1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
|
"""Utility functions and classes for testing."""
import logging
from typing import Union
from PIL import Image
from pypdf import PageObject
from pypdf.generic import DictionaryObject, IndirectObject
class PositionedText:
"""
Specify a text with coordinates, font-dictionary and font-size.
The font-dictionary may be None in case of an unknown font.
"""
def __init__(self, text, x, y, font_dict, font_size) -> None:
# TODO: \0-replace: Encoding issue in some files?
self.text = text.replace("\0", "")
self.x = x
self.y = y
self.font_dict = font_dict
self.font_size = font_size
def get_base_font(self) -> str:
"""
Gets the base font of the text.
Return UNKNOWN in case of an unknown font.
"""
if (self.font_dict is None) or "/BaseFont" not in self.font_dict:
return "UNKNOWN"
return self.font_dict["/BaseFont"]
class Rectangle:
"""Specify a rectangle."""
def __init__(self, x, y, w, h) -> None:
self.x = x.as_numeric()
self.y = y.as_numeric()
self.w = w.as_numeric()
self.h = h.as_numeric()
def contains(self, x, y) -> bool:
return (
self.x <= x <= (self.x + self.w)
and self.y <= y <= (self.y + self.h)
)
def extract_text_and_rectangles(
page: PageObject, rect_filter=None
) -> tuple[list[PositionedText], list[Rectangle]]:
"""
Extracts texts and rectangles of a page of type pypdf._page.PageObject.
This function supports simple coordinate transformations only.
The optional rect_filter-lambda can be used to filter wanted
rectangles.
rect_filter has Rectangle as argument and must return a boolean.
It returns a tuple containing a list of extracted texts and
a list of extracted rectangles.
"""
logger = logging.getLogger("extract_text_and_rectangles")
rectangles = []
texts = []
def print_op_b(op, args, cm_matrix, tm_matrix) -> None:
if logger.isEnabledFor(logging.DEBUG):
logger.debug(f"before: {op} at {cm_matrix}, {tm_matrix}")
if op == b"re":
if logger.isEnabledFor(logging.DEBUG):
logger.debug(f" add rectangle: {args}")
w = args[2]
h = args[3]
r = Rectangle(args[0], args[1], w, h)
if (rect_filter is None) or rect_filter(r):
rectangles.append(r)
def print_visi(text, cm_matrix, tm_matrix, font_dict, font_size) -> None:
if text.strip() != "":
if logger.isEnabledFor(logging.DEBUG):
logger.debug(f"at {cm_matrix}, {tm_matrix}, font size={font_size}")
texts.append(
PositionedText(
text, tm_matrix[4], tm_matrix[5], font_dict, font_size
)
)
visitor_before = print_op_b
visitor_text = print_visi
page.extract_text(
visitor_operand_before=visitor_before, visitor_text=visitor_text
)
return texts, rectangles
def extract_table(
texts: list[PositionedText], rectangles: list[Rectangle]
) -> list[list[list[PositionedText]]]:
"""
Extracts a table containing text.
It is expected that each cell is marked by a rectangle-object.
It is expected that the page contains one table only.
It is expected that the table contains at least 3 columns and 2 rows.
A list of rows is returned.
Each row contains a list of cells.
Each cell contains a list of PositionedText-elements.
"""
logger = logging.getLogger("extractTable")
# Step 1: Count number of x- and y-coordinates of rectangles.
# Remove duplicate rectangles. The new list is rectangles_filtered.
col2count = {}
row2count = {}
key2rectangle = {}
rectangles_filtered = []
for r in rectangles:
# Coordinates may be inaccurate, we have to round.
# cell: x=72.264, y=386.57, w=93.96, h=46.584
# cell: x=72.271, y=386.56, w=93.96, h=46.59
key = f"{round(r.x, 0)} {round(r.y, 0)} {round(r.w, 0)} {round(r.h, 0)}"
if key in key2rectangle:
# Ignore duplicate rectangles
continue
key2rectangle[key] = r
if r.x not in col2count:
col2count[r.x] = 0
if r.y not in row2count:
row2count[r.y] = 0
col2count[r.x] += 1
row2count[r.y] += 1
rectangles_filtered.append(r)
# Step 2: Look for texts in rectangles.
rectangle2texts = {}
for text in texts:
for r in rectangles_filtered:
if r.contains(text.x, text.y):
if r not in rectangle2texts:
rectangle2texts[r] = []
rectangle2texts[r].append(text)
break
# PDF: y = 0 is expected at the bottom of the page.
# So the header-row is expected to have the highest y-value.
rectangles.sort(key=lambda r: (-r.y, r.x))
# Step 3: Build the list of rows containing list of cell-texts.
rows = []
row_nr = 0
col_nr = 0
curr_y = None
curr_row = None
for r in rectangles_filtered:
if col2count[r.x] < 3 or row2count[r.y] < 2:
# We expect at least 3 columns and 2 rows.
continue
if curr_y is None or r.y != curr_y:
# next row
curr_y = r.y
col_nr = 0
row_nr += 1
curr_row = []
rows.append(curr_row)
col_nr += 1
if logger.isEnabledFor(logging.DEBUG):
logger.debug(f"cell: x={r.x}, y={r.y}, w={r.w}, h={r.h}")
if r not in rectangle2texts:
curr_row.append("")
continue
cell_texts = list(rectangle2texts[r])
curr_row.append(cell_texts)
return rows
def extract_cell_text(cell_texts: list[PositionedText]) -> str:
"""Joins the text-objects of a cell."""
return ("".join(t.text for t in cell_texts)).strip()
def get_image_data(
image: Image.Image, band: Union[int, None] = None
) -> Union[tuple[tuple[int, ...], ...], tuple[float, ...]]:
try:
return image.get_flattened_data(band=band)
except AttributeError:
# For Pillow < 12.1.0
return tuple(image.getdata(band=band))
class ReaderDummy:
def __init__(self, strict=False) -> None:
self.strict = strict
def get_object(self, indirect_reference):
class DummyObj:
def get_object(self) -> "DummyObj":
return self
return DictionaryObject()
def get_reference(self, obj):
return IndirectObject(idnum=1, generation=1, pdf=self)
|
