-
Notifications
You must be signed in to change notification settings - Fork 0
/
pdfParser.py
361 lines (299 loc) · 13.4 KB
/
pdfParser.py
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
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
import io
import os
import datetime
from pdfminer.converter import HTMLConverter
from pdfminer.layout import LAParams
from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
from pdfminer.pdfpage import PDFPage
from bs4 import BeautifulSoup
import pandas as pd
import numpy as np
import re
import pytesseract
from pdf2image import convert_from_path
import PyPDF2
from pdfminer.pdfparser import PDFSyntaxError
def pdftohtml(pdf_file):
current_time = currentTime()
output_file_path = f"processed_{current_time}.html"
rsrcmgr = PDFResourceManager()
codec = 'utf-8'
laparams = LAParams()
with io.BytesIO() as output_stream:
converter = HTMLConverter(rsrcmgr, output_stream, codec=codec, laparams=laparams)
with open(pdf_file, 'rb') as input_file:
for page in PDFPage.get_pages(input_file):
try:
interpreter = PDFPageInterpreter(rsrcmgr, converter)
interpreter.process_page(page)
except PDFSyntaxError:
pass
html_content = output_stream.getvalue().decode()
with open(output_file_path, 'w') as output_file:
output_file.write(html_content)
return output_file_path
def getPDF():
for i in range(3):
input_file_path = input("Enter the PDF file path: ")
if os.path.isfile(input_file_path):
return input_file_path
else:
print("File not found.")
print("Maximum number of attempts reached. Exiting program.")
exit()
def isSearchable():
for i in range(3):
isSearchable = input("Is the given PDF searchable? (Enter y or n): ")
if isSearchable == 'y':
return True
elif isSearchable == 'n':
return False
else:
print("Invalid Entry!!")
print("Maximum number of attempts reached. Exiting program.")
exit()
def currentTime():
return datetime.datetime.now().strftime("%d-%m-%Y %H:%M:%S")
def fix_unordered_lists(html):
# Match any <span> tag that contains bullet characters
pattern = r"<span[^>]*>(?:•|•|•|\u2022|-|–|–|⋄|\n)\s*([^<]*)</span>"
# Replace matched spans with bullet tags
html = re.sub(pattern, r"<bullet>\1</bullet>", html, flags=re.MULTILINE)
list_items = []
current_list_item = ''
# Find all occurrences of the </bullet> tag
bullet_closing_indices = [m.end() for m in re.finditer('</bullet>', html)]
# Iterate through the bullet_closing_indices and get the text between bullet tags
for i, closing_index in enumerate(bullet_closing_indices):
opening_index = html.find('<bullet>', closing_index)
if opening_index == -1:
opening_index = len(html)
text = html[closing_index:opening_index]
text = text.strip()
p = re.compile(r'<span[^>]*>|<\/span>')
text = p.sub('', text)
if not text or text.isspace():
continue
text = f"<li>{text}</li>"
if i == 0:
# If this is the first item, start a new list
list_items.append([text])
elif opening_index == len(html):
# If this is the last item and there is no opening <bullet> tag, add it to the current list
current_list_item += text
list_items[-1].append(current_list_item)
elif opening_index > bullet_closing_indices[i - 1] + 1:
# If there is a gap between this bullet and the previous one, start a new list
list_items.append([text])
current_list_item = ''
else:
# Otherwise, add it to the current list
current_list_item += text
list_items[-1].append(current_list_item)
if len(list_items) == 0:
result = ""
else:
result = '<ul>\n' + '\n'.join([''.join(l) for l in list_items]) + '\n</ul>'
firstBulletTagIndex = html.find('<bullet>', 0)
openingText = html.split('<bullet>', 1)[0]
result = openingText + result
# Replace <span> tags with <p> tags
result = re.sub(r'<span\s+[^>]*>(.*?)<\/span>', r'<p>\1</p>', result, flags=re.DOTALL)
return result
def fix_ordered_lists(html):
inputHTML = html
# Match ordered list indices
pattern = r"(?:<p>|<br\/>)(\(\d+\)|\d+\)|\d+\.)"
# Replace matched pattern with bullet tags
html = re.sub(pattern, r"<bullet>\1", html)
pattern = r"<bullet>\s*(.*?)\s*(?=<bullet>|\Z)(?:</p>)?"
matches = re.findall(pattern, html, re.DOTALL)
# Wrap each match inside <li> tags
list_items = []
prev_num = None
for match in matches:
num_match = re.match(r"^\(?(\d+)\)?(\(\d+\)|\d+\)|\d+\.)*", match.strip())
if num_match:
curr_num = int(num_match.group(1))
if prev_num is None or curr_num == prev_num + 1:
if prev_num is None:
list_items.append("<ol>")
list_items.append(f"<li>{match.strip()}</li>")
else:
list_items.append("</ol>")
list_items.append("<ol>")
list_items.append(f"<li>{match.strip()}</li>")
prev_num = curr_num
if list_items:
list_items.append("</ol>")
# Join the list items into a single string
list_html = "\n".join(list_items)
pattern = r'</p>(?!(\s*<p>|$))'
list_html = re.sub(pattern, '', list_html)
if not list_html:
return inputHTML
return list_html
def htmltocsv(html_file_name):
with open(html_file_name) as f:
html_text = f.read()
soup = BeautifulSoup(html_text, 'html.parser')
stack = []
headings = []
hCounter = 0
shCounter = 0
latest_type = -1 # Title - 0, Heading - 1, Subheading - 2, Content - 3
titleParsed = False
for div in soup.find_all(lambda div: div.name == 'div' and 'left' in div.get('style').lower()):
# Fetching value of left attribute
style = div.get('style').lower()
left_value = None
for style_prop in style.split(';'):
if 'left' in style_prop:
left_value = int(style_prop.split(':')[1].strip('px'))
break
if left_value is not None:
span_tags = div.find_all('span')
for tag in span_tags:
# Fetching value of font-size attribute
sty = tag.get('style').lower()
font_size = None
for style_prop in sty.split(';'):
if 'font-size' in style_prop:
font_size = int(style_prop.split(':')[1].strip('px'))
break
content = tag.get_text().strip()
raw = str(tag)
if tag in soup.find_all(lambda tag: tag.name == 'span' and '-bold' in tag.get('style').lower() and font_size is not None and font_size >= 14) and not titleParsed:
# Title
headings.append({'data': content, 'type': 'title'})
latest_type = 0
titleParsed = True
elif tag in soup.find_all(lambda tag: tag.name == 'span' and '-bold' in tag.get('style').lower()):
if latest_type == 0:
# Heading
headings.append({'data': content, 'type': 'heading'})
latest_type = 1
stack.append({'H'+str(hCounter) : left_value})
hCounter+= 1
elif latest_type == 1:
# Subheading
headings.append({'data': content, 'type': 'subheading'})
latest_type = 2
stack.append({'SH'+str(shCounter) : left_value})
shCounter+= 1
elif latest_type == 3:
isParsed = False
while len(stack) != 0:
top = stack[-1]
key = list(top.keys())[0]
val = top[key]
if left_value < val:
stack.pop()
elif left_value == val:
if key.startswith('H'):
stack.append({'H'+str(hCounter) : left_value})
hCounter+= 1
headings.append({'data': content, 'type': 'heading'})
latest_type = 1
isParsed = True
elif key.startswith('SH'):
stack.append({'SH'+str(shCounter) : left_value})
shCounter+= 1
headings.append({'data': content, 'type': 'subheading'})
latest_type = 2
isParsed = True
break
elif left_value > val:
if key.startswith('H'):
stack.append({'SH'+str(shCounter) : left_value})
shCounter+= 1
headings.append({'data': content, 'type': 'subheading'})
latest_type = 2
isParsed = True
elif key.startswith('SH'):
if latest_type == 3:
latest_dict = headings[-1]
latest_dict['data'] += raw
headings[-1] = latest_dict
else:
headings.append({'data': raw, 'type': 'content'})
latest_type = 3
isParsed = True
break
if len(stack) == 0 and isParsed == False:
stack.append({'H'+str(hCounter) : left_value})
hCounter+= 1
headings.append({'data': content, 'type': 'heading'})
latest_type = 1
elif tag not in soup.find_all(lambda tag: tag.name == 'span' and '-bold' in tag.get('style').lower()):
# Content
if latest_type == 3:
latest_dict = headings[-1]
latest_dict['data'] += raw
headings[-1] = latest_dict
else:
headings.append({'data': raw, 'type': 'content'})
latest_type = 3
# Create a DataFrame
df = pd.DataFrame(headings)
# Add a new column with the title text
df['Topic'] = df['data'].where(df['type'] == 'title', '').ffill()
# Add a new column with the heading text
df['Categ'] = df['data'].where(df['type'] == 'heading', '').ffill()
# Add a new column with the subheading text
df['Sub_cat'] = df['data'].where(df['type'] == 'subheading', '').ffill()
# Add a new column with the content text
df['Text'] = df['data'].where(df['type'] == 'content', '')
# Drop 'data' column
df.drop('data', axis=1, inplace=True)
# Drop 'type' column
df.drop('type', axis=1, inplace=True)
df = df.replace(r'^\s*$',np.nan,regex=True)
df['Topic'].fillna(method="ffill",inplace=True)
df['Categ'].fillna(method="ffill",inplace=True)
df['Sub_cat'] = df.groupby('Categ', sort=False)['Sub_cat'].apply(lambda x: x.ffill())
current_time = currentTime()
df['Crawl_datetime'] = current_time
df = df[df['Text'].notna()]
df = df[["Crawl_datetime","Topic","Categ","Sub_cat","Text"]]
for i, row in df.iterrows():
processed_text = fix_unordered_lists(str(row['Text']))
df.loc[i, 'Processed_Text'] = processed_text
df.drop('Text', axis=1, inplace=True)
df.rename(columns={'Processed_Text': 'Text'}, inplace=True)
df = df.replace(r'^\s*$',np.nan,regex=True)
df = df[df['Text'].notna()]
for i, row in df.iterrows():
processed_text = fix_ordered_lists(str(row['Text']))
df.loc[i, 'Processed_Text'] = processed_text
df.drop('Text', axis=1, inplace=True)
df.rename(columns={'Processed_Text': 'Text'}, inplace=True)
csv_file_name = f"parser_{current_time}.csv"
# Save the DataFrame to a CSV file
df.to_csv(csv_file_name, index=False)
print("The CSV file path is:",csv_file_name)
def getSearchable(pdf_file_path):
# Convert each page of the PDF file to an image
images = convert_from_path(pdf_file_path)
# Create a new PDF file
output_pdf = PyPDF2.PdfWriter()
# Loop through each image and perform OCR using PyTesseract
for i, image in enumerate(images):
# Get a searchable PDF
pdf = pytesseract.image_to_pdf_or_hocr(image, extension='pdf')
# Add the page to the output PDF file
output_page = PyPDF2.PdfReader(io.BytesIO(pdf)).pages[0]
output_pdf.add_page(output_page)
current_time = currentTime()
pdf_file_name = f"searchable_{current_time}.pdf"
with open(pdf_file_name, 'wb') as f:
output_pdf.write(f)
return pdf_file_name
def parser():
pdf_file_path = getPDF()
print("The PDF file path is:", pdf_file_path)
if not isSearchable():
pdf_file_path = getSearchable(pdf_file_path)
processedHTML = pdftohtml(pdf_file_path)
htmltocsv(processedHTML)
parser()