File size: 4,961 Bytes
aeb6dbc
 
 
 
 
 
 
 
 
 
 
 
8bc2fc9
aeb6dbc
 
 
82adb1c
d624a4e
872406f
d624a4e
aeb6dbc
 
 
 
 
 
6e567cd
 
 
aeb6dbc
 
 
 
 
 
 
6e567cd
aeb6dbc
 
 
6e567cd
8bc2fc9
6e567cd
 
aeb6dbc
6e567cd
 
 
aeb6dbc
6e567cd
aeb6dbc
 
 
 
 
 
0404a52
 
aeb6dbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
872406f
 
 
 
aeb6dbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0842c1e
aeb6dbc
d624a4e
aeb6dbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0404a52
aeb6dbc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#  Licensed under the Apache License, Version 2.0 (the "License");
#  you may not use this file except in compliance with the License.
#  You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
#  Unless required by applicable law or agreed to in writing, software
#  distributed under the License is distributed on an "AS IS" BASIS,
#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#  See the License for the specific language governing permissions and
#  limitations under the License.
#
import logging
from tika import parser
from io import BytesIO
import re

from deepdoc.parser.utils import get_text
from rag.app import naive
from rag.nlp import rag_tokenizer, tokenize
from deepdoc.parser import PdfParser, ExcelParser, PlainParser, HtmlParser


class Pdf(PdfParser):
    def __call__(self, filename, binary=None, from_page=0,
                 to_page=100000, zoomin=3, callback=None):
        from timeit import default_timer as timer
        start = timer()
        callback(msg="OCR started")
        self.__images__(
            filename if not binary else binary,
            zoomin,
            from_page,
            to_page,
            callback
        )
        callback(msg="OCR finished ({:.2f}s)".format(timer() - start))

        start = timer()
        self._layouts_rec(zoomin, drop=False)
        callback(0.63, "Layout analysis ({:.2f}s)".format(timer() - start))
        logging.debug("layouts cost: {}s".format(timer() - start))

        start = timer()
        self._table_transformer_job(zoomin)
        callback(0.65, "Table analysis ({:.2f}s)".format(timer() - start))

        start = timer()
        self._text_merge()
        callback(0.67, "Text merged ({:.2f}s)".format(timer() - start))
        tbls = self._extract_table_figure(True, zoomin, True, True)
        self._concat_downward()

        sections = [(b["text"], self.get_position(b, zoomin))
                    for i, b in enumerate(self.boxes)]
        for (img, rows), poss in tbls:
            if not rows:
                continue
            sections.append((rows if isinstance(rows, str) else rows[0],
                             [(p[0] + 1 - from_page, p[1], p[2], p[3], p[4]) for p in poss]))
        return [(txt, "") for txt, _ in sorted(sections, key=lambda x: (
            x[-1][0][0], x[-1][0][3], x[-1][0][1]))], None


def chunk(filename, binary=None, from_page=0, to_page=100000,
          lang="Chinese", callback=None, **kwargs):
    """
        Supported file formats are docx, pdf, excel, txt.
        One file forms a chunk which maintains original text order.
    """

    eng = lang.lower() == "english"  # is_english(cks)

    if re.search(r"\.docx$", filename, re.IGNORECASE):
        callback(0.1, "Start to parse.")
        sections, tbls = naive.Docx()(filename, binary)
        sections = [s for s, _ in sections if s]
        for (_, html), _ in tbls:
            sections.append(html)
        callback(0.8, "Finish parsing.")

    elif re.search(r"\.pdf$", filename, re.IGNORECASE):
        pdf_parser = Pdf() if kwargs.get(
            "parser_config", {}).get(
            "layout_recognize", True) else PlainParser()
        sections, _ = pdf_parser(
            filename if not binary else binary, to_page=to_page, callback=callback)
        sections = [s for s, _ in sections if s]

    elif re.search(r"\.xlsx?$", filename, re.IGNORECASE):
        callback(0.1, "Start to parse.")
        excel_parser = ExcelParser()
        sections = excel_parser.html(binary, 1000000000)

    elif re.search(r"\.(txt|md|markdown)$", filename, re.IGNORECASE):
        callback(0.1, "Start to parse.")
        txt = get_text(filename, binary)
        sections = txt.split("\n")
        sections = [s for s in sections if s]
        callback(0.8, "Finish parsing.")

    elif re.search(r"\.(htm|html)$", filename, re.IGNORECASE):
        callback(0.1, "Start to parse.")
        sections = HtmlParser()(filename, binary)
        sections = [s for s in sections if s]
        callback(0.8, "Finish parsing.")

    elif re.search(r"\.doc$", filename, re.IGNORECASE):
        callback(0.1, "Start to parse.")
        binary = BytesIO(binary)
        doc_parsed = parser.from_buffer(binary)
        sections = doc_parsed['content'].split('\n')
        sections = [s for s in sections if s]
        callback(0.8, "Finish parsing.")

    else:
        raise NotImplementedError(
            "file type not supported yet(doc, docx, pdf, txt supported)")

    doc = {
        "docnm_kwd": filename,
        "title_tks": rag_tokenizer.tokenize(re.sub(r"\.[a-zA-Z]+$", "", filename))
    }
    doc["title_sm_tks"] = rag_tokenizer.fine_grained_tokenize(doc["title_tks"])
    tokenize(doc, "\n".join(sections), eng)
    return [doc]


if __name__ == "__main__":
    import sys

    def dummy(prog=None, msg=""):
        pass

    chunk(sys.argv[1], from_page=0, to_page=10, callback=dummy)