from collections import namedtuple import html import re from typing import Dict, Optional, Any import pygments.formatters from literate.annot import Span, fill_annot, cut_annot from literate.file import File, Line, Diff from literate.points import Point, map_points, merge_points, \ annot_starts, annot_ends, cut_annot_at_points import literate.diff import literate.highlight import literate.marks def mark_class(f: File, node_id: int) -> str: '''Get the CSS class suffix to use for `node_id`'s mark indicator.''' added, removed, kept = f.mark_labels.get(node_id, ((), (), ())) if len(added) == 0 and len(removed) == 0: return 'kept' elif len(added) > 0 and len(removed) == 0 and len(kept) == 0: return 'ins' elif len(removed) > 0 and len(added) == 0 and len(kept) == 0: return 'del' else: return 'chg' def mark_desc(f: File, node_id: int) -> str: '''Get a user-facing description of the mark on `node_id`. The returned descriptions look like: item #123: added 'foo'; removed 'bar', 'baz'; kept 'target' ''' m = f.marks[node_id] if m.id == 0xffffffff: id_str = '#DUMMY' elif (m.id & 0x80000000) != 0: id_str = '#NEW-%d' % (m.id & ~0x80000000) else: id_str = '#%d' % m.id if m.name is not None and m.name != '': mark_str = '%s %s "%s"' % (m.kind, id_str, m.name) else: mark_str = '%s %s' % (m.kind, id_str) added, removed, kept = f.mark_labels.get(node_id, ((), (), ())) parts = [] if len(added) > 0: parts.append('added %s' % ', '.join(repr(l) for l in added)) if len(removed) > 0: parts.append('removed %s' % ', '.join(repr(l) for l in removed)) if len(kept) > 0: parts.append('kept %s' % ', '.join(repr(l) for l in kept)) return '%s: %s' % (mark_str, '; '.join(parts)) def render_line(line: Line, f: File, opts: Dict[str, Any]) -> str: '''Render HTML output for a single line of a file. `f` should be the file containing `line`.''' parts = [] # Helpers for adding common types of parts, such as span start/end tags. def mark_arrow(m, start): if start: sym = '▶' else: sym = '◀' parts.append('%s' % (mark_class(f, m), html.escape(mark_desc(f, m)), sym)) def emit_text(start, end): if end == len(line.text) and line.text.endswith('\n'): # Avoid emitting the line's trailing `\n`. It normally gets # ignored by the browser, but sometimes we emit a mark indicator # afterward, which we'd like to stay on the same line when # possible. end = end - 1 if end > start: parts.append(line.text[start : end]) def start_span(cls): parts.append('' % cls) def end_span(): parts.append('') # Build a list of "events", which are places where we should insert markup # tags into the text of the line. `events` is a list of `Point`s, each # labeled with details of the tag to be inserted at the point. # We rely on the bias of `merge_points` (when two input sequences have # points at the same position, the ones from the first input appear first) # to establish an ordering between different event types and ensure proper # nesting of HTML tags. # # The most interesting case here is for `line.intra`. The CSS for # intraline edit markers uses a `border` attribute, which means a pair of # adjacent intraline edit spans is visibly distinct from a single longer # span. (This is unlike `highlight`, whose CSS only sets text color.) We # deliberately break intraline edit spans around mark indicators, but # otherwise try to keep them intact. events = [] # Mark starts and ends. events = merge_points( map_points(line.mark_ends, lambda l: ('m_e', l)), events, map_points(line.mark_starts, lambda l: ('m_s', l)), ) # Intraline edits. if line.intra is not None and opts['diff-style'] != 'only-new': intra = cut_annot_at_points(line.intra, events) events = merge_points( map_points(annot_ends(intra), lambda l: ('i_e', l)), events, map_points(annot_starts(intra), lambda l: ('i_s', l)), ) # Syntax highlighting. highlight = cut_annot_at_points(line.highlight, events) events = merge_points( map_points(annot_ends(highlight), lambda l: ('hl_e', l)), events, map_points(annot_starts(highlight), lambda l: ('hl_s', l)), ) # Build up the output text. We emit special tags at each point in # `events`, and otherwise emit substrings from `line.text`. last_pos = 0 if line.hunk_start_marks: # We sort all starting labels descending and all ending labels # ascending, in hopes that higher-numbered labels enclose # lower-numbered ones. There's no guarantee, of course. for m in sorted(line.hunk_start_marks, reverse=True): mark_arrow(m, True) # These hunk-start mark indicators often appear before the # whitespace of indented lines. To avoid throwing off the # indentation, we let them "eat up" as much whitespace as is # available. Any start/end events in the eaten whitespace will be # pushed back until after the marks. (Hopefully there are not many # events inside a purely whitespace span.) if last_pos < len(line.text) and line.text[last_pos] == ' ': last_pos += 1 hl_mode = opts['highlight-mode'] for p in events: if p.pos > last_pos: emit_text(last_pos, p.pos) last_pos = p.pos kind, label = p.label if kind in 'm_s': for m in sorted(label, reverse=True): mark_arrow(m, True) elif kind == 'm_e': for m in sorted(label): mark_arrow(m, False) elif kind == 'i_s': start_span('diff-intra-%s' % label) elif kind == 'i_e': end_span() elif kind == 'hl_s': cls = literate.highlight.token_css_class(label, hl_mode) if cls is not None: start_span(cls) elif kind == 'hl_e': cls = literate.highlight.token_css_class(label, hl_mode) if cls is not None: end_span() if len(line.text) > last_pos: emit_text(last_pos, len(line.text)) if line.hunk_end_marks: for m in sorted(line.hunk_end_marks): mark_arrow(m, False) return ''.join(parts) def annotate_irrelevant(f: File, start: str, end: str): start_re = re.compile(start) end_re = re.compile(end) result = [] start_line = None for i, l in enumerate(f.lines): if start_line is None and start_re.match(l.text): start_line = i if start_line is not None and end_re.match(l.text): result.append(Span(start_line, i + 1, None)) start_line = None f.set_drop_irrelevant_lines(result) def prepare_files(files: [File]): '''Run single-file initialization steps on each file in `files`.''' for i, f in enumerate(files): print('preparing file %d (%s)' % (i, f.path)) literate.highlight.highlight_file(f) literate.marks.mark_file(f) def make_diff(f1: File, f2: File, opts: Dict[str, Any]) -> Diff: '''Construct a diff between two files, and run diff initialization steps.''' print(' diffing file %s' % f1.path) f1 = f1.copy() f2 = f2.copy() # These options get handled during diff construction because they can vary # across refactoring blocks, while each `File` in `prepare_files` is # typically used in 2 different blocks. if opts['irrelevant-start-regex'] and opts['irrelevant-end-regex']: start = opts['irrelevant-start-regex'] end = opts['irrelevant-end-regex'] annotate_irrelevant(f1, start, end) annotate_irrelevant(f2, start, end) d = literate.diff.diff_files(f1, f2) literate.marks.init_mark_labels(d) literate.marks.init_keep_mark_lines(d) literate.diff.build_diff_hunks(d, opts['diff-style'] == 'context') literate.diff.build_output_lines(d) literate.marks.init_hunk_boundary_marks(d) return d def render_diff(old_files: Dict[str, File], new_files: Dict[str, File], opts: Dict[str, Any]) -> Optional[str]: '''Render a diff between each file in `new_files` and the corresponding one in `old_files`. The result is either a string of HTML source, or `None` if nothing changed.''' if opts['hide-diff']: return None omit_old = opts['diff-style'] == 'only-new' cols = 4 if not omit_old else 2 file_names = sorted(new_files.keys()) # Is the diff empty? empty = True parts = [] parts.append('\n' % literate.highlight.get_highlight_class(opts)) parts.append('') if not omit_old: parts.append('') parts.append('') parts.append('\n') for file_idx, f in enumerate(file_names): # `make_diff` copies the files, then updates the copies. We want # references to the new copies only. diff = make_diff(old_files[f], new_files[f], opts) old = diff.old_file new = diff.new_file if len(diff.hunks) > 0: empty = False if opts['show-filename']: parts.append('' % (cols, f)) else: if file_idx > 0: parts.append('' % cols) first = True for hunk in diff.hunks: if not first: parts.append('' % cols) first = False for ol in hunk.output_lines: old_cls = 'diff-old' if ol.changed else '' new_cls = 'diff-new' if ol.changed and not omit_old else '' parts.append('') if not omit_old: if ol.old_line is not None: parts.append('' % (old_cls, ol.old_line + 1)) parts.append('') else: parts.append('') if ol.new_line is not None: parts.append('' % (new_cls, ol.new_line + 1)) parts.append('') else: parts.append('') parts.append('\n') parts.append('
%s
%d
' % old_cls)
                        parts.append(render_line(old.lines[ol.old_line], old, opts))
                        parts.append('
%d
' % new_cls)
                    parts.append(render_line(new.lines[ol.new_line], new, opts))
                    parts.append('
\n') if empty: return None return ''.join(parts) def get_styles() -> str: '''Return CSS styles for displaying generated diffs.''' parts = [] parts.append('table.highlight { width: 800px; table-layout: fixed; }') parts.append('.highlight td { overflow-wrap: break-word; }') parts.append('.highlight td.line-num { text-align: right; }') parts.append('.highlight pre { display: inline; white-space: pre-wrap; }') # Diff highlight colors for use with `monokai` color scheme parts.append('.diff-old { background-color: rgba(255, 0, 0, 0.25); }') parts.append('.diff-new { background-color: rgba(0, 255, 0, 0.20); }') parts.append('.diff-intra-del { border: solid 1px #cc0000; }') parts.append('.diff-intra-chg { border: solid 1px #cccc00; }') parts.append('.diff-intra-ins { border: solid 1px #00cc00; }') # Monokai colors for mark indicators (taken from pygments.styles.monokai) parts.append('.mark-kept { color: #66d9ef; border: solid 1px; }') parts.append('.mark-ins { color: #a6e22e; border: solid 1px; }') parts.append('.mark-del { color: #f92672; border: solid 1px; }') parts.append('.mark-chg { color: #e6db74; border: solid 1px; }') # Colors for light-background color schemes, like `friendly` #parts.append('.diff-old { background-color: #ffcccc; }') #parts.append('.diff-new { background-color: #ccffcc; }') #parts.append('.diff-intra-del { border: solid 1px #880000; }') #parts.append('.diff-intra-chg { border: solid 1px #888800; }') #parts.append('.diff-intra-ins { border: solid 1px #008800; }') # Compatibility with manual/mdbook rendering parts.append('.diff { font-size: 0.875em; }') parts.append('.diff td { padding: 2px; }') parts.append('.diff td pre { margin: 0; }') parts.append('table.diff tbody tr:nth-child(2n) { background: inherit; }') return '\n'.join(parts) + '\n' def get_pygments_styles(fmt=None) -> str: fmt = fmt or pygments.formatters.get_formatter_by_name( 'html', nowrap=True, style='monokai') return fmt.get_style_defs('.highlight')