menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Move HOCR_OK_LANGS to hocrtransform.py

Browse Source
This commit is contained in:
James R. Barlow
2021-05-23 01:15:53 -07:00
parent c87221a4e6
commit f3715daf15
2 changed files with 57 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/ocrmypdf/_validation.py
+1 -54
View File
@@ -32,65 +32,12 @@ from ocrmypdf.helpers import (
monotonic,
safe_symlink,
)
from ocrmypdf.hocrtransform import HOCR_OK_LANGS
from ocrmypdf.subprocess import check_external_program
# -------------
# External dependencies
# According to Wikipedia these languages are supported in the ISO-8859-1 character
# set, meaning reportlab can generate them and they are compatible with hocr,
# assuming Tesseract has the necessary languages installed. Note that there may
# not be language packs for them.
HOCR_OK_LANGS = frozenset(
[
# Languages fully covered by Latin-1:
'afr', # Afrikaans
'alb', # Albanian
'ast', # Leonese
'baq', # Basque
'bre', # Breton
'cos', # Corsican
'eng', # English
'eus', # Basque
'fao', # Faoese
'gla', # Scottish Gaelic
'glg', # Galician
'glv', # Manx
'ice', # Icelandic
'ind', # Indonesian
'isl', # Icelandic
'ita', # Italian
'ltz', # Luxembourgish
'mal', # Malay Rumi
'mga', # Irish
'nor', # Norwegian
'oci', # Occitan
'por', # Portugeuse
'roh', # Romansh
'sco', # Scots
'sma', # Sami
'spa', # Spanish
'sqi', # Albanian
'swa', # Swahili
'swe', # Swedish
'tgl', # Tagalog
'wln', # Walloon
# Languages supported by Latin-1 except for a few rare characters that OCR
# is probably not trained to recognize anyway:
'cat', # Catalan
'cym', # Welsh
'dan', # Danish
'deu', # German
'dut', # Dutch
'est', # Estonian
'fin', # Finnish
'fra', # French
'hun', # Hungarian
'kur', # Kurdish
'nld', # Dutch
'wel', # Welsh
]
)
DEFAULT_LANGUAGE = 'eng' # Enforce English hegemony
log = logging.getLogger(__name__)
src/ocrmypdf/hocrtransform.py
+56
View File
@@ -41,6 +41,62 @@ from reportlab.lib.colors import black, cyan, magenta, red
from reportlab.lib.units import inch
from reportlab.pdfgen.canvas import Canvas
# According to Wikipedia these languages are supported in the ISO-8859-1 character
# set, meaning reportlab can generate them and they are compatible with hocr,
# assuming Tesseract has the necessary languages installed. Note that there may
# not be language packs for them.
HOCR_OK_LANGS = frozenset(
[
# Languages fully covered by Latin-1:
'afr', # Afrikaans
'alb', # Albanian
'ast', # Leonese
'baq', # Basque
'bre', # Breton
'cos', # Corsican
'eng', # English
'eus', # Basque
'fao', # Faoese
'gla', # Scottish Gaelic
'glg', # Galician
'glv', # Manx
'ice', # Icelandic
'ind', # Indonesian
'isl', # Icelandic
'ita', # Italian
'ltz', # Luxembourgish
'mal', # Malay Rumi
'mga', # Irish
'nor', # Norwegian
'oci', # Occitan
'por', # Portugeuse
'roh', # Romansh
'sco', # Scots
'sma', # Sami
'spa', # Spanish
'sqi', # Albanian
'swa', # Swahili
'swe', # Swedish
'tgl', # Tagalog
'wln', # Walloon
# Languages supported by Latin-1 except for a few rare characters that OCR
# is probably not trained to recognize anyway:
'cat', # Catalan
'cym', # Welsh
'dan', # Danish
'deu', # German
'dut', # Dutch
'est', # Estonian
'fin', # Finnish
'fra', # French
'hun', # Hungarian
'kur', # Kurdish
'nld', # Dutch
'wel', # Welsh
]
)
Element = ElementTree.Element