeholder_str = "".join(s.raw for s in so_slices)
# The Jinja templater sometimes returns source-only slices with
# gaps between. For example, in this section:
#
# {% else %}
# JOIN
# {{action}}_raw_effect_sizes
# USING
# ({{ states }})
# {% endif %}
#
# we might get {% else %} and {% endif %} slices, without the
# 4 lines between. This indicates those lines were not executed
# In this case, generate a placeholder where the skipped code is
# omitted but noted with a brief string, e.g.:
#
# "{% else %}... [103 unused template characters] ...{% endif %}".
#
# This is more readable -- it would be REALLY confusing for a
# placeholder to include code that wasn't even executed!!
if len(so_slices) >= 2:
has_gap = False
gap_placeholder_parts = []
last_slice = None
# For each slice...
for so_slice in so_slices:
# If it's not the first slice, was there a gap?
if last_slice:
end_last = last_slice.source_idx + len(
last_slice.raw)
chars_skipped = so_slice.source_idx - end_last
if chars_skipped:
# Yes, gap between last_slice and so_slice.
has_gap = True
# Generate a string documenting the gap.
if chars_skipped >= 10:
gap_placeholder_parts.append(
f"... [{chars_skipped} unused template "
"characters] ...")
else:
gap_placeholder_parts.append("...")
# Now add the slice's source.
gap_placeholder_parts.append(so_slice.raw)
last_slice = so_slice
if has_gap:
placeholder_str = "".join(gap_placeholder_parts)
lexer_logger.debug(
" Overlap Length: %s. PS: %s, LS: %s, p_str: %r, templ_str: %r",
existing_len,
placeholder_slice,
source_slice,
placeholder_str,
templ_str,
)
# Calculate potential indent/dedent
block_slices = sum(
s.slice_type.startswith("block_") for s in so_slices)
block_balance = sum(s.slice_type == "block_start"
for s in so_slices) - sum(
s.slice_type == "block_end"
for s in so_slices)
lead_dedent = so_slices[0].slice_type in ("block_end",
"block_mid")
trail_indent = so_slices[-1].slice_type in ("block_start",
"block_mid")
add_indents = self.config.get("template_blocks_indent",
"indentation")
lexer_logger.debug(
" Block Slices: %s. Block Balance: %s. Lead: %s, Trail: %s, "
"Add: %s",
block_slices,
block_balance,
lead_dedent,
trail_indent,
add_indents,
)
# Add a dedent if appropriate.
if lead_dedent and add_indents:
lexer_logger.debug(" DEDENT")
segment_buffer.append(
Dedent(pos_marker=PositionMarker.from_point(
placeholder_slice.start,
element.template_slice.start,
templated_file,
)))
# Always add a placeholder
segment_buffer.append(
TemplateSegment(
pos_marker=PositionMarker(
placeholder_slice,
slice(
element.template_slice.start,
element.template_slice.start,
),
templated_file,
),
source_str=placeholder_str,
block_type=so_slices[0].slice_type
if len(so_slices) == 1 else "compound",
))
lexer_logger.debug(" Placeholder: %s, %r",
segment_buffer[-1], placeholder_str)
# Add an indent if appropriate.
if trail_indent and add_indents:
lexer_logger.debug(" INDENT")
segment_buffer.append(
Indent(
is_template=True,
pos_marker=PositionMarker.from_point(
placeholder_slice.stop,
element.template_slice.start,
templated_file,
),
))
# Add the actual segment
segment_buffer.append(
element.to_segment(pos_marker=PositionMarker(
source_slice,
element.template_slice,
templated_file,
), ))
# Generate placeholders for any source-only slices that *follow*
# the last element. This happens, for example, if a Jinja templated
# file ends with "{% endif %}", and there's no trailing newline.
if idx == len(elements) - 1:
so_slices = [
so for so in source_only_slices
if so.source_idx >= source_slice.stop
]
for so_slice in so_slices:
segment_buffer.append(
TemplateSegment(
pos_marker=PositionMarker(
slice(so_slice.source_idx,
so_slice.end_source_idx()),
slice(
element.template_slice.stop,
element.template_slice.stop,
),
templated_file,
),
source_str=so_slice.raw,
block_type=so_slice.slice_type,
))
# Add an end of file marker
segment_buffer.append(
EndOfFile(pos_marker=segment_buffer[-1].pos_marker.
end_point_marker() if segment_buffer else PositionMarker.
from_point(0, 0, templated_file)))
# Convert to tuple before return
return tuple(segment_buffer)
