Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/tesseract/src/training/unicharset/lang_model_helpers.h @ 2:b50eed0cc0ef upstream

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents
children
comparison
equal deleted inserted replaced
1:1d09e1dec1d9 2:b50eed0cc0ef
1 // Copyright 2017 Google Inc. All Rights Reserved.
2 // Author: rays@google.com (Ray Smith)
3 // Purpose: Collection of convenience functions to simplify creation of the
4 // unicharset, recoder, and dawgs for an LSTM model.
5
6 // Licensed under the Apache License, Version 2.0 (the "License");
7 // you may not use this file except in compliance with the License.
8 // You may obtain a copy of the License at
9 // http://www.apache.org/licenses/LICENSE-2.0
10 // Unless required by applicable law or agreed to in writing, software
11 // distributed under the License is distributed on an "AS IS" BASIS,
12 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 // See the License for the specific language governing permissions and
14 // limitations under the License.
15 #ifndef TESSERACT_TRAINING_LANG_MODEL_HELPERS_H_
16 #define TESSERACT_TRAINING_LANG_MODEL_HELPERS_H_
17
18 #include "export.h"
19
20 #include "serialis.h"
21 #include "tessdatamanager.h"
22 #include "unicharset.h"
23
24 #include <string>
25
26 namespace tesseract {
27
28 // Helper makes a filename (<output_dir>/<lang>/<lang><suffix>) and writes data
29 // to the file, using writer if not null, otherwise, a default writer.
30 // Default writer will overwrite any existing file, but a supplied writer
31 // can do its own thing. If lang is empty, returns true but does nothing.
32 // NOTE that suffix should contain any required . for the filename.
33 TESS_UNICHARSET_TRAINING_API
34 bool WriteFile(const std::string &output_dir, const std::string &lang, const std::string &suffix,
35 const std::vector<char> &data, FileWriter writer);
36 // Helper reads a file with optional reader and returns a string.
37 // On failure emits a warning message and returns and empty string.
38 TESS_UNICHARSET_TRAINING_API
39 std::string ReadFile(const std::string &filename, FileReader reader = nullptr);
40
41 // Helper writes the unicharset to file and to the traineddata.
42 bool WriteUnicharset(const UNICHARSET &unicharset, const std::string &output_dir,
43 const std::string &lang, FileWriter writer, TessdataManager *traineddata);
44 // Helper creates the recoder from the unicharset and writes it to the
45 // traineddata, with a human-readable form to file at:
46 // <output_dir>/<lang>/<lang>.charset_size=<num> for some num being the size
47 // of the re-encoded character set. The charset_size file is written using
48 // writer if not null, or using a default file writer otherwise, overwriting
49 // any existing content.
50 // If pass_through is true, then the recoder will be a no-op, passing the
51 // unicharset codes through unchanged. Otherwise, the recoder will "compress"
52 // the unicharset by encoding Hangul in Jamos, decomposing multi-unicode
53 // symbols into sequences of unicodes, and encoding Han using the data in the
54 // radical_table_data, which must be the content of the file:
55 // langdata/radical-stroke.txt.
56 bool WriteRecoder(const UNICHARSET &unicharset, bool pass_through, const std::string &output_dir,
57 const std::string &lang, FileWriter writer, std::string *radical_table_data,
58 TessdataManager *traineddata);
59
60 // The main function for combine_lang_model.cpp.
61 // Returns EXIT_SUCCESS or EXIT_FAILURE for error.
62 // unicharset: can be a hand-created file with incomplete fields. Its basic
63 // and script properties will be set before it is used.
64 // script_dir: should point to the langdata (github repo) directory.
65 // version_str: arbitrary version label.
66 // Output files will be written to <output_dir>/<lang>/<lang>.*
67 // If pass_through_recoder is true, the unicharset will be used unchanged as
68 // labels in the classifier, otherwise, the unicharset will be "compressed" to
69 // make the recognition task simpler and faster.
70 // The words/puncs/numbers lists may be all empty. If any are non-empty then
71 // puncs must be non-empty.
72 // lang_is_rtl indicates that the language is generally written from right
73 // to left (eg Arabic/Hebrew).
74 TESS_UNICHARSET_TRAINING_API
75 int CombineLangModel(const UNICHARSET &unicharset, const std::string &script_dir,
76 const std::string &version_str, const std::string &output_dir,
77 const std::string &lang, bool pass_through_recoder,
78 const std::vector<std::string> &words, const std::vector<std::string> &puncs,
79 const std::vector<std::string> &numbers, bool lang_is_rtl, FileReader reader,
80 FileWriter writer);
81
82 } // namespace tesseract
83
84 #endif // TESSERACT_TRAINING_LANG_MODEL_HELPERS_H_