Coverage for liitos / labels.py: 83.44%

1from collections.abc import Iterable 

2 

3from liitos import log 

4 

5NO_LABEL = 'no-label-found-ERROR' 

6 

7 

8def is_include_graphics(text: str) -> bool: 

9 """Only DRY.""" 

10 ntoken = r'\pandocbounded' 

11 pos_after = len(ntoken) 

12 if len(text) > len(ntoken) and text.startswith(ntoken) and text[pos_after] in ('[', '{'): 

13 return True 

14 token = r'\includegraphics' # nosec B105 

15 pos_after = len(token) 

16 return len(text) > len(token) and text.startswith(token) and text[pos_after] in ('[', '{') 

17 

18 

19def extract_image_path(include_graphics_line: str) -> str: 

20 """We had bugs, so we isolate in a function to simplify following upstream fashions.""" 

21 igl = include_graphics_line 

22 if igl and 'pandocbounded{' in igl and ',alt={' in igl: 

23 return igl.split('}]{', 1)[1].rstrip().rstrip('}') 

24 if igl and 'pandocbounded{' in igl: 24 ↛ 25line 24 didn't jump to line 25 because the condition on line 24 was never true

25 return igl.split('{', 2)[2].rstrip().rstrip('}') 

26 if igl and '{' in igl: 

27 return igl.split('{', 1)[1].rstrip().rstrip('}') 

28 else: 

29 return 'IMAGE_PATH_NOT_FOUND' 

30 

31 

32def inject(incoming: Iterable[str], lookup: dict[str, str] | None = None) -> list[str]: 

33 """Later alligator.""" 

34 outgoing = [] 

35 modus = 'copy' 

36 label = NO_LABEL 

37 figure: list[str] = [] 

38 caption: list[str] = [] 

39 precondition = r'\begin{figure}' 

40 precondition_met = False 

41 for slot, line in enumerate(incoming): 

42 if line.startswith(precondition) and not precondition_met: 

43 log.info(f'start of a figure environment at line #{slot + 1}') 

44 precondition_met = True 

45 outgoing.append(line) 

46 continue 

47 

48 if modus == 'copy': 

49 if is_include_graphics(line) and not precondition_met: 

50 log.warning(f'graphics include outside of a figure environment at line #{slot + 1}') 

51 log.debug(f'line#{slot + 1}|{line.rstrip()}') 

52 log.info('trying to fix temporarily ... watch for marker MISSING-CAPTION-IN-MARKDOWN') 

53 adhoc_label = 'FIX-AT-SOURCE' 

54 try: 

55 lab = line.split('{', 1)[1] 

56 lab = lab.rsplit('.', 1)[0] 

57 lab = lab.rsplit('/', 1)[1] 

58 adhoc_label = r'\label{fig:' + lab + '}' 

59 log.info(adhoc_label) 

60 except Exception as err: 

61 log.error(f'failed to extract generic label from {line.strip()} with err: {err}') 

62 captain = 'MISSING-CAPTION-IN-MARKDOWN' 

63 try: 

64 token = extract_image_path(line) 

65 log.debug(f'- looking up key ({token}) for captions ...') 

66 if lookup is not None: 

67 cand = lookup.get(token, None) 

68 if cand is not None: 

69 captain = cand 

70 log.debug(f' + found ({captain})') 

71 else: 

72 log.debug(f' ? no match for key ({token})') 

73 else: 

74 log.debug(' + no lut?') 

75 except Exception as err: 

76 log.error( 

77 f'failed to extract file path token for caption lookup from {line.strip()} with err: {err}' 

78 ) 

79 outgoing.append(r'\begin{figure}' + '\n') 

80 outgoing.append(r'\centering' + '\n') 

81 outgoing.append(line) 

82 outgoing.append(r'\caption{' + captain + ' ' + adhoc_label + '}' + '\n') 

83 outgoing.append(r'\end{figure}' + '\n') 

84 elif is_include_graphics(line) and precondition_met: 

85 log.info(f'within a figure environment at line #{slot + 1}') 

86 log.info(line.rstrip()) 

87 modus = 'figure' 

88 figure = [line] 

89 try: 

90 lab = line.split('{', 1)[1] 

91 lab = lab.rsplit('.', 1)[0] 

92 lab = lab.rsplit('/', 1)[1] 

93 label = r'\label{fig:' + lab + '}' 

94 log.info(label) 

95 except Exception as err: 

96 log.error(f'failed to extract generic label from {line.strip()} with err: {err}') 

97 caption = [] 

98 else: 

99 log.debug(f'- copying {slot + 1 :3d}|{line.rstrip()}') 

100 outgoing.append(line) 

101 continue 

102 

103 if modus == 'figure': 103 ↛ 128line 103 didn't jump to line 128 because the condition on line 103 was always true

104 if line.startswith(r'\caption{'): 

105 log.info(f'- found the caption start at line #{slot + 1}') 

106 caption.append(line) 

107 if not line.strip().endswith('}'): 107 ↛ 108line 107 didn't jump to line 108 because the condition on line 107 was never true

108 log.info(f'- multi line caption at line #{slot + 1}') 

109 modus = 'caption' 

110 elif line.startswith(r'\end{figure}'): 110 ↛ 124line 110 didn't jump to line 124 because the condition on line 110 was always true

111 log.info(f'end of figure env detected at line #{slot + 1}') 

112 outgoing.extend(figure) 

113 caption_text = '\n'.join(caption) 

114 if r'\label{' not in caption_text: 114 ↛ 116line 114 didn't jump to line 116 because the condition on line 114 was always true

115 caption_text = f"{caption_text.rstrip().rstrip('}')} {label}" + '}' 

116 if caption_text == caption_text.rstrip(): 116 ↛ 118line 116 didn't jump to line 118 because the condition on line 116 was always true

117 caption_text += '\n' 

118 outgoing.append(caption_text) 

119 outgoing.append(line) 

120 modus = 'copy' 

121 label = NO_LABEL 

122 precondition_met = False 

123 else: 

124 log.debug(f'- figure continues at line #{slot + 1}') 

125 figure.append(line) 

126 continue 

127 

128 if modus == 'caption': 

129 caption.append(line) 

130 if line.strip().endswith(r'}'): 

131 log.info(f'- caption read at line #{slot + 1}') 

132 modus = 'figure' 

133 continue 

134 

135 return outgoing