Date: prev next · Thread: first prev next last
2013 Archives by date, by thread · List index


Hello Caolan,


Caolán McNamara wrote
No idea, IMO it would be worth bibisecting this to find where it changed
to re-examine if it was intentional or not. Would then likely be a good
candidate for a unit test to lock in whichever is the right behaviour.

If someone did this intentionally he should say it. ;)

As a Windows user I can’t bibisect, but I investigated. (I’ll try to
bibisect when I have the time.)

doProofreading is called only in the file
core/linguistic/source/gciterator.cxx
It’s used to call the grammar checker for a given language.
Few changes last months:
http://cgit.freedesktop.org/libreoffice/core/log/linguistic/source/gciterator.cxx

For testing purpose, I change the file Lightproof.py in LibreOffice
4.0\share\extensions\dict-en

I added the line:

    sys.stdout = open("D:\_lightproof_stdout.txt", "w")
    
before the class Lightproof.
And I added the lines:

        print("PARAGRAPH: ", rText)
        print(nStartOfSentencePos, nSuggestedSentenceEndPos,
rText[nStartOfSentencePos:nSuggestedSentenceEndPos])
        sys.stdout.flush()

at the beginning of the method doProofreading.
(See the code at the end of this mail.)

Then I opened LibreOffice and I copy/paste a paragraph in document whose
language is set to English (USA).

The paragraph is:

Called up for active service in 1939, Jackson served with No. 23 Squadron in
Australia before he was posted to the Middle East in November 1940. A
charity hockey game held to benefit former player Bill Heindl, Jr. in 1980
was the only occasion that hockey legends Bobby Orr and Wayne. The total
floor area of St. Michael's Cathedral is 2,740 square metres.


The result:
PARAGRAPH:  
0 0 
PARAGRAPH:  Called up for active service in 1939, Jackson served with No. 23
Squadron in Australia before he was posted to the Middle East in November
1940. A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne. The
total floor area of St. Michael's Cathedral is 2,740 square metres.
0 61 Called up for active service in 1939, Jackson served with No.
PARAGRAPH:  Called up for active service in 1939, Jackson served with No. 23
Squadron in Australia before he was posted to the Middle East in November
1940. A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne. The
total floor area of St. Michael's Cathedral is 2,740 square metres.
62 144 23 Squadron in Australia before he was posted to the Middle East in
November 1940.
PARAGRAPH:  Called up for active service in 1939, Jackson served with No. 23
Squadron in Australia before he was posted to the Middle East in November
1940. A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne. The
total floor area of St. Michael's Cathedral is 2,740 square metres.
145 284 A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne.
PARAGRAPH:  Called up for active service in 1939, Jackson served with No. 23
Squadron in Australia before he was posted to the Middle East in November
1940. A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne. The
total floor area of St. Michael's Cathedral is 2,740 square metres.
285 312 The total floor area of St.
PARAGRAPH:  Called up for active service in 1939, Jackson served with No. 23
Squadron in Australia before he was posted to the Middle East in November
1940. A charity hockey game held to benefit former player Bill Heindl, Jr.
in 1980 was the only occasion that hockey legends Bobby Orr and Wayne. The
total floor area of St. Michael's Cathedral is 2,740 square metres.
313 356 Michael's Cathedral is 2,740 square metres.


So the small paragraph is split this way:

  Called up for active service in 1939, Jackson served with No.
  23 Squadron in Australia before he was posted to the Middle East in
November 1940.
  A charity hockey game held to benefit former player Bill Heindl, Jr. in
1980 was the only occasion that hockey legends Bobby Orr and Wayne.
  The total floor area of St.
  Michael's Cathedral is 2,740 square metres.

1 paragraph of 3 sentences, split in 5 parts.
It splits after No., after St., but strangely not after Jr.

A paragraph is passed to the grammar checker as many times as the number of
sentences found. In this case, 5 times.

It’s not a proper behavior for the English language, and it provokes false
alarms in the French grammar checker. I didn’t test with other languages,
but I assume it’s probably not a good thing for them either, as we can’t be
sure a dot is the end of a sentence.

In LO 3.6, each paragraph was passed only once to the GC, and it was up to
the GC to split the paragraph or not.

Regards,
Olivier


** the modified Lightproof file **

# -*- encoding: UTF-8 -*-
# Lightproof grammar checker for LibreOffice and OpenOffice.org
# 2009-2012 (c) László Németh (nemeth at numbertext org), license: MPL 1.1 /
GPLv3+ / LGPLv3+

import uno, unohelper, os, sys, traceback
from lightproof_impl_en import locales
from lightproof_impl_en import pkg
import lightproof_impl_en
import lightproof_handler_en

from com.sun.star.linguistic2 import XProofreader, XSupportedLocales
from com.sun.star.linguistic2 import ProofreadingResult,
SingleProofreadingError
from com.sun.star.lang import XServiceInfo, XServiceName,
XServiceDisplayName
from com.sun.star.lang import Locale
# reload in obj.reload in Python 3
try:
    from obj import reload
except:
    pass

try:
    sys.stdout = open("D:\_lightproof_stdout.txt", "w")
except:
    pass

class Lightproof( unohelper.Base, XProofreader, XServiceInfo, XServiceName,
XServiceDisplayName, XSupportedLocales):

    def __init__( self, ctx, *args ):
        self.ctx = ctx
        self.ServiceName = "com.sun.star.linguistic2.Proofreader"
        self.ImplementationName = "org.libreoffice.comp.pyuno.Lightproof." +
pkg
        self.SupportedServiceNames = (self.ServiceName, )
        self.locales = []
        for i in locales:
            l = locales[i]
            self.locales += [Locale(l[0], l[1], l[2])]
        self.locales = tuple(self.locales)
        currentContext = uno.getComponentContext()
        lightproof_impl_en.SMGR = currentContext.ServiceManager
        lightproof_impl_en.spellchecker = \
           
lightproof_impl_en.SMGR.createInstanceWithContext("com.sun.star.linguistic2.SpellChecker",
currentContext)
        lightproof_handler_en.load(currentContext)

    # XServiceName method implementations
    def getServiceName(self):
        return self.ImplementationName

    # XServiceInfo method implementations
    def getImplementationName (self):
        return self.ImplementationName

    def supportsService(self, ServiceName):
        return (ServiceName in self.SupportedServiceNames)

    def getSupportedServiceNames (self):
        return self.SupportedServiceNames

    # XSupportedLocales
    def hasLocale(self, aLocale):
        if aLocale in self.locales:
            return True
        for i in self.locales:
            if (i.Country == aLocale.Country or i.Country == "") and
aLocale.Language == i.Language:
                return True
        return False

    def getLocales(self):
        return self.locales

    # XProofreader
    def isSpellChecker(self):
        return False

    def doProofreading(self, nDocId, rText, rLocale, nStartOfSentencePos,
nSuggestedSentenceEndPos, rProperties):

        print("PARAGRAPH: ", rText)
        print(nStartOfSentencePos, nSuggestedSentenceEndPos,
rText[nStartOfSentencePos:nSuggestedSentenceEndPos])
        sys.stdout.flush()
        
        aRes = uno.createUnoStruct(
"com.sun.star.linguistic2.ProofreadingResult" )
        aRes.aDocumentIdentifier = nDocId
        aRes.aText = rText
        aRes.aLocale = rLocale
        aRes.nStartOfSentencePosition = nStartOfSentencePos
        aRes.nStartOfNextSentencePosition = nSuggestedSentenceEndPos
        aRes.aProperties = ()
        aRes.xProofreader = self
        aRes.aErrors = ()
        if len(rProperties) > 0 and rProperties[0].Name == "Update":
            try:
                import lightproof_compile_en
                try:
                    code = lightproof_compile_en.c(rProperties[0].Value,
rLocale.Language, True)
                except Exception as e:
                    aRes.aText, aRes.nStartOfSentencePosition = e
                    return aRes
                path = lightproof_impl_en.get_path()
                f = open(path.replace("_impl", ""), "w")
                f.write("dic = %s" % code["rules"])
                f.close()
                if pkg in lightproof_impl_en.langrule:
                    mo = lightproof_impl_en.langrule[pkg]
                    reload(mo)
                    lightproof_impl_en.compile_rules(mo.dic)
                    lightproof_impl_en.langrule[pkg] = mo
                if "code" in code:
                    f = open(path, "r")
                    ft = f.read()
                    f.close()
                    f = open(path, "w")
                    f.write(ft[:ft.find("# [code]") + 8] + "\n" +
code["code"])
                    f.close()
                    try:
                        reload(lightproof_impl_en)
                    except Exception as e:
                        aRes.aText = e.args[0]
                        if e.args[1][3] == "": # "expected an indented
block" (end of file)
                            aRes.nStartOfSentencePosition =
len(rText.split("\n"))
                        else:
                            aRes.nStartOfSentencePosition =
rText.split("\n").index(e.args[1][3][:-1]) + 1
                        return aRes
                aRes.aText = ""
                return aRes
            except:
                if 'PYUNO_LOGLEVEL' in os.environ:
                    print(traceback.format_exc())

        l = rText[nSuggestedSentenceEndPos:nSuggestedSentenceEndPos+1]
        while l == " ":
            aRes.nStartOfNextSentencePosition =
aRes.nStartOfNextSentencePosition + 1
            l =
rText[aRes.nStartOfNextSentencePosition:aRes.nStartOfNextSentencePosition+1]
        if aRes.nStartOfNextSentencePosition == nSuggestedSentenceEndPos and
l!="":
            aRes.nStartOfNextSentencePosition = nSuggestedSentenceEndPos + 1
        aRes.nBehindEndOfSentencePosition =
aRes.nStartOfNextSentencePosition

        try:
            aRes.aErrors = lightproof_impl_en.proofread( nDocId, rText,
rLocale, \
                nStartOfSentencePos, aRes.nBehindEndOfSentencePosition,
rProperties)
        except Exception as e:
            if len(rProperties) > 0 and rProperties[0].Name == "Debug" and
len(e.args) == 2:
                aRes.aText, aRes.nStartOfSentencePosition = e
            else:
                if 'PYUNO_LOGLEVEL' in os.environ:
                    print(traceback.format_exc())
        return aRes

    def ignoreRule(self, rid, aLocale):
        lightproof_impl_en.ignore[rid] = 1

    def resetIgnoreRules(self):
        lightproof_impl_en.ignore = {}

    # XServiceDisplayName
    def getServiceDisplayName(self, aLocale):
        return lightproof_impl_en.name

g_ImplementationHelper = unohelper.ImplementationHelper()
g_ImplementationHelper.addImplementation( Lightproof, \
    "org.libreoffice.comp.pyuno.Lightproof." + pkg,
    ("com.sun.star.linguistic2.Proofreader",),)

g_ImplementationHelper.addImplementation(
lightproof_handler_en.LightproofOptionsEventHandler, \
    "org.libreoffice.comp.pyuno.LightproofOptionsEventHandler." + pkg,
    ("com.sun.star.awt.XContainerWindowEventHandler",),)





--
View this message in context: 
http://nabble.documentfoundation.org/Grammar-checker-Undocumented-change-in-the-API-for-LO-4-tp4030639p4039657.html
Sent from the Dev mailing list archive at Nabble.com.

Context


Privacy Policy | Impressum (Legal Info) | Copyright information: Unless otherwise specified, all text and images on this website are licensed under the Creative Commons Attribution-Share Alike 3.0 License. This does not include the source code of LibreOffice, which is licensed under the Mozilla Public License (MPLv2). "LibreOffice" and "The Document Foundation" are registered trademarks of their corresponding registered owners or are in actual use as trademarks in one or more countries. Their respective logos and icons are also subject to international copyright laws. Use thereof is explained in our trademark policy.