from pdfplumber import utils from pdfplumber.container import Container from six import string_types import re lt_pat = re.compile(r"^LT") class Page(Container): cached_properties = Container.cached_properties + [ "_layout" ] def __init__(self, pdf, page_obj, page_number=None, initial_doctop=0): self.pdf = pdf self.page_obj = page_obj self.page_number = page_number self.page_obj.rotate = self.page_obj.attrs.get("Rotate", 0) % 360 self.layout = self.pdf.process_page(self.page_obj) self.initial_doctop = self.decimalize(initial_doctop) cropbox = page_obj.attrs.get("CropBox", page_obj.attrs.get("MediaBox")) self.cropbox = tuple(map(self.decimalize, cropbox)) self.bbox = self.layout.bbox x0, bottom, x1, top = self.bbox width = x1 - x0 height = top - bottom def decimalize(self, x): return utils.decimalize(x, self.pdf.precision) @property def width(self): return self.bbox[2] - self.bbox[0] @property def height(self): return self.bbox[3] - self.bbox[1] @property def objects(self): if hasattr(self, "_objects"): return self._objects self._objects = self.parse_objects() return self._objects def parse_objects(self): objects = {} d = self.decimalize h = self.height idc = self.initial_doctop pno = self.page_number def process_object(obj): attr = dict((k, d(v)) for k, v in obj.__dict__.items() if isinstance(v, (float, int, string_types)) and k[0] != "_") kind = re.sub(lt_pat, "", obj.__class__.__name__).lower() attr["object_type"] = kind attr["page_number"] = pno if hasattr(obj, "get_text"): attr["text"] = obj.get_text() if attr.get("y0") != None: attr["top"] = h - attr["y1"] attr["bottom"] = h - attr["y0"] attr["doctop"] = idc + attr["top"] if objects.get(kind) == None: objects[kind] = [] objects[kind].append(attr) if hasattr(obj, "_objs"): for child in obj._objs: process_object(child) for obj in self.layout._objs: process_object(obj) return objects def filter_edges(self, orientation, edge_type=None, min_length=1): if orientation not in ("v", "h"): raise Exception("Orientation must be 'v' or 'h'") def test(e): dim = "height" if e["orientation"] == "v" else "width" et = (e["object_type"] == edge_type if edge_type != None else True) return et & ( (e["orientation"] == orientation) & (e[dim] >= min_length) ) edges = filter(test, self.edges) return edges def get_edge_positions(self, orientation, edge_type=None, min_length=1, tolerance=1): edges = self.filter_edges(orientation, edge_type=edge_type, min_length=min_length) pos_var = "x0" if orientation == "v" else "top" edges_uniq = set(e[pos_var] for e in edges) edges_clust = utils.cluster_list(edges_uniq, tolerance=tolerance) edge_means = list(sorted(sum(c) / len(c) for c in edges_clust)) return edge_means TABLE_STRATEGIES = ("lines", "lines_strict", "gutters") def extract_table(self, v="lines", h="lines", line_min_height=1, line_min_width=1, gutter_min_width=5, gutter_min_height=5, x_tolerance=utils.DEFAULT_X_TOLERANCE, y_tolerance=utils.DEFAULT_Y_TOLERANCE): """ For the purposes of this method, "lines" refers to all two-dimensional lines, including "rect_edge" objects. To use only actual "line" objects, choose the "lines_strict" strategy. """ def use_strategy(param, name): # If dividers are explicitly passed if isinstance(param, (list, tuple)): return param else: if param not in self.TABLE_STRATEGIES: msg = "{0} must be list/tuple of ints/floats or one of {1}"\ .format(name, self.TABLE_STRATEGIES) raise ValueError(msg) if param == "lines": if name == "v": return self.get_edge_positions("v", min_length=line_min_height, tolerance=x_tolerance) if name == "h": return self.get_edge_positions("h", min_length=line_min_width, tolerance=y_tolerance) if param == "lines_strict": if name == "v": return self.get_edge_positions("v", edge_type="line", min_length=line_min_height, tolerance=x_tolerance) if name == "h": return self.get_edge_positions("h", edge_type="line", min_length=line_min_width, tolerance=y_tolerance) if param == "gutters": if name == "v": return utils.find_gutters(self.chars, "v", min_size=gutter_min_width) if name == "h": return utils.find_gutters(self.chars, "h", min_size=gutter_min_height) h = list(map(self.decimalize, use_strategy(h, "h"))) v = list(map(self.decimalize, use_strategy(v, "v"))) table = utils.extract_table(self.chars, v, h, x_tolerance=x_tolerance, y_tolerance=y_tolerance) return table def extract_text(self, x_tolerance=utils.DEFAULT_X_TOLERANCE, y_tolerance=utils.DEFAULT_Y_TOLERANCE): return utils.extract_text(self.chars, x_tolerance=x_tolerance, y_tolerance=y_tolerance) def extract_words(self, x_tolerance=utils.DEFAULT_X_TOLERANCE, y_tolerance=utils.DEFAULT_Y_TOLERANCE): return utils.extract_words(self.chars, x_tolerance=x_tolerance, y_tolerance=y_tolerance) def crop(self, bbox, strict=False): return CroppedPage(self, bbox, strict=strict) def filter(self, fn): return FilteredPage(self, fn) class CroppedPage(Page): def __init__(self, parent_page, bbox, strict=False): self.parent_page = parent_page self.page_number = parent_page.page_number self.page_number = parent_page.page_number self.decimalize = parent_page.decimalize self.bbox = bbox self.strict = strict @property def objects(self): if hasattr(self, "_objects"): return self._objects if self.strict: kwargs = { "strict": True } else: kwargs = { "crop": True } self._objects = utils.within_bbox( self.parent_page.objects, self.bbox, **kwargs) return self._objects class FilteredPage(Page): def __init__(self, parent_page, test_function): self.parent_page = parent_page self.test_function = test_function self.page_number = parent_page.page_number self.page_number = parent_page.page_number self.decimalize = parent_page.decimalize @property def objects(self): if hasattr(self, "_objects"): return self._objects self._objects = utils.filter_objects( self.parent_page.objects, self.test_function ) return self._objects