import inspect import itertools import logging import re import string from operator import itemgetter from typing import ( Any, Callable, Dict, Generator, List, Match, Optional, Pattern, Tuple, Union, ) from .._typing import T_bbox, T_dir, T_num, T_obj, T_obj_iter, T_obj_list from .clustering import cluster_objects from .generic import to_list from .geometry import objects_to_bbox logger = logging.getLogger(__name__) DEFAULT_X_TOLERANCE = 3 DEFAULT_Y_TOLERANCE = 3 DEFAULT_X_DENSITY = 7.25 DEFAULT_Y_DENSITY = 13 DEFAULT_LINE_DIR: T_dir = "ttb" DEFAULT_CHAR_DIR: T_dir = "ltr" LIGATURES = { "ff": "ff", "ffi": "ffi", "ffl": "ffl", "fi": "fi", "fl": "fl", "st": "st", "ſt": "st", } def get_line_cluster_key(line_dir: T_dir) -> Callable[[T_obj], T_num]: return { "ttb": lambda x: x["top"], "btt": lambda x: -x["bottom"], "ltr": lambda x: x["x0"], "rtl": lambda x: -x["x1"], }[line_dir] def get_char_sort_key(char_dir: T_dir) -> Callable[[T_obj], Tuple[T_num, T_num]]: return { "ttb": lambda x: (x["top"], x["bottom"]), "btt": lambda x: (-(x["top"] + x["height"]), -x["top"]), "ltr": lambda x: (x["x0"], x["x0"]), "rtl": lambda x: (-x["x1"], -x["x0"]), }[char_dir] BBOX_ORIGIN_KEYS = { "ttb": itemgetter(1), "btt": itemgetter(3), "ltr": itemgetter(0), "rtl": itemgetter(2), } POSITION_KEYS = { "ttb": itemgetter("top"), "btt": itemgetter("bottom"), "ltr": itemgetter("x0"), "rtl": itemgetter("x1"), } def validate_directions(line_dir: T_dir, char_dir: T_dir, suffix: str = "") -> None: valid_dirs = set(POSITION_KEYS.keys()) if line_dir not in valid_dirs: raise ValueError( f"line_dir{suffix} must be one of {valid_dirs}, not {line_dir}" ) if char_dir not in valid_dirs: raise ValueError( f"char_dir{suffix} must be one of {valid_dirs}, not {char_dir}" ) if set(line_dir) == set(char_dir): raise ValueError( f"line_dir{suffix}={line_dir} is incompatible " f"with char_dir{suffix}={char_dir}" ) class TextMap: """ A TextMap maps each unicode character in the text to an individual `char` object (or, in the case of layout-implied whitespace, `None`). """ def __init__( self, tuples: List[Tuple[str, Optional[T_obj]]], line_dir_render: T_dir, char_dir_render: T_dir, ) -> None: validate_directions(line_dir_render, char_dir_render, "_render") self.tuples = tuples self.line_dir_render = line_dir_render self.char_dir_render = char_dir_render self.as_string = self.to_string() def to_string(self) -> str: cd = self.char_dir_render ld = self.line_dir_render base = "".join(map(itemgetter(0), self.tuples)) if cd == "ltr" and ld == "ttb": return base else: lines = base.split("\n") if ld in ("btt", "rtl"): lines = list(reversed(lines)) if cd == "rtl": lines = ["".join(reversed(line)) for line in lines] if ld in ("rtl", "ltr"): max_line_length = max(map(len, lines)) if cd == "btt": lines = [ (" " * (max_line_length - len(line))) + line for line in lines ] else: lines = [ line + (" " * (max_line_length - len(line))) for line in lines ] return "\n".join( "".join(line[i] for line in lines) for i in range(max_line_length) ) else: return "\n".join(lines) def match_to_dict( self, m: Match[str], main_group: int = 0, return_groups: bool = True, return_chars: bool = True, ) -> Dict[str, Any]: subset = self.tuples[m.start(main_group) : m.end(main_group)] chars = [c for (text, c) in subset if c is not None] x0, top, x1, bottom = objects_to_bbox(chars) result = { "text": m.group(main_group), "x0": x0, "top": top, "x1": x1, "bottom": bottom, } if return_groups: result["groups"] = m.groups() if return_chars: result["chars"] = chars return result def search( self, pattern: Union[str, Pattern[str]], regex: bool = True, case: bool = True, return_groups: bool = True, return_chars: bool = True, main_group: int = 0, ) -> List[Dict[str, Any]]: if isinstance(pattern, Pattern): if regex is False: raise ValueError( "Cannot pass a compiled search pattern *and* regex=False together." ) if case is False: raise ValueError( "Cannot pass a compiled search pattern *and* case=False together." ) compiled = pattern else: if regex is False: pattern = re.escape(pattern) flags = re.I if case is False else 0 compiled = re.compile(pattern, flags) gen = re.finditer(compiled, self.as_string) # Remove zero-length matches (can happen, e.g., with optional # patterns in regexes) and whitespace-only matches filtered = filter(lambda m: bool(m.group(main_group).strip()), gen) return [ self.match_to_dict( m, return_groups=return_groups, return_chars=return_chars, main_group=main_group, ) for m in filtered ] def extract_text_lines( self, strip: bool = True, return_chars: bool = True ) -> List[Dict[str, Any]]: """ `strip` is analogous to Python's `str.strip()` method, and returns `text` attributes without their surrounding whitespace. Only relevant when the relevant TextMap is created with `layout` = True Setting `return_chars` to False will exclude the individual character objects from the returned text-line dicts. """ if strip: pat = r" *([^\n]+?) *(\n|$)" else: pat = r"([^\n]+)" return self.search( pat, main_group=1, return_chars=return_chars, return_groups=False ) class WordMap: """ A WordMap maps words->chars. """ def __init__(self, tuples: List[Tuple[T_obj, T_obj_list]]) -> None: self.tuples = tuples def to_textmap( self, layout: bool = False, layout_width: T_num = 0, layout_height: T_num = 0, layout_width_chars: int = 0, layout_height_chars: int = 0, layout_bbox: T_bbox = (0, 0, 0, 0), x_density: T_num = DEFAULT_X_DENSITY, y_density: T_num = DEFAULT_Y_DENSITY, x_shift: T_num = 0, y_shift: T_num = 0, y_tolerance: T_num = DEFAULT_Y_TOLERANCE, line_dir: T_dir = DEFAULT_LINE_DIR, char_dir: T_dir = DEFAULT_CHAR_DIR, line_dir_rotated: Optional[T_dir] = None, char_dir_rotated: Optional[T_dir] = None, char_dir_render: Optional[T_dir] = None, line_dir_render: Optional[T_dir] = None, use_text_flow: bool = False, presorted: bool = False, expand_ligatures: bool = True, ) -> TextMap: """ Given a list of (word, chars) tuples (i.e., a WordMap), return a list of (char-text, char) tuples (i.e., a TextMap) that can be used to mimic the structural layout of the text on the page(s), using the following approach for top-to-bottom, left-to-right text: - Sort the words by (top, x0) if not already sorted. - Cluster the words by top (taking `y_tolerance` into account), and iterate through them. - For each cluster, divide (top - y_shift) by `y_density` to calculate the minimum number of newlines that should come before this cluster. Append that number of newlines *minus* the number of newlines already appended, with a minimum of one. - Then for each cluster, iterate through each word in it. Divide each word's x0, minus `x_shift`, by `x_density` to calculate the minimum number of characters that should come before this cluster. Append that number of spaces *minus* the number of characters and spaces already appended, with a minimum of one. Then append the word's text. - At the termination of each line, add more spaces if necessary to mimic `layout_width`. - Finally, add newlines to the end if necessary to mimic to `layout_height`. For other line/character directions (e.g., bottom-to-top, right-to-left), these steps are adjusted. """ _textmap: List[Tuple[str, Optional[T_obj]]] = [] if not len(self.tuples): return TextMap( _textmap, line_dir_render=line_dir_render or line_dir, char_dir_render=char_dir_render or char_dir, ) expansions = LIGATURES if expand_ligatures else {} if layout: if layout_width_chars: if layout_width: raise ValueError( "`layout_width` and `layout_width_chars` cannot both be set." ) else: layout_width_chars = int(round(layout_width / x_density)) if layout_height_chars: if layout_height: raise ValueError( "`layout_height` and `layout_height_chars` cannot both be set." ) else: layout_height_chars = int(round(layout_height / y_density)) blank_line = [(" ", None)] * layout_width_chars else: blank_line = [] num_newlines = 0 line_cluster_key = get_line_cluster_key(line_dir) char_sort_key = get_char_sort_key(char_dir) line_position_key = POSITION_KEYS[line_dir] char_position_key = POSITION_KEYS[char_dir] y_origin = BBOX_ORIGIN_KEYS[line_dir](layout_bbox) x_origin = BBOX_ORIGIN_KEYS[char_dir](layout_bbox) words_sorted_line_dir = ( self.tuples if presorted or use_text_flow else sorted(self.tuples, key=lambda x: line_cluster_key(x[0])) ) tuples_by_line = cluster_objects( words_sorted_line_dir, lambda x: line_cluster_key(x[0]), y_tolerance, preserve_order=presorted or use_text_flow, ) for i, line_tuples in enumerate(tuples_by_line): if layout: line_position = line_position_key(line_tuples[0][0]) y_dist_raw = line_position - (y_origin + y_shift) adj = -1 if line_dir in ["btt", "rtl"] else 1 y_dist = y_dist_raw * adj / y_density else: y_dist = 0 num_newlines_prepend = max( # At least one newline, unless this iis the first line int(i > 0), # ... or as many as needed to get the imputed "distance" from the top round(y_dist) - num_newlines, ) for i in range(num_newlines_prepend): if not len(_textmap) or _textmap[-1][0] == "\n": _textmap += blank_line _textmap.append(("\n", None)) num_newlines += num_newlines_prepend line_len = 0 line_tuples_sorted = ( line_tuples if presorted or use_text_flow else sorted(line_tuples, key=lambda x: char_sort_key(x[0])) ) for word, chars in line_tuples_sorted: if layout: char_position = char_position_key(word) x_dist_raw = char_position - (x_origin + x_shift) adj = -1 if char_dir in ["btt", "rtl"] else 1 x_dist = x_dist_raw * adj / x_density else: x_dist = 0 num_spaces_prepend = max(min(1, line_len), round(x_dist) - line_len) _textmap += [(" ", None)] * num_spaces_prepend line_len += num_spaces_prepend for c in chars: letters = expansions.get(c["text"], c["text"]) for letter in letters: _textmap.append((letter, c)) line_len += 1 # Append spaces at end of line if layout: _textmap += [(" ", None)] * (layout_width_chars - line_len) # Append blank lines at end of text if layout: num_newlines_append = layout_height_chars - (num_newlines + 1) for i in range(num_newlines_append): if i > 0: _textmap += blank_line _textmap.append(("\n", None)) # Remove terminal newline if _textmap[-1] == ("\n", None): _textmap = _textmap[:-1] return TextMap( _textmap, line_dir_render=line_dir_render or line_dir, char_dir_render=char_dir_render or char_dir, ) class WordExtractor: def __init__( self, x_tolerance: T_num = DEFAULT_X_TOLERANCE, y_tolerance: T_num = DEFAULT_Y_TOLERANCE, x_tolerance_ratio: Union[int, float, None] = None, y_tolerance_ratio: Union[int, float, None] = None, keep_blank_chars: bool = False, use_text_flow: bool = False, vertical_ttb: bool = True, # Should vertical words be read top-to-bottom? horizontal_ltr: bool = True, # Should words be read left-to-right? line_dir: T_dir = DEFAULT_LINE_DIR, char_dir: T_dir = DEFAULT_CHAR_DIR, line_dir_rotated: Optional[T_dir] = None, char_dir_rotated: Optional[T_dir] = None, extra_attrs: Optional[List[str]] = None, split_at_punctuation: Union[bool, str] = False, expand_ligatures: bool = True, ): self.x_tolerance = x_tolerance self.y_tolerance = y_tolerance self.x_tolerance_ratio = x_tolerance_ratio self.y_tolerance_ratio = y_tolerance_ratio self.keep_blank_chars = keep_blank_chars self.use_text_flow = use_text_flow self.horizontal_ltr = horizontal_ltr self.vertical_ttb = vertical_ttb if vertical_ttb is False: logger.warning( "vertical_ttb is deprecated and will be removed;" " use line_dir/char_dir instead." ) if horizontal_ltr is False: logger.warning( "horizontal_ltr is deprecated and will be removed;" " use line_dir/char_dir instead." ) self.line_dir = line_dir self.char_dir = char_dir # Default is to "flip" the directions for rotated text self.line_dir_rotated = line_dir_rotated or char_dir self.char_dir_rotated = char_dir_rotated or line_dir validate_directions(self.line_dir, self.char_dir) validate_directions(self.line_dir_rotated, self.char_dir_rotated, "_rotated") self.extra_attrs = [] if extra_attrs is None else extra_attrs # Note: string.punctuation = '!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~' self.split_at_punctuation = ( string.punctuation if split_at_punctuation is True else (split_at_punctuation or "") ) self.expansions = LIGATURES if expand_ligatures else {} def get_char_dir(self, upright: int) -> T_dir: # Note: This can be simplified and reincorporated into .merge_chars and # .iter_chars_to_lines once .vertical_ttb and .horizontal_ltr # deprecation is complete. if not upright and not self.vertical_ttb: return "btt" elif upright and not self.horizontal_ltr: return "rtl" return self.char_dir if upright else self.char_dir_rotated def merge_chars(self, ordered_chars: T_obj_list) -> T_obj: x0, top, x1, bottom = objects_to_bbox(ordered_chars) doctop_adj = ordered_chars[0]["doctop"] - ordered_chars[0]["top"] upright = ordered_chars[0]["upright"] char_dir = self.get_char_dir(upright) word = { "text": "".join( self.expansions.get(c["text"], c["text"]) for c in ordered_chars ), "x0": x0, "x1": x1, "top": top, "doctop": top + doctop_adj, "bottom": bottom, "upright": upright, "height": bottom - top, "width": x1 - x0, "direction": char_dir, } for key in self.extra_attrs: word[key] = ordered_chars[0][key] return word def char_begins_new_word( self, prev_char: T_obj, curr_char: T_obj, direction: T_dir, x_tolerance: T_num, y_tolerance: T_num, ) -> bool: """This method takes several factors into account to determine if `curr_char` represents the beginning of a new word: - Whether the text is "upright" (i.e., non-rotated) - Whether the user has specified that horizontal text runs left-to-right (default) or right-to-left, as represented by self.horizontal_ltr - Whether the user has specified that vertical text the text runs top-to-bottom (default) or bottom-to-top, as represented by self.vertical_ttb - The x0, top, x1, and bottom attributes of prev_char and curr_char - The self.x_tolerance and self.y_tolerance settings. Note: In this case, x/y refer to those directions for non-rotated text. For vertical text, they are flipped. A more accurate terminology might be "*intra*line character distance tolerance" and "*inter*line character distance tolerance" An important note: The *intra*line distance is measured from the *end* of the previous character to the *beginning* of the current character, while the *inter*line distance is measured from the *top* of the previous character to the *top* of the next character. The reasons for this are partly repository-historical, and partly logical, as successive text lines' bounding boxes often overlap slightly (and we don't want that overlap to be interpreted as the two lines being the same line). The upright-ness of the character determines the attributes to compare, while horizontal_ltr/vertical_ttb determine the direction of the comparison. """ # Note: Due to the grouping step earlier in the process, # curr_char["upright"] will always equal prev_char["upright"]. if direction in ("ltr", "rtl"): x = x_tolerance y = y_tolerance ay = prev_char["top"] cy = curr_char["top"] if direction == "ltr": ax = prev_char["x0"] bx = prev_char["x1"] cx = curr_char["x0"] else: ax = -prev_char["x1"] bx = -prev_char["x0"] cx = -curr_char["x1"] else: x = y_tolerance y = x_tolerance ay = prev_char["x0"] cy = curr_char["x0"] if direction == "ttb": ax = prev_char["top"] bx = prev_char["bottom"] cx = curr_char["top"] else: ax = -prev_char["bottom"] bx = -prev_char["top"] cx = -curr_char["bottom"] return bool( # Intraline test (cx < ax) or (cx > bx + x) # Interline test or abs(cy - ay) > y ) def iter_chars_to_words( self, ordered_chars: T_obj_iter, direction: T_dir, ) -> Generator[T_obj_list, None, None]: current_word: T_obj_list = [] def start_next_word( new_char: Optional[T_obj], ) -> Generator[T_obj_list, None, None]: nonlocal current_word if current_word: yield current_word current_word = [] if new_char is None else [new_char] xt = self.x_tolerance xtr = self.x_tolerance_ratio yt = self.y_tolerance ytr = self.y_tolerance_ratio for char in ordered_chars: text = char["text"] if not self.keep_blank_chars and text.isspace(): yield from start_next_word(None) elif text in self.split_at_punctuation: yield from start_next_word(char) yield from start_next_word(None) elif current_word and self.char_begins_new_word( current_word[-1], char, direction, x_tolerance=(xt if xtr is None else xtr * current_word[-1]["size"]), y_tolerance=(yt if ytr is None else ytr * current_word[-1]["size"]), ): yield from start_next_word(char) else: current_word.append(char) # Finally, after all chars processed if current_word: yield current_word def iter_chars_to_lines( self, chars: T_obj_iter ) -> Generator[Tuple[T_obj_list, T_dir], None, None]: chars = list(chars) upright = chars[0]["upright"] line_dir = self.line_dir if upright else self.line_dir_rotated char_dir = self.get_char_dir(upright) line_cluster_key = get_line_cluster_key(line_dir) char_sort_key = get_char_sort_key(char_dir) # Cluster by line subclusters = cluster_objects( chars, line_cluster_key, (self.y_tolerance if line_dir in ("ttb", "btt") else self.x_tolerance), ) for sc in subclusters: # Sort within line chars_sorted = sorted(sc, key=char_sort_key) yield (chars_sorted, char_dir) def iter_extract_tuples( self, chars: T_obj_iter ) -> Generator[Tuple[T_obj, T_obj_list], None, None]: grouping_key = itemgetter("upright", *self.extra_attrs) grouped_chars = itertools.groupby(chars, grouping_key) for keyvals, char_group in grouped_chars: line_groups = ( [(char_group, self.char_dir)] if self.use_text_flow else self.iter_chars_to_lines(char_group) ) for line_chars, direction in line_groups: for word_chars in self.iter_chars_to_words(line_chars, direction): yield (self.merge_chars(word_chars), word_chars) def extract_wordmap(self, chars: T_obj_iter) -> WordMap: return WordMap(list(self.iter_extract_tuples(chars))) def extract_words( self, chars: T_obj_list, return_chars: bool = False ) -> T_obj_list: if return_chars: return list( {**word, "chars": word_chars} for word, word_chars in self.iter_extract_tuples(chars) ) else: return list(word for word, word_chars in self.iter_extract_tuples(chars)) def extract_words( chars: T_obj_list, return_chars: bool = False, **kwargs: Any ) -> T_obj_list: return WordExtractor(**kwargs).extract_words(chars, return_chars) TEXTMAP_KWARGS = inspect.signature(WordMap.to_textmap).parameters.keys() WORD_EXTRACTOR_KWARGS = inspect.signature(WordExtractor).parameters.keys() def chars_to_textmap(chars: T_obj_list, **kwargs: Any) -> TextMap: kwargs.update( { "presorted": True, "layout_bbox": kwargs.get("layout_bbox") or objects_to_bbox(chars), } ) extractor = WordExtractor( **{k: kwargs[k] for k in WORD_EXTRACTOR_KWARGS if k in kwargs} ) wordmap = extractor.extract_wordmap(chars) textmap = wordmap.to_textmap( **{k: kwargs[k] for k in TEXTMAP_KWARGS if k in kwargs} ) return textmap def extract_text( chars: T_obj_list, line_dir_render: Optional[T_dir] = None, char_dir_render: Optional[T_dir] = None, **kwargs: Any, ) -> str: chars = to_list(chars) if len(chars) == 0: return "" if kwargs.get("layout"): textmap_kwargs = { **kwargs, **{"line_dir_render": line_dir_render, "char_dir_render": char_dir_render}, } return chars_to_textmap(chars, **textmap_kwargs).as_string else: extractor = WordExtractor( **{k: kwargs[k] for k in WORD_EXTRACTOR_KWARGS if k in kwargs} ) words = extractor.extract_words(chars) line_dir_render = line_dir_render or extractor.line_dir char_dir_render = char_dir_render or extractor.char_dir line_cluster_key = get_line_cluster_key(extractor.line_dir) x_tolerance = kwargs.get("x_tolerance", DEFAULT_X_TOLERANCE) y_tolerance = kwargs.get("y_tolerance", DEFAULT_Y_TOLERANCE) lines = cluster_objects( words, line_cluster_key, y_tolerance if line_dir_render in ("ttb", "btt") else x_tolerance, ) return TextMap( [ (char, None) for char in ( "\n".join(" ".join(word["text"] for word in line) for line in lines) ) ], line_dir_render=line_dir_render, char_dir_render=char_dir_render, ).as_string def collate_line( line_chars: T_obj_list, tolerance: T_num = DEFAULT_X_TOLERANCE, ) -> str: coll = "" last_x1 = None for char in sorted(line_chars, key=itemgetter("x0")): if (last_x1 is not None) and (char["x0"] > (last_x1 + tolerance)): coll += " " last_x1 = char["x1"] coll += char["text"] return coll def extract_text_simple( chars: T_obj_list, x_tolerance: T_num = DEFAULT_X_TOLERANCE, y_tolerance: T_num = DEFAULT_Y_TOLERANCE, ) -> str: clustered = cluster_objects(chars, itemgetter("doctop"), y_tolerance) return "\n".join(collate_line(c, x_tolerance) for c in clustered) def dedupe_chars( chars: T_obj_list, tolerance: T_num = 1, extra_attrs: Optional[Tuple[str, ...]] = ("fontname", "size"), ) -> T_obj_list: """ Removes duplicate chars — those sharing the same text and positioning (within `tolerance`) as other characters in the set. Use extra_args to be more restrictive with the properties shared by the matching chars. """ key = itemgetter(*("upright", "text"), *(extra_attrs or tuple())) pos_key = itemgetter("doctop", "x0") def yield_unique_chars(chars: T_obj_list) -> Generator[T_obj, None, None]: sorted_chars = sorted(chars, key=key) for grp, grp_chars in itertools.groupby(sorted_chars, key=key): for y_cluster in cluster_objects( list(grp_chars), itemgetter("doctop"), tolerance ): for x_cluster in cluster_objects( y_cluster, itemgetter("x0"), tolerance ): yield sorted(x_cluster, key=pos_key)[0] deduped = yield_unique_chars(chars) return sorted(deduped, key=chars.index)