python/pyocr: 0.4.6 -> 0.4.7

Upstream changes:

  * Tesseract 4.00.00alpha:
    * Version parsing: Ignore suffix (so '4.00.00alpha' == (4, 0, 0))
    * Libtesseract: Load libtesseract.so.4 instead of libtesseract.so.3
      if available
  * Support for Tesseract 3.05.00:
    * Builders: Split field 'tess_conf' into 'tess_flags' and 'tess_conf'
    * Libtesseract: If available, use
                    TessBaseAPIDetectOrientationScript() instead of
                    TessBaseAPIDetectOS
  * Libtesseract:
    * Workaround: Prevents possible segfault in image_to_string() when
                  the target language is not available

Full upstream change log can be found at:

https://github.com/openpaperwork/pyocr/blob/b006123d1d002711b9/ChangeLog

The tesseract.patch for supporting Tesseract version 3.05.00 has been
applied upstream and we can safely drop it.

We now use substituteInPlace in conjunction with a patch to insert the
relevant store paths instead of sed, so it's less fragile whenever we
have upstream changes in handling of these paths.

I've tested this by reverting 48a941e29f and applying a build
fix patch of Cuneiform 1.1.0 from Arch Linux, because right now
Cuneiform is an experimental version that can't be fixed on behalf of
pyocr (the reason is that pyocr needs to get a list of languages, which
doesn't work in that version anymore).

In addition to that I've successfully built paperwork-backend which by
now is the one package which depends on pyocr. However, I didn't do
runtime tests of Paperwork.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @7c6f434c
This commit is contained in:
aszlig 2017-09-02 05:18:38 +02:00
parent 3086fc7f83
commit ca1ea69972
No known key found for this signature in database
GPG Key ID: 1DE8E48E57DB5436
3 changed files with 129 additions and 329 deletions

View File

@ -4,30 +4,35 @@
buildPythonPackage rec { buildPythonPackage rec {
name = "pyocr-${version}"; name = "pyocr-${version}";
version = "0.4.6"; version = "0.4.7";
# Don't fetch from PYPI because it doesn't contain tests. # Don't fetch from PYPI because it doesn't contain tests.
src = fetchFromGitHub { src = fetchFromGitHub {
owner = "jflesch"; owner = "jflesch";
repo = "pyocr"; repo = "pyocr";
rev = version; rev = version;
sha256 = "0amyhkkm400qzbw65ivyzrzxl2r7vxqgsgqm7ml95m7gwkwhnzz0"; sha256 = "1iw73r8yrgjf8g00yzpz62ymqbf89cqhyhl9g430srmsrq7mn2yd";
}; };
patches = [ ./tesseract.patch ]; NIX_CUNEIFORM_CMD = "${cuneiform}/bin/cuneiform";
NIX_CUNEIFORM_DATA = "${cuneiform}/share/cuneiform";
NIX_LIBTESSERACT_PATH = "${tesseract}/lib/libtesseract.so";
NIX_TESSDATA_PREFIX = "${tesseract}/share/tessdata";
NIX_TESSERACT_CMD = "${tesseract}/bin/tesseract";
patches = [ ./paths.patch ];
postPatch = '' postPatch = ''
sed -i \ substituteInPlace src/pyocr/cuneiform.py \
-e 's,^\(TESSERACT_CMD *= *\).*,\1"${tesseract}/bin/tesseract",' \ --subst-var NIX_CUNEIFORM_CMD \
-e 's,^\(CUNEIFORM_CMD *= *\).*,\1"${cuneiform}/bin/cuneiform",' \ --subst-var NIX_CUNEIFORM_CMD
-e '/^CUNIFORM_POSSIBLE_PATHS *= *\[/,/^\]$/ {
c CUNIFORM_POSSIBLE_PATHS = ["${cuneiform}/share/cuneiform"]
}' src/pyocr/{tesseract,cuneiform}.py
sed -i -r \ substituteInPlace src/pyocr/tesseract.py \
-e 's,"libtesseract\.so\.3","${tesseract}/lib/libtesseract.so",' \ --subst-var NIX_TESSERACT_CMD
-e 's,^(TESSDATA_PREFIX *=).*,\1 "${tesseract}/share/tessdata",' \
src/pyocr/libtesseract/tesseract_raw.py substituteInPlace src/pyocr/libtesseract/tesseract_raw.py \
--subst-var NIX_TESSDATA_PREFIX \
--subst-var NIX_LIBTESSERACT_PATH
# Disable specific tests that are probably failing because of this issue: # Disable specific tests that are probably failing because of this issue:
# https://github.com/jflesch/pyocr/issues/52 # https://github.com/jflesch/pyocr/issues/52

View File

@ -0,0 +1,111 @@
diff --git a/src/pyocr/cuneiform.py b/src/pyocr/cuneiform.py
index a461d92..1f2b914 100644
--- a/src/pyocr/cuneiform.py
+++ b/src/pyocr/cuneiform.py
@@ -27,13 +27,9 @@ from . import error
from . import util
-# CHANGE THIS IF CUNEIFORM IS NOT IN YOUR PATH, OR IS NAMED DIFFERENTLY
-CUNEIFORM_CMD = 'cuneiform'
+CUNEIFORM_CMD = '@NIX_CUNEIFORM_CMD@'
-CUNEIFORM_DATA_POSSIBLE_PATHS = [
- "/usr/local/share/cuneiform",
- "/usr/share/cuneiform",
-]
+CUNEIFORM_DATA_POSSIBLE_PATHS = ['@NIX_CUNEIFORM_DATA@']
LANGUAGES_LINE_PREFIX = "Supported languages: "
LANGUAGES_SPLIT_RE = re.compile("[^a-z]")
diff --git a/src/pyocr/libtesseract/tesseract_raw.py b/src/pyocr/libtesseract/tesseract_raw.py
index b4e7bda..47505f7 100644
--- a/src/pyocr/libtesseract/tesseract_raw.py
+++ b/src/pyocr/libtesseract/tesseract_raw.py
@@ -1,55 +1,13 @@
import ctypes
import logging
import os
-import sys
from ..error import TesseractError
logger = logging.getLogger(__name__)
-TESSDATA_PREFIX = os.getenv('TESSDATA_PREFIX', None)
-libnames = []
-
-if getattr(sys, 'frozen', False):
- # Pyinstaller integration
- libnames += [os.path.join(sys._MEIPASS, "libtesseract-4.dll")]
- libnames += [os.path.join(sys._MEIPASS, "libtesseract-3.dll")]
- tessdata = os.path.join(sys._MEIPASS, "data")
- if not os.path.exists(os.path.join(tessdata, "tessdata")):
- logger.warning(
- "Running from container, but no tessdata ({}) found !".format(
- tessdata
- )
- )
- else:
- TESSDATA_PREFIX = tessdata
-
-
-if sys.platform[:3] == "win":
- libnames += [
- # Jflesch> Don't they have the equivalent of LD_LIBRARY_PATH on
- # Windows ?
- "../vs2010/DLL_Release/libtesseract302.dll",
- "libtesseract302.dll",
- "C:\\Program Files (x86)\\Tesseract-OCR\\libtesseract-4.dll",
- "C:\\Program Files (x86)\\Tesseract-OCR\\libtesseract-3.dll",
- ]
-else:
- libnames += [
- "libtesseract.so.4",
- "libtesseract.so.3",
- ]
-
-
-g_libtesseract = None
-
-for libname in libnames:
- try:
- g_libtesseract = ctypes.cdll.LoadLibrary(libname)
- break
- except OSError:
- pass
+g_libtesseract = ctypes.cdll.LoadLibrary('@NIX_LIBTESSERACT_PATH@')
class PageSegMode(object):
@@ -326,12 +284,11 @@ def init(lang=None):
try:
if lang:
lang = lang.encode("utf-8")
- prefix = None
- if TESSDATA_PREFIX:
- prefix = TESSDATA_PREFIX.encode("utf-8")
+ prefix = os.getenv('TESSDATA_PREFIX', '@NIX_TESSDATA_PREFIX@')
+ os.environ['TESSDATA_PREFIX'] = prefix
g_libtesseract.TessBaseAPIInit3(
ctypes.c_void_p(handle),
- ctypes.c_char_p(prefix),
+ ctypes.c_char_p(prefix.encode('utf-8')),
ctypes.c_char_p(lang)
)
g_libtesseract.TessBaseAPISetVariable(
diff --git a/src/pyocr/tesseract.py b/src/pyocr/tesseract.py
index c935881..7139ffe 100755
--- a/src/pyocr/tesseract.py
+++ b/src/pyocr/tesseract.py
@@ -31,8 +31,7 @@ from .builders import DigitBuilder # backward compatibility
from .error import TesseractError # backward compatibility
from .util import digits_only
-# CHANGE THIS IF TESSERACT IS NOT IN YOUR PATH, OR IS NAMED DIFFERENTLY
-TESSERACT_CMD = 'tesseract.exe' if os.name == 'nt' else 'tesseract'
+TESSERACT_CMD = '@NIX_TESSERACT_CMD@'
TESSDATA_EXTENSION = ".traineddata"

View File

@ -1,316 +0,0 @@
This patch is required for pyocr to work with Tesseract version 3.05.00
and has been submitted upstream at the following URL:
https://github.com/jflesch/pyocr/pull/62
diff --git a/src/pyocr/builders.py b/src/pyocr/builders.py
index 73c964d..20f390c 100644
--- a/src/pyocr/builders.py
+++ b/src/pyocr/builders.py
@@ -240,8 +240,10 @@ class BaseBuilder(object):
cuneiform_args : Arguments passed to the Cuneiform command line.
"""
- def __init__(self, file_extensions, tesseract_configs, cuneiform_args):
+ def __init__(self, file_extensions, tesseract_flags, tesseract_configs,
+ cuneiform_args):
self.file_extensions = file_extensions
+ self.tesseract_flags = tesseract_flags
self.tesseract_configs = tesseract_configs
self.cuneiform_args = cuneiform_args
@@ -298,7 +300,7 @@ class TextBuilder(BaseBuilder):
def __init__(self, tesseract_layout=3, cuneiform_dotmatrix=False,
cuneiform_fax=False, cuneiform_singlecolumn=False):
file_ext = ["txt"]
- tess_conf = ["-psm", str(tesseract_layout)]
+ tess_flags = ["-psm", str(tesseract_layout)]
cun_args = ["-f", "text"]
# Add custom cuneiform parameters if needed
for par, arg in [(cuneiform_dotmatrix, "--dotmatrix"),
@@ -306,7 +308,7 @@ class TextBuilder(BaseBuilder):
(cuneiform_singlecolumn, "--singlecolumn")]:
if par:
cun_args.append(arg)
- super(TextBuilder, self).__init__(file_ext, tess_conf, cun_args)
+ super(TextBuilder, self).__init__(file_ext, tess_flags, [], cun_args)
self.tesseract_layout = tesseract_layout
self.built_text = []
@@ -540,9 +542,11 @@ class WordBoxBuilder(BaseBuilder):
def __init__(self, tesseract_layout=1):
file_ext = ["html", "hocr"]
- tess_conf = ["hocr", "-psm", str(tesseract_layout)]
+ tess_flags = ["-psm", str(tesseract_layout)]
+ tess_conf = ["hocr"]
cun_args = ["-f", "hocr"]
- super(WordBoxBuilder, self).__init__(file_ext, tess_conf, cun_args)
+ super(WordBoxBuilder, self).__init__(file_ext, tess_flags, tess_conf,
+ cun_args)
self.word_boxes = []
self.tesseract_layout = tesseract_layout
@@ -614,9 +618,11 @@ class LineBoxBuilder(BaseBuilder):
def __init__(self, tesseract_layout=1):
file_ext = ["html", "hocr"]
- tess_conf = ["hocr", "-psm", str(tesseract_layout)]
+ tess_flags = ["-psm", str(tesseract_layout)]
+ tess_conf = ["hocr"]
cun_args = ["-f", "hocr"]
- super(LineBoxBuilder, self).__init__(file_ext, tess_conf, cun_args)
+ super(LineBoxBuilder, self).__init__(file_ext, tess_flags, tess_conf,
+ cun_args)
self.lines = []
self.tesseract_layout = tesseract_layout
diff --git a/src/pyocr/libtesseract/tesseract_raw.py b/src/pyocr/libtesseract/tesseract_raw.py
index 0c2259a..f7ab309 100644
--- a/src/pyocr/libtesseract/tesseract_raw.py
+++ b/src/pyocr/libtesseract/tesseract_raw.py
@@ -263,11 +263,22 @@ if g_libtesseract:
]
g_libtesseract.TessDeleteText.restype = None
- g_libtesseract.TessBaseAPIDetectOS.argtypes = [
- ctypes.c_void_p, # TessBaseAPI*
- ctypes.POINTER(OSResults),
- ]
- g_libtesseract.TessBaseAPIDetectOS.restype = ctypes.c_bool
+ if hasattr(g_libtesseract, 'TessBaseAPIDetectOrientationScript'):
+ g_libtesseract.TessBaseAPIDetectOrientationScript.argtypes = [
+ ctypes.c_void_p, # TessBaseAPI*
+ ctypes.POINTER(ctypes.c_int), # orient_deg
+ ctypes.POINTER(ctypes.c_float), # orient_conf
+ ctypes.POINTER(ctypes.c_char_p), # script_name
+ ctypes.POINTER(ctypes.c_float), # script_conf
+ ]
+ g_libtesseract.TessBaseAPIDetectOrientationScript.restype = \
+ ctypes.c_bool
+ else:
+ g_libtesseract.TessBaseAPIDetectOS.argtypes = [
+ ctypes.c_void_p, # TessBaseAPI*
+ ctypes.POINTER(OSResults),
+ ]
+ g_libtesseract.TessBaseAPIDetectOS.restype = ctypes.c_bool
def init(lang=None):
@@ -526,15 +537,37 @@ def detect_os(handle):
global g_libtesseract
assert(g_libtesseract)
- results = OSResults()
- r = g_libtesseract.TessBaseAPIDetectOS(
- ctypes.c_void_p(handle),
- ctypes.pointer(results)
- )
- if not r:
- raise TesseractError("detect_orientation failed",
- "TessBaseAPIDetectOS() failed")
- return {
- "orientation": results.best_orientation_id,
- "confidence": results.best_oconfidence,
- }
+ # Use the new API function if it is available, because since Tesseract
+ # 3.05.00 the old API function _always_ returns False.
+ if hasattr(g_libtesseract, 'TessBaseAPIDetectOrientationScript'):
+ orientation_deg = ctypes.c_int(0)
+ orientation_confidence = ctypes.c_float(0.0)
+
+ r = g_libtesseract.TessBaseAPIDetectOrientationScript(
+ ctypes.c_void_p(handle),
+ ctypes.byref(orientation_deg),
+ ctypes.byref(orientation_confidence),
+ None, # script_name
+ None # script_confidence
+ )
+
+ if not r:
+ raise TesseractError("detect_orientation failed",
+ "TessBaseAPIDetectOrientationScript() failed")
+ return {
+ "orientation": round(orientation_deg.value / 90),
+ "confidence": orientation_confidence.value,
+ }
+ else: # old API (before Tesseract 3.05.00)
+ results = OSResults()
+ r = g_libtesseract.TessBaseAPIDetectOS(
+ ctypes.c_void_p(handle),
+ ctypes.pointer(results)
+ )
+ if not r:
+ raise TesseractError("detect_orientation failed",
+ "TessBaseAPIDetectOS() failed")
+ return {
+ "orientation": results.best_orientation_id,
+ "confidence": results.best_oconfidence,
+ }
diff --git a/src/pyocr/tesseract.py b/src/pyocr/tesseract.py
index 99b0121..658c96b 100755
--- a/src/pyocr/tesseract.py
+++ b/src/pyocr/tesseract.py
@@ -22,6 +22,8 @@ import os
import subprocess
import sys
import tempfile
+import contextlib
+import shutil
from . import builders
from . import error
@@ -62,9 +64,11 @@ class CharBoxBuilder(builders.BaseBuilder):
def __init__(self):
file_ext = ["box"]
+ tess_flags = []
tess_conf = ["batch.nochop", "makebox"]
cun_args = []
- super(CharBoxBuilder, self).__init__(file_ext, tess_conf, cun_args)
+ super(CharBoxBuilder, self).__init__(file_ext, tess_flags, tess_conf,
+ cun_args)
self.tesseract_layout = 1
@staticmethod
@@ -173,18 +177,19 @@ def detect_orientation(image, lang=None):
TesseractError --- if no script detected on the image
"""
_set_environment()
- with temp_file(".bmp") as input_file:
- command = [TESSERACT_CMD, input_file.name, 'stdout', "-psm", "0"]
+ with temp_dir() as tmpdir:
+ command = [TESSERACT_CMD, "input.bmp", 'stdout', "-psm", "0"]
if lang is not None:
command += ['-l', lang]
if image.mode != "RGB":
image = image.convert("RGB")
- image.save(input_file.name)
+ image.save(os.path.join(tmpdir, "input.bmp"))
proc = subprocess.Popen(command, stdin=subprocess.PIPE, shell=False,
startupinfo=g_subprocess_startup_info,
creationflags=g_creation_flags,
+ cwd=tmpdir,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT)
proc.stdin.close()
@@ -224,8 +229,8 @@ def get_available_builders():
]
-def run_tesseract(input_filename, output_filename_base, lang=None,
- configs=None):
+def run_tesseract(input_filename, output_filename_base, cwd=None, lang=None,
+ flags=None, configs=None):
'''
Runs Tesseract:
`TESSERACT_CMD` \
@@ -238,6 +243,8 @@ def run_tesseract(input_filename, output_filename_base, lang=None,
input_filename --- image to read
output_filename_base --- file name in which must be stored the result
(without the extension)
+ cwd --- Run Tesseract in the specified working directory or use current
+ one if None
lang --- Tesseract language to use (if None, none will be specified)
config --- List of Tesseract configs to use (if None, none will be
specified)
@@ -252,10 +259,13 @@ def run_tesseract(input_filename, output_filename_base, lang=None,
if lang is not None:
command += ['-l', lang]
+ if flags is not None:
+ command += flags
+
if configs is not None:
command += configs
- proc = subprocess.Popen(command,
+ proc = subprocess.Popen(command, cwd=cwd,
startupinfo=g_subprocess_startup_info,
creationflags=g_creation_flags,
stdout=subprocess.PIPE,
@@ -301,11 +311,18 @@ class ReOpenableTempfile(object):
self.name = None
-def temp_file(suffix):
- ''' Returns a temporary file '''
- if os.name == 'nt': # Windows
- return ReOpenableTempfile(suffix)
- return tempfile.NamedTemporaryFile(prefix='tess_', suffix=suffix)
+@contextlib.contextmanager
+def temp_dir():
+ """
+ A context manager for maintaining a temporary directory
+ """
+ # NOTE: Drop this as soon as we don't support Python 2.7 anymore, because
+ # since Python 3.2 there is a context manager called TemporaryDirectory().
+ path = tempfile.mkdtemp(prefix='tess_')
+ try:
+ yield path
+ finally:
+ shutil.rmtree(path)
def image_to_string(image, lang=None, builder=None):
@@ -329,23 +346,20 @@ def image_to_string(image, lang=None, builder=None):
if builder is None:
builder = builders.TextBuilder()
- with temp_file(".bmp") as input_file:
- with temp_file('') as output_file:
- output_file_name_base = output_file.name
-
+ with temp_dir() as tmpdir:
if image.mode != "RGB":
image = image.convert("RGB")
- image.save(input_file.name)
- (status, errors) = run_tesseract(input_file.name,
- output_file_name_base,
+ image.save(os.path.join(tmpdir, "input.bmp"))
+ (status, errors) = run_tesseract("input.bmp", "output", cwd=tmpdir,
lang=lang,
+ flags=builder.tesseract_flags,
configs=builder.tesseract_configs)
if status:
raise TesseractError(status, errors)
output_file_name = "ERROR"
for file_extension in builder.file_extensions:
- output_file_name = ('%s.%s' % (output_file_name_base,
+ output_file_name = ('%s.%s' % (os.path.join(tmpdir, "output"),
file_extension))
if not os.access(output_file_name, os.F_OK):
continue
diff --git a/tests/tests_libtesseract.py b/tests/tests_libtesseract.py
index ccddd07..1ac2a4b 100644
--- a/tests/tests_libtesseract.py
+++ b/tests/tests_libtesseract.py
@@ -33,8 +33,9 @@ class TestContext(unittest.TestCase):
(3, 3, 0),
(3, 4, 0),
(3, 4, 1),
+ (3, 5, 0),
), ("Tesseract does not have the expected version"
- " (3.4.0) ! Some tests will be skipped !"))
+ " (3.5.0) ! Some tests will be skipped !"))
def test_langs(self):
langs = libtesseract.get_available_languages()
diff --git a/tests/tests_tesseract.py b/tests/tests_tesseract.py
index e29c512..fa4d483 100644
--- a/tests/tests_tesseract.py
+++ b/tests/tests_tesseract.py
@@ -27,8 +27,9 @@ class TestContext(unittest.TestCase):
(3, 3, 0),
(3, 4, 0),
(3, 4, 1),
+ (3, 5, 0),
), ("Tesseract does not have the expected version"
- " (3.4.0) ! Some tests will be skipped !"))
+ " (3.5.0) ! Some tests will be skipped !"))
def test_langs(self):
langs = tesseract.get_available_languages()