Link Grammar Parser Older Changes
Summary of Older Changes
Version 5.10.5 (17 June 2022)
This version is the first to contain the final version of the Thai
dictionary.
- Updated Docker files. #1288
- English dict: broader handling of ellipsis.
- Updated Thai dicts: #1292, #1294
- Fix Thai regex's to work even with basic C++ regex lib. #1297
- Performance improvements. #1298
Version 5.10.4 (4 March 2022)
This is a notable and important release, as it is the first to
include a complete Thai dictionary!
- English dict: fix relative clause, per mailing list.
- Remove assorted length restrictions on word-size. #1283
- Add missing files for building link-generator on Windows. #1285
- Strip the internally added "._I" from subscripted idioms. #1287
- New: Provisional Thai dictionary. #1279
Version 5.10.3 (14 February 2022)
- Remove `node.js/package-lock.json` from tarball distribution. #1251
- Fix Windows MSVC build break. #1253
- Fix memory leak in the "!" link-parser command. #1256
- Add C++ regex support. It is now the default for MSVC builds. #1258
- Fix spell-guess for run-on words. #1249
- Port link-generator to MS-Windows. #1269
- Fix apostrophe handling for link-generator w/sqlite3 dicts. #1276
Version 5.10.2 (16 September 2021)
- Fix python install path.
- Fix size in brand-new `link-generator` (hits 32-bit & ARM) #1247
Version 5.10.1 (7 September 2021)
- Fix perl bindings build fail. #1248
Version 5.10.0 (4 September 2021)
The minor version number has been bumped because of a change to the
link types used for idioms. Subscripts with an underbar are now
reserved.
Users of the unsupervised language learning project will need
this version. It contains fixes for random corpora generation.
- Expanded English vocabulary
- Support dictionary "#define allow-duplicate-words true". #1204
- Fix crash for sentences containing wildcard words. #1206
- Connector names starting with "ID" are no longer reserved. #1208
- Connector names starting with underbar are reserved for internal use.
- ".I" subscripts are no longer reserved; "._" subscripts are reserved.
These last three changes introduce linkage incompatibilities.
- Fix parsing with nulls when using an sqlite3 dictionary.
- Fix regexes for NetBSD when using libc regexes. #1223
- English dict: fix many "how?" questions.
- English dict: fix conditional sentences #1240
Version 5.9.1 (28 April 2021)
Emergency bug fix.
- Fix build break when SQLite3 is not installed. #1195
Version 5.9.0 (25 April 2021)
An experimental sentence generator has been added. This generator will
create new grammatical sentences, based on a "fill in the blanks"
approach to specifying a template sentence. The dictionary is scanned
for any suitable words that might fit into wild-card locations; the
resulting sentences are then printed. This is particularly useful for
generating random corpora of grammatically valid sentences.
- Use #define for custom configuration in dictionaries. #1128
- Panic-mode fixes and extensions. In link-parser see !help panic_variables.
- English dict: fix silly mistake with "I love cats and dogs".
- Disable maintainer-mode in `configure.ac`.
- Fix very rare crash/corruption introduced in v.5.8.1 #1142
- English dict: fix problems with "just/only".
- English dict: work on hesitation markers.
- Fix multi-threading mem-leak. #1149
- Provide emscripten javascript wrapper for the command-line parser.
- Public API shared library entry points exported automatically. #1182
- Provide bindings for the Vala programming language.
- Increase number of allowed idiom expressions. #1187
- Replace O(n^2) idiom loading algo by an O(n log n) algo. #1194
- Disable SAT solver by default.
- New tool: Sentence generator! This is an experimental prototype.
Version 5.8.1 (8 January 2021)
Assorted fixes.
- Fix macOS/SunOS build break.
- English dict: fix numerical identifiers used as adjectives.
- English dict: fix post-posed Latin adjectival modifiers.
- Merge upstream gentoo patches. #1102
- Make -O3 default for CFLAGS/CXXFLAGS, but overridable by the user.
- English dict: fix look_at, listen_to person-action
- English dict: fix verb "felt" with object-action.
- English dict: fix why-perform-action questions.
- Fix race condition in spell-guesser. #1122 #1123
Version 5.8.0 (28 February 2020)
Notable changes include: inclusion of javascript node.js bindings;
the obsoleting of python2, improved English dictionaries, and most
interestingly, an experimental interface for dialects. With this
interface, one can provide alternative weightings that emphasize
the type of speech that might be common in limited geographical
areas, and would not be considered to be commonplace. For example,
one can provide weightings for Irish-American, urban English, and
newspaper-headline English which might otherwise interfere with
ordinary parsing of mainstream English.
- Java bindings: Remove the obsolete senses API.
- swig-4.0 compatibility bug fix.
- English dict: Fixes to support questions ending in WH-words.
- Copy (merge) Richard van der Dys `node.js` bindings.
- English dict: Provide head and tail markers for all conjunctions.
- Remove the Python 2 bindings.
- Add dialect support to the library.
- English dict: support for archaic/poetic abbreviations
- English dict: introduce OH link for vocatives/invocations.
- English dict: improved parsing of imperatives.
- Add !!word/ link-parser command for displaying extended word dict info.
Version 5.7.0 (13 September 2019)
This version has one quite remarkable change: the parsing of long
sentences has been improved by a factor of 3x or 4x, and thus, the
parse speed of many "typical" texts is doubled, or more. Two other
important fixes are for broken 32-bit support, and for Windows.
- Minor efficiency improvements to the SQL-backed dictionary.
- Incompatible change to the Exp traversal API.
- Remove the obsolete and unsupported "corpus statistics" code.
- Major performance improvement (3x-4x) for long sentences. #996
- Fix broken build on Windows.
- Convert Windows build to Visual Studio 2019.
- Fix a bug that causes random results on 32-bit systems. #1000
- Fix a bug that could cause missing linkages on some systems. #1007
Version 5.6.2 (24 June 2019)
This adds a missing shared-library symbol that broke the opencog build!
- Bug-fix the SQL-backed dictionary.
- Add missing public symbol to shlib export list.
- English dict: additions of paraphrasing verbs.
Version 5.6.1 (27 May 2019)
Important! This is an important update, as it more than
doubles the performance across a broad range of different input texts.
Kudos to Amir for this amazing work, as he took something that seemed
quite fast to begin with, and squeezed out an honest factor of two
from it! This is unusual in mature software.
- Performance improvement (approx 20%) in expressions #882.
- Performance improvement (approx 10%) by disjunct/connectors pools #896.
- Performance improvement (4-10% for English) by faster power-pruning #903.
- Fix a bug in trailing connectors encoding (may cause bad linkages).
- Fix inability to form linkage when first word is disconnected.
- English dict: fix use of quotations with paraphrasing verbs.
- English dict: fix broken usage of "have not".
- Performance improvement (approx 16%) for long sentences. #931
- Performance improvement (approx 20%) for long sentences. #939
Version 5.6.0 (4 January 2019)
- Improve Windows support.
- Fix dict cost reading under user locales with comma decimal separator.
- Support using the pcre2 regex package (configured by default if available).
- Add "-with-regexlib=pcre2|tre|regex|c" to "configure".
- Fix and document building on FreeBSD.
- Major documentation update for building with Cygwin.
- Revise the manpage.
- Remove the experimental Viterbi code.
- Revise the SAT parser cost model to align it with the classic parser.
- Implement a strict check on connector name.
- Major revision of the SWIG interface file; wrap all the library functions.
- Fix linkage_get_disjunct_*() when parse-option display_morphology is true.
- Change library and python-bindings default for display_morphology to true.
- Drastic speedup for long sentences (hash encoding of trailing connectors).
- English dict: Support locative replies/declarations.
- English dict: broaden support for misc paraphrasing verbs.
- English dict: fix relativized paraphrasing.
- English dict: fix comparative-style conjunctions.
Version 5.5.1 (27 July 2018)
- Fix broken Java bindings build.
- English dict: Fix clause openers with questions.
- English dict: Various misc fixes.
- English dict: Various paraphrasing verbs
- Bring the SQL-backed dict to production state.
- Convert MSVC build to MSVC15 (Visual Studio 2017).
- Restore the repeatability of the produced linkages.
Version 5.5.0 (29 April 2018)
- Fix accidental API breakage that impacts OpenCog.
- Fix memory leak when parsing with null links.
- Python bindings: Add an optional parse-option argument to parse().
- Add an extended version API and use it in "link-parser --version".
- Fix spurious errors if the last dict line is a comment.
- Fix garbage report if EOF encountered in a quoted dict word.
- Fix garbage report if whitespace encountered in a quoted dict word.
- Add a per-command help in link-parser.
- Add a command line completion in link-parser.
- Enable build of word-graph printing support by default.
- Add idiom lookup in link-parser's dict lookup command (!!idiom_here).
- Improve handling of quoted words (e.g. single words in "scare quotes").
- Fix random selection of linkages so that it's actually random.
Version 5.4.4 (11 March 2018)
- Dictionary loading now thread safe.
- Fix post-nominal modifiers used with pronouns.
- Fix comparative openers.
- Fix given-name single-letter abbreviations.
- Fix conjoined questions and conjoined WH-statements.
- Fix conditional sentences.
- Fix misc comparatives.
- Fix crash on invalid UTF-8 input.
- Fix many predicative adjective uses.
- Fix many paraphrasing-type constructions.
- Minor cleanup of word-lists.
- New dict definition LENGTH-LIMIT-n to limit connector link length to n.
- Speed up parsing of Russian by factor of 2x.
- Add assorted technical vocabulary (#680)
- Fix conjoined infinitives.
Version 5.4.3 (4 January 2018)
- Fix man page installation (actually broken from 5.3.0).
- Add "thither" to the English dictionary.
- Fix printing inf loop for very narrow screen widths.
- Some Windows code clean up.
- Remove trailing blanks from the linkage diagram.
- Fix square area and cubic volume measurements (English dict).
- Fix assorted exclamations and responses (English dict).
- Fix displaying random linkages on Windows.
- Fix unit tokenization to remove ambiguity.
- Fix utf8-related bug on Windows that could affect printing.
- Add missing affix file, needed for the 'any' language.
Version 5.4.2 (19 October 2017)
- Fix man page build (broken in 5.4.1).
Version 5.4.1 (18 October 2017)
- Fix man page installation (broken in 5.3.8).
- Add affix-class MPUNC for splitting at intra-word punctuation.
- Fix crash when there is no PP info.
- Fix a stack buffer overflow.
- Eliminate hard-wired linkage diagram size limitations.
- Fix an unintended clipping of the linkage-limit option to 250000.
Version 5.4.0 (26 July 2017)
Notable: This reorganizes the source code into subdirectories,
grouped according to the processing stage. This should make it easier
to understand what the major components are, and which files & functions
are a part of each component.
- Fix for missing locale info in Windows XP.
- Empty out the post-processing tables for the any, ady, amy languages
- Remove left_print_string() from the API.
- Recover pp_lexer.l from ancient version 2.2!
- Fix unusual crash in post-processing for the "any" language.
- Remove three deprecated post-processing functions from API.
- Major reorganization of code base into more modular directories.
- Revive the sqlite3 dictionary into operational form.
- Add double-quotes to splittable punctuation for the "any" language.
- Add API functions to get linkage word positions in the sentence.
- Fix printing of diagrams containing Chinese or other wide glyphs.
- Fix `make distclean` when ant not installed.
Version 5.3.16 (15 April 2017)
- Fix python3 unit tests.
- Restore tty state after ctrl-C, ctrl-Z of the app.
- Check for <stdatomic.hs> before use (in Java bindings).
- Fix handling of punctuation in the "any" language.
Version 5.3.15 (12 Feb 2017)
- Fix Windows compilation; the new wcwidth files were omitted.
- Fix "amy" language for 3-way random morphology splits.
- Remove the "EMPTY_WORD" device; it is no longer needed.
- Enable morphology display in the java bindings, by default.
- Improve random sampling in the "ady" language.
- Improve performance by 10% to 15% for all languages.
- Fix broken multi-threading support in the Java bindings.
- Fix the `make uninstall` targets.
Version 5.3.14 (19 Jan 2017)
- Fix printing widths for Unicode-9 CJK ideographs and emoji's.
- Fix broken randomization in the "any" language.
- Add UTF-8 support to the random morpheme splitter (amy).
- Create an "ady" language for two-part morphology splits.
- Improved error notification facility (experimental).
Version 5.3.13: (19 November 2016)
Emergency fix: remove accidental dependency on zlib and python.
- Fix fatal errors w/ zlib-dev and python dependencies.
Version 5.3.12: (17 November 2016)
Notable: Both python2 and python 3 bindings are built by default.
- Fix bug in 'any' language (opencog/relex/issues/248).
- Preliminary support for common typos in English.
- Enable both python2 & python3 bindings by default.
- Fix locale_t use for the newly introduced Cygwin 2.6.0.
- Include in the distribution the missing make-check.py (for Windows).
- Minisat configuration improvements + fix a problem on Gentoo.
- When using the bundled minisat, link it statically, don't install it.
Version 5.3.11: (28 September 2016)
Notable: A conflict of the bundled version of minisat with the
system-provided version is minimized: LG will now use the system-provided
version, if it is available (and not install the bundled version).
- Re-enable postscript header printing!
- Cleanup python API (in a non-backwards-compat fashion).
- Fix certain adverbial uses of "only".
- Fix some interjective openers to questions.
- Fix serious error with subject-verb inversion to past participle.
- Remove most calls to exit() from the library.
- Update the SAT solver code to use MiniSAT 2.2.
- Use the system minisat2 library if available, instead of the bundled one.
Version 5.3.10: (14 September 2016)
Notable: Fixes a build-break for OSX! Also, a large restructuring
of the English-language dictionaries to handle a greater variety of
sentences with "as" and "so" in them.
- Implement `make installcheck`.
- Pull #371: Simplification of API when handling disconnected words.
- Fix SAT parser crashes.
- Expand default list of Java JDK search paths.
- Fix python bindings: after timeout, no further parsing is performed.
- Fix various adverbial, conjunctive uses of "as", "so".
- Extended list of exclamations.
- Remove CC link, add VC link, for clauses to coordinating conjunctions.
- Fixes for the verb "dare", "someone or other", etc.
- Fix OSX build break, concerning undefined locale_t.
- Pull #385: Fix ancient bug that made dictionary debugging difficult.
Version 5.3.9: (27 August 2016)
Emergency release to fix a fatal error in the previous
release!
- Pull req #354: Major changes to support Cygwin.
- Pull req #356: SAT parser bug fix.
- General python binding cleanup.
- Fatal error: Unable to open default dictionary.
Version 5.3.8: (15 August 2016)
The big change in this release is the support for python2 and
python3 bindings, large improvements in Windows support, and
the use of locales in dictionaries, which should help avoid
locale-related difficulties (for example, capitalization is
locale-dependent; and so mis-set locales break Turkish).
- Disambiguate "Bob" the given name from "bob" the verb.
- Pull req #300: Crash while parsing certain Russian sentences.
- Pull req #301: MSVC compiler error, and warnings.
- Pull req #304: Python failure when no parses are possible.
- Pull req #309: Add MSVC14 support, remove MSVC9, MSVC12.
- Pull req #317: Make Java build reproducible.
- Remove (obsolete) binreloc support.
- Enable both python2.7 and python3.4 bindings to be built.
- Improved Cygwin and MinGW support (as well as improved MSVC support).
- Dictionaries now specify the appropriate locale.
Version 5.3.7: (7 May 2016)
- Fix another MacOS build break, regarding library exports.
Version 5.3.6: (1 May 2016)
- Add missing `parses-quotes-en.txt` file that python tests need.
- Fix build break related to lg_fgetc when libeditline is missing.
Version 5.3.5: (28 April 2016)
Fix strange Apple Mac OSX behavior.
- Modified (hacked) Kazakh.
- MacOS bug fix: fgetc behaves oddly in OSX, see bug #293.
Version 5.3.4: (16 March 2016)
- Fix broken handling of apostrophe (issue #281).
- Revamp the README file; describe transitivity.
- Revised Turkish dictionary from Tatiana Batura, et al.
- Prototype Kazakh dictionary from Tatiana Batura, et al.
- Parse priority tweaks for the OpenCog chatbot.
- Fix Windows printing problem affecting some utf8 codepoints (issue #285).
Version 5.3.3: (23 December 2015)
Fix build break for Apple Mac OSX.
- Improve support for quoted phrases.
- Fixes for assorted zero-infinitive speech acts.
- Add 37 paraphrasing verbs.
- Add Greek mythological names.
- A few dozen more common computing terms added to dictionary.
- Misc coordination and question fixes.
- Misc abbreviations.
- Vietnamese dictionaries!
- Major overhaul of subject-verb inversion.
- Performance improvements on long sentences. (pull #247)
- Change default setting of 'islands_ok' back to false (bug #140).
- Fix for build break on Mac OSX el_capitan w/clang (bug #255).
- Disable perl bindings by default; use Lingua::LinkParser
Version 5.3.2: (4 December 2015)
Fix build break for Apple Mac OSX.
- Performance improvements, esp. for long sentences.
- Use std=c11 (the 2011 C standard) by default.
- Partial Irish English support.
- A few dozen common computing terms added to dictionary.
- Fix for build break on Mac OSX.
Version 5.3.1: (22 November 2015)
Fix build break.
- Fix build break with SAT solver.
Version 5.3.0: (22 November 2015)
This is a major release of the parser, with many important changes in
it. Most fundamentally, the tokenizer has been completely redesigned;
the tokenizer is the device that splits sentences in sequences of words
and (for non-English languages) morphemes.
Another very important change: The python bindings are completely
redesigned, and not in a backwards-compatible way. The new python
bindings are much closer to how the parsing process should be thought
about in the abstract.
There are also various fixes: the SAT solver is no longer crippled.
Assorted performance speedups have been implemented, especially
affecting longer sentences. Assorted bugs and cleanup has been
performed.
- Major redesign of the python bindings.
- Major redesign of sentence tokenization (the "wordgraph" design)
- Verb 'steal' is optionally transitive.
- Fixes for misc MSVC warnings.
- Hebrew dictionary expansion.
- Enhanced diagram printing, giving more space for link names.
- Minor work on phonetic agreement for 'a' vs. 'an'.
- Add ability to histogram the costs of different parses.
- Improve support for splitting sentences.
- Change default setting of 'islands_ok' to true.
- Improve performance on long sentences.
- Fix rare crash due to memory corruption on long sentences.
- Random morphology generation can be enabled at runtime.
- Remove obsolete, unmaintained MacOSX build file.
- Extensive updates to man page.
- Fix crash on long sentences (issue #137).
- Fix a memory leak in language bindings (issue #138).
- Remove bogus post-processor API function.
- Fix broken domain letter printing.
- New regex-file feature - negative regex'es.
- Correct the handling of morphology stems with non-LL links.
- Fix !!LEFT-WALL and !!RIGHT-WALL
- SAT solver now linked statically.
- Assorted SAT solver cleanup and improvements.
- Performance improvement in fast matcher: 15% faster on fixes.batch.
Version 5.2.5 (1 February 2015)
Fixes for assorted breaks.
- Fix contracted "is" verb.
- Fix crash in batch mode (issue #63).
- Fix Python bindings so that setting PYTHONPATH is not required.
- Fix "... as I expected him to be."
- Fix incorrect printing for some Russian linkages. (pull #75)
- Fix warning from glibc version 2.20 (bug #77)
Version 5.2.4 (12 January 2015)
Fixes for assorted breaks.
- Fix usage of 'less'.
- Fix MS Windows random number portability API.
- Fix mis-handled dictionary subscript dot (issue #53).
- Fix crash on dots used as fill (issue #57).
- Fix memory leak (issue #54).
Version 5.2.3 (4 January 2015)
Fixes for assorted build breaks.
- Fix broken check for editline UTF8 support.
- Work around broken perl binding definition for clang.
Version 5.2.2 (3 January 2015)
Fixes for assorted build breaks.
- Fix OSX build break.
- MSVC12 project file fixes.
- Check for UTF8 support in libedit ("undefined reference to el_wgets")
- Enable the 'make check' target for the multi-threading unit test.
- Misc verb fixes.
Version 5.2.1 (28 December 2014)
Failed to run all of the tests when creating 5.2.0. So try again.
- Prototype Indonesian dictionary from Hendy Irawan.
- Fix crash on long sentences.
Version 5.2.0 (27 December 2014)
This is a major release of the parser, with many important changes in
it. The internals of the parser have been re-organized, resulting in
a speedup of 2x to 4x for typical English texts. Multiple multi-
threading bugs were fixed, and there is now a simple multi-threading
unit test. A memory leak was fixed, and a memory over-consumption
bug was fixed. These changes were enabled by the final removal of the
"fat link" code from the parser.
- y'all, ain't, gonna, y'gotta: Beverly Hillbillies basilect.
- Permanent removal of the fat-link code.
- Remove deprecated constituent tree code.
- Windows: add terminal screen resizing support.
- Windows: a build fix.
- reign, rule, run, leave, come: can take predicative adjective.
- Rework costs for many verb-derived adjectives.
- Handle (predicative) adjectival modifiers for assorted perfect verbs.
- Fixes for various color names.
- Fixes for various affirmative answers.
- Add 100 missing verbs.
- Add preliminary lxc-docker (docker.io) support.
- Remove MSVC6 support.
- Fix memleak introduced in version 5.1.0
- Speedup of 1.7x to 4x (depending on text) from linkage processing redesign.
- Fix multi-threading safety bug.
- Fix link-and-domain printing alignment (to handle utf8 char widths).
- Windows: fixes for MSVC12 support.
- Fix memory consumption bug (EMPTY_WORD) introduced in version 4.7.10.
- Get rid of xrealloc, which clashes with libbfd symbol xrealloc.
- Add multi-threaded parsing unit test.
Version 5.1.3 (7 October 2014)
This release continues with fixes for build-breaks for Apple OSX.
- More fixes for build breaks on Apple OSX.
- Minor fixes involving "to do"
Version 5.1.2 (4 October 2014)
The most serious fix in this release is a build-break fix for
Apple OSX Mavericks.
- Fix greeting: "How do you do?"
- Fix indirect object in 'what' questions: 'To what do you owe your success?'
- Fix assorted questions with verb "to be".
- Compile fixes for Apple OSX version "Mavericks"
Version 5.1.1 (23 September 2014)
The most serious fix in this release is a fix involving parse
ranking in the Java API, which was causing RelEx to generate
incorrect parse rankings for certain sentences.
- Minor post-processing cleanup.
- English dict: Fix questions with "it".
- swig: add missing API functions sentence_split, dictionary_get_lang
- Swap order of post-processing and bad morphology rejection.
- Fix handling of ellipsis when there's missing whitespace.
- Java: API bugfix/change: costs should have been doubles not ints.
- Fat-linkage code: fix it so it compiles again.
- Sat-solver: re-enable it so that it runs.
Version 5.1.0 (29 August 2014)
This version includes a number of important changes. One of these is
that the connectors can now be given a direction (head and tail
indicators), so that link-grammar dependencies can now be true,
hierarchical dependency arrows. This is of marginal importance for
English, where dependency directions are implicit, but is vital for
free-word-order languages, where bi-directional links are not enough.
Another important change is that costs can now be arbitrary floating
point numbers. This is particularly useful for providing fine-grained
parse ranking. The LG cost system assigns a "cost" to every connector,
and the sum-total of costs for a sentence determines the parse ranking.
Since costs are additive, they behave as entropies (log P -- the
logarithm of a probability: probabilities are multiplicative, logarithms
are additive).
Under the covers, there's been some major work on the tokenization
(splitting sentences into words) and morphology (splitting words into
morphemes) code. This work is ongoing, and should eventually result in
much better support for non-English languages.
Other notable changes include an updated Russian dictionary, and an
assortment of changes to the English dictionary. An intriguing step
towards phonology: LG can now distinguish between the use of the
determiners "a" and "an" preceding nouns that start with consonants
or vowels. Whether fancier phonology support is possible is a curious
question.
- Updated Russian dictionaries from Sergei Protasov.
- Added morphology-based unknown-word handling for Russian, from Sergei.
- Fix up fat-linkage code, which was recently broken...
- API cleanup: many command-line options never belonged in the API.
- New emoticon support was clobbering certain dictionary words.
- Fix: "Go to spot X", "It happens at time T."
- Add a dozen missing verbs.
- Minor work on greetings.
- Add mechanism for denoting fractional costs in the file-backed dict.
- Fix: broken handling of gerunds (due to bad verb-wall connectors)
- Major redesign of morpheme splitting mechanism (from AmirP)
- Minor extensions to support numeric formulas, e.g. 1 + 1 = 2.
- Remove fat linkage support from the SAT solver.
- Enable build of SAT solver by default.
- Fix multiple bugs with unit stripping.
- Add bounds-checking to the C API.
- Fix the old disjunct-printing implementation.
- Add support for easy-to-use link direction indicator.
- Add random morphology generator tool.
- Partial support for phonetic use of "a" vs. "an" for English.
- Rework how coordination between conjunctions works: "either... or ...", etc.
- Major redesign of tokenization mechanism (from AmirP)
Version 5.0.8 (30 April 2014)
- Fix handling of initial letters in ordered lists.
- Fix another serious error in constituent printing, introduced in 4.8.0.
- New emoticon support was clobbering certain number expressions.
- Misc English dict fixes, more verb-wall connectors.
Version 5.0.7 (29 April 2014)
- Compile fixes in SAT solver.
- Add missing verb-wall connectors for is, hasn't, haven't, hadn't, etc.
- Remove verb-wall connector for imperative verbs.
- Fix serious error in constituent printing, introduced in 5.0.3
- Fix old bug in command-line handling of options.
- Fix parsing of various ordered lists, including some tables of contents.
Version 5.0.6 (18 April 2014)
- Fix: JSON output format missing brace; from Matt Kruse.
- Fix: Serious error in Russian morphology printing.
Version 5.0.5 (17 April 2014)
- Fix packaging bug with the English dictionary.
Version 5.0.4 (16 April 2014)
- Expanded unit tests for capitalization.
- Fix who questions: "Who are they?", "Who are you?", etc.
- Provide verb-wall linkage for many questions.
- Add Biblical naming idioms: "Lud, son of Shem, ..."
- Fix MacOSX build break.
- Fix the 'make clean' target to not remove critical files.
- Fix broken emoticon support in English dict.
- Remove obsolete entity detection tokens from English dict.
- Fix broken equation parsing.
Version 5.0.3 (13 April 2014)
- Minor memory usage optimization
- Fix unit test: suppress printing of empty word, and of morphology.
- Fix: Swig and python were meant to be optional, not required!
Version 5.0.2 (10 April 2014)
- Expanded unit tests
- Fix another sqlite3-dev build break
Version 5.0.1 (9 April 2014)
- Dictionary debugging print fixes from Amir P
- Print summary of parse statistics when in batch mode (from AmirP)
- Generalize the notion of prefix/suffix to arbitrary classes (Amir P)
- Fixes for German adjectives.
- Fix build break when sqlite3-dev not installed.
- Fix regression in Russian morphology handling.
Version 5.0.0 (1 April 2014)
We are proud to announce a major new release of the Link Grammar Parser!
It contains many important changes and new additions. One of the most
significant changes is that the license has been changed from the BSD
license to the LGPL. This was done to enable considerably more
flexibility in accepting contributions to the project: it seems that
few are particularly interested in contributing to a BSD-licensed project.
This change has enabled folding in some new work:
- Arabic and Persian dictionaries! These were previously maintained
as separate add-ons. Including them as part of the distribution
should make it easier for interested users.
- A new 'bindings' directory, containing code for Java, Python, Common
Lisp, OCaML and AutoIt programming languages. The Python bindings
are an updated version of the older pylinkgrammar-0.2.13 bindings.
A SWIG interface file should make it easy to create other language
bindings as well.
- Improved morphology support. This will be invisible to most users,
but it lays the groundwork for add Hebrew support to the parser.
- Expanded Lithuanian support. This remains a simplistic prototype, but
it now performs a more sophisticated morphological analysis.
- Experimental Turkish and Hebrew dictionaries.
- A demo of the JSON parser server: it shows how to run the server,
which will accept accept raw sentences on a socket, and returns the
parsed forms.
- Some slightly incompatible changes to the API: it was time for some
housekeeping.
- Misc minor updates to the English Language dictionaries.
- Preliminary work for SQL-backed dynamic dictionaries. This should
enable certain types of automated language learning.
The full ChangeLog is shown below.
- License upgrade to LGPLv2.1
- Arabic dictionaries, from Jon Dehdari
- Persian dictionaries, from Jon Dehdari
- Support for Hebrew tokenization, from Amir P.
- Fix wild-card matching for user-supplied word lookup.
- Prototype Turkish dictionary from Can Bruce.
- Re-arrange programming language bindings directory.
- Adopt the orphaned/unsupported pylinkgrammar Python bindings.
- Deprecate the obsolete CNode interface.
- Provide low-level perl bindings.
- Adopt the orphaned/unsupported OCaML bindings.
- Support affirmative replies: "Who did it?" "John's evil twin."
- Expanded Lithuanian dictionary.
- Minor disjunct printing fixes.
- Fix: "Mary is too XXX to talk to."
- Prototype Hebrew dictionary from Amir P.
- Change !suffixes flag to !morphology.
- Introduce a bi-directional connector, for free-word-order languages.
- Introduce a symmetric-AND operator, for free-word-order languages.
- Add demo shell script for running the JSON parse server.
- Bugfix: Java server failing when input sentence has commas in it!
- New !test and !debug commands for selective debugging support.
- Print post-processing rejection message, when !bad is enabled.
- Remove some deprecated functions for C API.
- Remove all deprecated functions from Java API.
- Initial support for an SQL-backed dynamic dictionary.
Version 4.8.6 (2 February 2014)
- Fix minor OSX compiler warnings.
- Check for presence of Java ant before assuming it is there.
- Fix crash on certain sentences containing equals sign.
- Fix parsing of lists (blah, blah and blah).
- Fix build break for uClibc systems (Gentoo).
- Allow ungrammatical usage of 'ages' instead of 'aged'.
- Fix crash on certain sentences containing words with periods.
Version 4.8.5 (5 January 2014)
- Update memory usage accounting; fix accounting bugs.
- Fix Java garbage collection bug.
- Fix numerous compiler warnings in the SAT-solver code.
- Fix build-break involving multiple declaration of 'Boolean'.
Version 4.8.4 (31 December 2013)
- Fix build break for Mac OSX.
Version 4.8.3 (30 December 2013)
- Create new msvc12 build files, restore old msvc9 files.
- Revert location of the Windows mbrtowc declaration.
- Add verb-wall connector for present participles.
- Fix build-time include file directory paths.
- Provide the 'any' language to enumerate all possible linkages.
- Fix recognition of U+00A0, c2 a0, NO-BREAK SPACE as whitespace.
- Improve parse-time performance of exceptionally long sentences.
- Fix crash on certain sentences containing equals sign.
Version 4.8.2 (25 November 2013)
Add missing file, needed for Java bindings.
- More MSWindows UTF-8/multi-byte fixes (for Russian).
- Add missing JSONUtils file.
Version 4.8.1 (21 November 2013)
Minor updates, unless you are using Java, or the Russian dictionaries
on Windows, in which case, you'll need this update.
- Ongoing work on Viterbi.
- Updated MSVC9 project files from Jand Hashemi (Lucky--)
- Fix important bug in Java services: return top parses, not random ones.
- Java: for the link-diagram string, do not limit to 80 char term width.
- Windows: UTF-8 fixes so that Russian works in most MSWindows locales.
Version 4.8.0 (24 October 2013)
This is the start of a new version series, containing an important
change to the English language dictionary. Three new link types are
introduced WV, CV and IV. These are used to connect the left-wall to
the primary verb of the sentence (WV), to connect the ruling clause
to the primary verb of a dependent clause (CV), and a similar link
for certain infinitive verbs (IV). The goal of these links is to
make it easier to locate verbs, and thus to provide a more direct
mapping from the link-grammar formalism to a dependency parse (as
dependency parses always put the verb at the root of a sentence).
These are not the first links that explicitly indicate root verbs:
several other link types already play this role: The AF, CP, Eq, COq
and B links already play this role. The new WV, CV and IV links
round out this capability and do so in a very general form. See
WV,
CV and
IV
for details.
With this release, we expect that all (non-auxiliary) verbs in a
sentence will be linked either to the wall, or to a controlling parent.
We also expect there to be some additional fixes and tightening-up
to occur in future releases, especially in regards to comparative
sentences.
This release also includes a variety of fixes to the Java API/server.
In addition, some ancient, deprecated C code was removed.
- Fix "he answered yes"
- Support bulleted, numbered lists.
- New link types from Lian Ruiting, for identifying the head-verb.
- Java: fix bug when totaling WordNet word-sense score.
- Java: add info to README about using the JSON parse server.
- Java: remove many deprecated functions.
- C API: remove some deprecated functions.
- Java: fix silent failure when library is not found.
- Java: Add support for fetching the ASCII-art diagram string.
- Java: Fix insane language selection initialization.
- Fix: "The pig runs SLOWER than the cat."
- Fix: conjoined superlatives: "... the longest and the farthest."
- Fix: "inside" can be used with conjunction: "near or inside..."
- Fix: conjoined question modifiers: "exactly when and precisely where..."
- Fix: issue 59: crash/corruption when dictionary opened twice.
- Fix: assorted exclamations!
Version 4.7.14 (20 June 2013)
MacOS build break
- Fix build break on MacOSX
Version 4.7.13 (17 June 2013)
Windows unicode handling fixes.
- Remove trailing carriage return in MSVC6 filenames.
- Performance improvement: avoid excess system timer calls.
- Unicode fixes for Windows.
- Fix: "... a purseful of pesos"; other currency fixes.
- MinGW, unicode fixes from Per Larsson
Version 4.7.12 (25 May 2013)
Major revision to the Russian dictionaries!
- Large fixes to the Russian dictionaries.
- Windows: Explicitly fail if Cygwin version is too old.
- Tweak the lt dict to work again with the modern parser.
- Make the fat linkages code be compile-time configurable.
- Disable fat linkages by default; mark as deprecated.
- Fix SAT-solver build; recent changes had broken it.
- Export read-dict.h as a public API.
- Ongoing development of the Viterbi prototype.
- Windows: some UTF8/widechar refactoring.
- Java bindings: add method to set the language.
- CMake: add version checking to the CMakefile
- Fix: failed handling of capitalized first word for Russian.
- Fix: stemming failures in many cases (for Russian dictionaries)
- Add flag to suppress stem-suffix printing.
- Windows: Fixes to MSVC6 build files.
- Fix: hash-table bug affecting Russian dictionaries
Version 4.7.11 (23 March 2013)
Emergency build break fix.
- Fix build bug introduced recently (build_disjuncts_for_dict_node)
- MSVC6 build environment fixes.
- Ongoing development of the Viterbi prototype.
Version 4.7.10 (9 March 2013)
Russian dictionaries!
- New: Russian dictionaries from Sergey Protasov!
- Improve prefix, suffix handling and stemming (needed for Russian)
- Fix: add "gurgle.n" to dictionary.
- Improve memory utilization (remove pointless malloc/free).
- Fix: UTF8 multi-byte diagram printing.
- Add UTF8 multi-byte editline support, when available.
- Include Common Lisp bindings, from Peter Szolovits (circa 2009).
- Ongoing work on pre-alpha Viterbi decoder.
- Fix: Bug 55: aspell header file included even if disabled.
- Updated MSVC6, MSVC9 project files.
- New, improved search for java jni.h
- Faster counting of disjuncts for the !! command.
- The !! command now shows regex entries and stem+suffix splits.
Version 4.7.9 (3 November 2012)
A large number of dictionary updates, plus some more MacOSX touchups.
- Fix: Poor comma choices: "The man, that you saw..." (issue #36)
- Fix: Add more java jni.h search paths.
- Fix: Warning in java JSON interface.
- Fix: weren't constructions: "If it weren't for Sally, ..."
- Fix: Better support for ellipsis ...
- Fix: strangely: "He is behaving very strangely"
- Fix: assorted usages of yes, no, maybe
- Fix: "being" in subordinate clauses: "The blade being dull, he ..."
- Fix: Build failure for Mac OSX.
- Fix: "... this time for real".
- Fix: "... the impudence to laugh."
- Fix: "we two", "us two"
- Fix: very old parse bug, using "enough"
- Fix: assorted profanity as synonyms to "else"
- Fix: whoever, whomever
- Fix: decade possessives: "It's a 60's hit"
- Fix: "... chances that ... could be ..."
- Fix: "Do you have it ready?"
- Fix: "What John wants are those cats"
- Fix: greetings
- Fix: misc verbs: wave, ink, kiss, bake, corral, crown, decant, rope,
slice, ink, butter, proclaim, make, give, bid, bade, think
- Fix: titles can behave as indefinite nouns.
- Fix: somewhere, near, nearby, halfway, about
- Fix: police-interrogation style questions
Version 4.7.8 (10 October 2012)
Broke the build, in the previous release. Ooooops!
- Ongoing work on pre-alpha Viterbi decoder.
- Fix include file paths so that 'make install' doesn't cause recompile.
- Fix build break due to bad include file paths in Makefiles.
Version 4.7.7 (6 October 2012)
This version makes a number of significant changes to the dictionary.
The most important of these is to indicate that expressions of the form
...a number of X act as if they were determiners;
correspondingly, the subject and object links move to a new location.
That is, the subject/object links now land on "X" instead of "number".
Other changes to the dictionary are listed below.
- Fix configure to not bomb if c++ is not found. (weird regex.h missing)
- Fix configure to explicitly require c++ only if sat-solver enabled.
- Dictionary: parse comma in "Monsters, Inc."
- Update README about Viterbi algo.
- Remove malloc-dbg.c from the MSVC6 project file.
- Add missing source files to MSVC6 project file.
- Use enums for constituent display style.
- Move command-line parser to its own directory.
- Portability fixes for Windows, from Bill Hayes.
- Fix: "Show results from today"
- Fix: conjoined imperatives.
- Fix: qualified conditional clauses
- Fix: conjoined irregular verbs: seem, appear, do, be, prove, have
- Disable fat link code via conditional compile.
- Fix: "a lot of", "a gaggle of" now treated as quantifying determiners
- Fix: likewise: "gallons of", "grams of" volume measures as determiners
- Update German dictionary so that it actually loads.
- Fix: assorted "but not", "not very" constructions.
- Pre-alpha implementation of Viterbi decoder started.
- Fix: conjunctions of optionally ditransitive verbs.
Version 4.7.6 (26 April 2012)
This version fixes a bungled header file in the last version: builds
broke because the public header file was including an internal header file.
- fast-match.c: unroll recursive call into loop (thx Valery Kholodkov).
- Fix accidental inclusion of internal header file from public header file.
Version 4.7.5 (16 April 2012)
This version includes fixes for a handful of dictionary bugs, an
all new manual page, and several fixes and improvements for the
build system.
- Remove several duplicate given names.
- Remove malloc-dbg.c from the MSVC9 project file.
- Remove several dozen duplicated idioms.
- New AM_SILENT_RULES for less noisy build!
- Clean up misc minor compiler warnings.
- Start using -O3 optimization for another 2% perf improvement.
- Fix configure.in --enable/--disable flags (bug #53)
- Some intransitive verbs are actually transitive in rare cases.
- Use enum not int for the cost-model type.
- Update the man page to reflect actual command usage.
Version 4.7.4 (13 February 2011)
This version includes fixes for several bugs/crashes introduced in
the previous version :-( It also includes changes that result in
dramatic speed improvements on long sentences, with an average of
1.1x to 2.0x improvement on "typical" texts.
- Dramatic speed improvement for long sentences (hash table fix).
- Fix: crash on certain sentences, when fat-link parsing enabled.
- Fix: fatal error introduced in last version (combinatorial explosion).
- Fix: mem leak introduced in last version (combinatorial explosion).
- Assorted fixes to the Boolean SAT solver.
Version 4.7.3 (7 February 2011)
This version fixes a bug where a "combinatorial explosion"
resulted in no parses being printed.
- Fix: "level" as indefinite noun: "... at knee level".
- Enable max disjunct cost as a controllable parser parameter.
- Don't just give up on combinatorial explosion, show something!
Version 4.7.2 (6 January 2011)
This version fixes a crash and an "almost infinite loop"; both might
be hit when using the (now deprecated) fat-links, and certain overflow
conditions are met.
- Fix: crash on certain sentences, when fat-link parsing enabled
- Fix: 3 corrupted nouns in words.n.1
- Fix: minimize cpu spin if constituent processing and-list overflows.
Version 4.7.1 (11 December 2010)
This version fixes several bugs introduced by the conjunction-handling
rework of 4.7.0. Several Windows-related compile fixes are included as
well.
- Fix: dictionary: Add "x" as synonym for "times".
- Fix: compilation failure due to lack of termios support on Windows.
- Fix: if not set, then force-set locale to UTF-8 in the java library.
- Fix: Explicitly include <stddef.h> for Windows builds.
- Fix: "John imagines himself lost".
- Fix: predicative adjectives w/misc verbs: "You are driving me crazy"
- Fix: "judge innocent", etc.
- Fix: many, many conjoined present tense verbs.
- Fix: conjoined verbs with negations.
- Fix: "to" as post-nominal modifier: "the inability to laugh"
Version 4.7.0 (12 September 2010)
This version introduces a major change to the way that conjunctions are
handled. The parse output for conjunctions is not compatible with the
old output, but should be easier to work with, solving many problems.
The rework results in a performance improvement of 1.3x to 2.7x, depending
on the text.
- Fix: Hunspell configuration on Fedora (bugtracker issue 47)
- Fix: 'turn' with adjective: "She turned him green" from wingedtachikoma
- Fix: comma-conjoined modifiers: "It tastes bitter, not sweet."
- Fix: conjoined question words: "When and where is the party?"
- Fix: recognize short, capitalized words (Los, La, etc.).
- Treat colon as synonym for is: "The answer: yes."
- Fix: begin with prepositions: "It all began in Chicago."
- Fix: "What does it come to?" and related.
- Fix: null infinitive: "I'd like to, I want to."
- Fix: "Because I said so."
- Fix: "sure" as preverbal adverb: "It sure is."
- Fix: Gerunds with determiners: "a running of the bulls"
- SJ link for conjoined nouns/noun phrases.
- Sort linkages according to whether fat linkage was used.
- Add flag to enable use of fat linkage during parsing.
(Fat links now disabled by default).
- Add male/female gender tags to misc nouns.
- Fix: misc optionally transitive verbs: mix, paint, boot
- Fix: word order: "look about fearfully", "look fearfully about", around
- Fix: recognize simple fractions
- Fix: "is" with uncountable nouns: "there is blood on your hands"
- Fix: Roman numeral suffixes e.g. "Henry VIII"
- Fix: regression in dates followed by punctuation. "In the 1950s, ..."
- Fix: verbs drank, drunk are optionally transitive.
- Fix: regression: "all the X", X can be plural or mass.
- Fix: verbs paint, color may be ditransitive: "paint the car bright green"
Version 4.6.7 (16 April 2010)
- Fix: configure handling of --enable-hunspell (bugtracker issue 46)
- Fix: conjoined modifier phrases: "the black and white cat sleeps."
- Fix: MSWindows+MingGW compilation issue.
- Update msvc9 Microsoft Visual C 2008 project, from Boris Iordanov.
Version 4.6.6 (19 March 2010)
- Minor fixes to handling of units.
- Fix: MSWindows+Cygwin/MingGW compilation issues.
- Use pkgconfig for editline (from Vincent Untz)
- Fix: "I want it back."
- Additional titular prefixes, suffixes: Maj. Ph.D.
- Add code to detect dictionary version number.
- Fix: MS VC missing strdup() function.
- Add AutoIt3 bindings from JRowe
Version 4.6.5 (3 November 2009)
- Fix: Superlatives without preceding determiners ("... likes you best")
- Fix: Take more care in distinguishing mass and count nouns.
- Fix: Old bug w/relative clauses: Rw+ is optional, not mandatory.
- Provide tags identifying relative, superlative adjectives.
- Remove BioLG NUMBER-AND-UNIT handling, its been superseded.
- Fix handling of parenthetical phrases/clauses.
- Fix: handling of capitalized first words ending in letter "s".
- Fix: support "filler-it" SF link for "It was reasoned that..."
- Fix: certain WH-word constructions: "I did not know why until recently"
- Fix: go: "there goes the greatest guy ever"
- Fix: opening coordinating conjunctions: "And you can also ..."
- Configurable Hunspell spell-checker dictionary location.
- Fix: Misc ordinal usage.
- Add support for aspell spell-checker.
Version 4.6.4 (11 October 2009)
- Restore nouns starting w/letters x-z, elided in version 4.5.9 ff.
- Add support for single-word interjections/exclamations!
- Fix: sometimes command line client fails to show all valid linkages.
- Misc fixes: such_that, upon, acted.v
- Fix: impersonal "be" linking to passive participle.
- Fix: handling of capitalized first words.
- Fix: duplication of certain parses involving transitive verbs.
Version 4.6.3 (4 October 2009)
- Fix compilation bug on FreeBSD.
- Fix: allow MX link to post-nominal ", to be ..., "
- Fix: add idiom "time and again"
- Fix: another BioLG regression in handling of possessives.
- Fix: handling of period at end of number at end of sentence.
- Fix: Capitalized words ending in s at start of sentence.
- Use corpus-statistics-based ranking by default, if available.
- Fix difficulties in build of corpus statistics module.
Version 4.6.2 (21 September 2009)
- Fix: "come across as authoritative".
- Improve Java location guessing in FreeBSD
- Fix for assert triggered by long sentences.
- Fix: long sequence of periods treated as unknown word.
- Add informational print showing dictionary location on startup.
- Remove duplicated {@MV+} in tend.v
- Automatically resize the display size to fit the current window size.
- Fix handling of punctuation at the end of a capitalized word.
- Fix misc verbs acting as adjectival modifiers: e.g. "given", "allied"
- Fix bug in BioLG code regarding the handling of possessives.
- Fix a (rare) crash in sentences with many conjunctions.
- Fix a crash involving long sequences of UTF8 punctuation marks.
Version 4.6.1 (31 August 2009)
- Stop printing annoying warning when !vars are used.
- Fix missing dict file units.2 problem
- Fix compilation bug on FreeBSD.
Version 4.6.0 (29 August 2009)
- Avoid used of bzero, add missing include directives (MacOSX problem)
- Reclassify a number of "medical" prepositions as adverbs.
- Add approx 100 adverbs & 300 adjectives.
- Add approx 250 verbs.
- Add approx 300 nouns.
- Add misc units.
- Add misc European connector words/patronymics.
- Reclassify 100's of transitive verbs as optionally-transitive.
- Add distinct tokenization step ("sentence_split") to public API.
This last change forces the minor-version-number bump.
Version 4.5.10 (25 August 2009)
Version 4.5.9 (25 August 2009)
- Modify error messages to indicate that they are from link-grammar.
- Add missing Java files that were forgotten last time around.
- Add greeting to command-line client startup.
- Print disjunct cost also, when requesting disjunct printing.
- Add missing color names as mass nouns.
- Fix: Reclassify musical instruments: "He plays piano"
- Add experimental word-clustering system.
- Add CMake build file
- Fix: "It takes longer than that."
- Fix: "He has done very well."
- Fix: a dozen optionally transitive verbs (swim, kill, etc.)
- Fix: "He's out running."
- Fix: "suddenly" is a "manner adverb", not a clausal adverb.
- Fix: Use Pg links to gerunds: "He feared hitting the wall."
- Fix: assorted numerical-range bugs.
- Fix: prep modifiers with distances: "It is a few miles out"
- Fix: Spelled-out dates: "It started in nineteen twelve"
- Fix: Misc date, time expression parsing e.g "Zero hour is here."
- Fix: Misc words, "ordered list", "screened out"
- Fix: Post-fixed numbers can act as determiners.
- Fix: "We bought the last 50 ft. of cable."
- Fix: opening directives to imperatives: "Finally, move it back."
- Fix: Improved simple equation parsing support.
- Fix: Add misc fixes from BioLG that were previously overlooked.
- Fix: "favorite" can take determiner "a" ("a favorite place")
- Fix: assorted clausal complements: "The emperor ordered it done."
- Fix: ordinals: "First on our list is ..."
- Fix: verb modifier "some of the time", "most places"
- Fix: Sit, stand take modifiers: "he stood still"
Version 4.5.8 (2 July 2009) includes the following changes:
- Fix: 'than anticipated', 'than was anticipated', etc.
- Fix: 'saw the wood'
- Fix: sometimes commas are used as if they were semicolons.
- Fix: 'We have quite enough work already, thank you!'
- Fix: allow 'and' as conjunction in entity names.
- Fix: 'I stared him down', 'They shouted him down', 'booed off'
- Fix: 'sound him out', 'look him over'
- Fix: 'Somewhere in the distance'
- Stub out list of names given to both men and women, to avoid duplication.
- Fix: 'I think so, too'
- Fixes for compilation under Cygwin.
- From Boris Iordanov: fixes to JSON java code.
- From Boris Iordanov: new java remote client code.
- Fix: Biological texts commonly have adj-noun-adj-noun chains
Version 4.5.7 (4 June 2009) includes the following changes:
- Fix 'make install' for windows (abi bug 12049)
- Fix multi-threaded bug when TRACK_SPACE_USAGE is defined.
- Add './configure --enable-mudflap' just for fun...
- Fix: "Walk tall", "Think quick"
- Fix: "... part no. 1234-56A"
- Fix regression from BioLG merge: "It cost $14 million."
- Fix come/came: "The dog came running..."
- Fix year abbreviations: "He drove a souped-up '57 Chevy"
- Fix sit, stand: "The dog stood still"
- Fix act up, act out: "He is acting out." "The motor is acting up."
- Fix notoriously, poorly: "The store was poorly stocked".
- Fix: "strong" can be adverb
- Add support for recognizing basic time zones during parsing.
- Fix: verbs acting as adjectival modifiers: "a very politicized deal."
- Fix: ".. nearly so well", "...almost so well".
- Fix financial ranges: "It will cost $10 million to $20 million to build."
- Expand handling of capitalized words that appear in entity names.
- Expand the list of characters that are recognized as quotes.
- Support usage of yes, no as sentence openers.
- Better support for directives, commands.
- Fix: "Ash Wednesday", "Fat Tuesday", etc.
- Fix: post-verbal adj: "she wiped the table dry"
- Fix: wish: "she wished me a happy birthday"
Version 4.5.6 (24 May 2009) includes the following changes:
- Bugfix: fix non-thread-safe usage.
- Changes to enable MinGW/Windows to compile.
- Update of MSVC6 build files
- Fix: pizza, fries, chopsticks.
- Export word-sense database to Java apps.
- Fix: "Was the man drunk or crazy or both?"
Version 4.5.5 (10 May 2009) includes the following changes:
- Bugfix: crash for zero-length sentences.
Version 4.5.4 (9 May 2009) includes the following changes:
- Fix: "sleep in": "A bed is something you sleep in."
- Fix: "drinking": "Let's go drinking."
- Fix go+bare infinitive: "Let's go shop", "Let's go swim"
- Fix: "Let's go for a swim." "Let's go for a smoke".
- Fix: "Let's not" "Let's not go" "Let's not cry"
- Fix: ... is : "All he ever does is complain."
- Fix: "You will die young/happy/unhappy"
- Fix: "You should exercise to stay fit."
- Fix: "We danced 'til dawn."
- Fix: "tell off": "She had told him off."
- Bugfix: sometimes spell checker would run even if turned off.
Version 4.5.3 (14 April 2009) includes the following changes:
- Haste makes waste! Revert a recent 'fix'.
Version 4.5.2 (14 April 2009) includes the following changes:
- Use re-entrant version of mbtowc in all code.
- Fix run-time breakage on Mac OSX and FreeBSD.
Version 4.5.1 (13 April 2009) includes the following changes:
- Fix Assertion failed: negative constituent length!
- Fix build break for Mac OSX.
- Force use of UTF-8 locale in the command-line tool.
Version 4.5.0 (10 April 2009) includes the following changes:
- Hack around missing SQLite3 pkgconfig on MacOS
- Fix adverbs: 'The motor ran hot', 'the door swung wide open', etc.
- Fix: 'at risk of breakdown', 'under threat of fire'
- Add regular-expression-based word guessing, from BioLG project.
This provides support for many scientific/biomedical terms.
- Add spell-guessing for unknown words.
- Fix UTF8 support to be correctly thread-safe.
- BioLG: fix post-numbering: 'it started on day one'
- BioLG: add number ranges: 'it takes 2 to 3 times the effort'
- BioLG: assorted adverb fixes, typical of scientific prose.
- BioLG: initiate, attach, localize etc are optionally transitive.
- BioLG: allow fork, branch, splice, export, etc to take particles.
- BioLG: extended use of Greek letters in biomedical text.
- BioLG: support parsing of Roman numerals.
- BioLG: support Greek-letter-number combinations.
- Fix: 'she was singing', etc.
- Enable WordNet word-sense identification based on syntactical usage.
Version 4.4.3 (11 March 2009) includes the following changes:
- Look at JAVA_HOME to guess location of jni.h
- Add dictionary maintenance utility script.
- Add prototype German dictionary.
- Fix sentences containing measurements, lengths, speeds. etc.
- Add 1400 new adjectives, many of scientific origin.
- Add 250 international currencies, including special utf8 symbols.
- Performance optimization in sentence tokenization.
- Add 1100 male, female given names (from Bruce Wilcox).
- Cleanup parsing of "in-" date expressions, e.g. "in the 1940's".
- Cleanup of assorted compiler warnings.
- Fix handling of special utf8 characters in dictionary.
- Fix hang when parsing certain utf8 input strings.
- Make sure that most strings in the API are declared const char *
- Add option to command-line client to print disjunct strings.
Version 4.4.2 (15 January 2009) includes the following changes:
- Fix: 'The keys are gone', 'It is a gone desire'
- Fix: 'He poured me a glass', 'He wired me the money', etc.
- Initial rough-in of corpus statistics based parse ranking.
- Initial rough-in of corpus-guided word-sense identification.
- Fix for googlecode bugs 38, 39 (Java, Ruby failure to load)
- Add new API routine to return disjunct string.
Version 4.4.1 (15 December 2008) includes the following changes:
- Balance the dictionary tree; this speeds word-lookup slightly.
- New MSVC6 build files from Evgenii Philippov.
- Fix java server classes to pass along the link-grammar version number.
Version 4.4.0 (7 December 2008) includes the following changes:
- fix: recognize curly-single-quote ’ where straight quote can be used.
- recognize and explicitly ignore emoticon types.
- Include MSVC6 build files.
- Apply patch needed for Ruby bindings.
- fix: "Where did they come from?", per Viswanath IIIT
- fix: "Where did they go to?"
- fix: "It gives me peace of mind."
- fix: many, many incorrectly identified mass nouns.
- fix: ladle.v "molten hot" "piping hot"
- fix: "It's a shame that...", "The crux of the plan is that..."
- Performance improvements (about 11%) to prunce.c from Bruce Wilcox
- fix: "He eats with me nightly."
- Add new public api function: linkgrammar_get_version()
- MSVC9 build files from Borislav Iordanov
- Java network-efficient client-server classes from Borislav Iordanov
Version 4.3.9 (8 October 2008) includes the following changes:
- Issue 13: "John is altogether amazingly quick."
- Nonstandard spelling "unequivocably"
- Dictionary fixes for 'marginally', etc. "That one is marginally better"
- Issue 7: Dictionary fixes for 'done': "I am done working"
- dictionary entries for walk-up drivethru car-wash
- dictionary: "I am through being mad", "It was a through flight", etc.
- Issue 11: "You are doing well"
- Issue 3: "I asked Jim a question", "I told Jim a story"
- Passive subjects with objects: "I was told that crap, too" "...was asked..."
- Fixes for Apple Mac OSX (crash on non-executable stack)
- Early version of Filip Maric's boolean SAT solver
- fix: "He talked quietly of revolt."
- fix: "It consists mostly of sand.", "He talks, mostly of revolution."
- fix: "He talked mostly to Ann.", "He talks a lot."
- fix: than_usual: "He is taking longer than usual."
- fix: a batch of new verbs from Roman Khlupin
- fix: Fix crash on Apple Macintosh by correctly identifying the platform.
- fix: "San Gabriel" "Block Island" "Great Southern Bank" "de la Rente"
- fix: "I biked Johnson Creek."
Version 4.3.8 (14 August 2008) includes the following changes:
- Build fixes for Windows (missing bzero and thread-safe rand)
- Fix several mem leaks, one in the Java jni code.
- Fix crash of java6 jvm, due to blown stack limits.
Version 4.3.7 (8 August 2008) includes the following changes:
- Expanded list of given names to include those from 2005 US Census
- Export a new java jni function, to get the inflected word.
- Add a TO+ connector so that "I tend to agree" parses correctly.
- Makefile fixes for OS/X Leopard
Version 4.3.6 (21 July 2008) includes the following changes:
- Fixes for Windows MS Visual-C builds.
- Fix parsing of "He walked the dog.", "He sailed the boat."
- Add support for right-apostrophe (’) which is a non-ASCII UTF8 char.
- Add support for other non-ASCII UTF8 punctuation.
- Fix crash on printing constituent tree of certain long sentences.
- Avoid recursive error reporting for UTF8 dictionary errors.
- Clarify error logging and error printing.
- Add java getVersion() to return link-grammar version string.
- Add more numbers to dict (e.g. twenty-seven, bazillion, half-dozen, etc.)
- Foodstuffs: bagels, lox, tacos, guacamole, roe, neufchatel, mayo, etc.
- Weights and measures: megabytes, °C, km² etc.
- Performance improvements in printing of link-tree.
- Convert assert into warning when no canonical linkages can be found.
- Convert assert into warning when constituent andlist overflows.
- Provide additional checks for constituent overflows.
- Convert most error printfs into a formal error reporting system.
- Remove all globals, library is now thread-safe.
- Fix crash when sentence has square bracket, and doing constituents.
Version 4.3.5 (29 April 2008) includes the following changes:
- Added ant build file to create the link-grammar jar file.
- Fix regression in command-line client of multiple-parse display.
- Use MB_LEN_MAX, not MB_CUR_MAX for UTF8 support.
- Fix a WIN32 compiler regression (no in-line support in Windows).
- Fix error in handling of UTF8 dictionaries.
- Fix strncat() misuse in error.c
- Fix capitalization errors in country names.
- Fix parsing of "he angled left, he dodged left, he turned left".
- Don't build the JNI library if Java isn't found. Fixes build on Windows.
- Fix install bug for NetBSD systems.
- Pre-detected entities cannot participate in G links.
- There is no UTF8 support in windows, so stub it out.
- Fix crash in constituent output, bug #22 in googlecode bugtracker.
- Some small steps taken to eventually make library thread-safe.
- There are three constituent string styles, enable all three.
- Make the command-line flag errors less cryptic.
- Add readline (BSD editline) support.
- Rename "grammar-parse" to the more logical "link-parser".
- Small man page updates.
- Export and cost, link cost via public API.
Version 4.3.4 (16 March 2008) includes the following changes:
- Fix regression of handling of capitalization at the start of sentences.
- Fix dictionary search path so that it respects command-line input.
- Fix rare but nasty crash when parsing long sentences in panic mode.
- Add a method to set the dictionary path.
- Fix all remaining compiler warnings.
- Make parser capable of handling UTF8 strings and dictionaries.
- Ongoing minor expansion of the Lithuanian (lt) dictionary.
Version 4.3.3 (27 February 2008) includes the following changes:
- Missing java is a warning, not an error.
- man page for grammar-parse.
- Removed cruft from the dictionary open routines.
- configure tries to guess some non-standard jni.h locations.
- Split up java library exports, should help Cygwin builds.
- Fix java library pre-linking bug.
- Minor English dictionary additions.
- Prototype Lithuanian (lt) dictionary.
Version 4.3.2 (2 February 2008) includes the following changes:
- Fix dictionary errors involving given names; e.g. any sentence
with the name "John" in it.
- Minor Windows build fixes.
Version 4.3.1 (31 January 2008) includes the following changes:
- Merger of extensive dictionary additions from Peter Szolovits.
This adds 15K new words, bringing the dictionary to 70K words
total.
Version 4.3.0 (30 January 2008) includes the following changes:
- New link types (Ct, Cta, Rn, Rw) for comparatives, so as to link
relative clauses: "John is bigger than Dave is", "John wants more
cookies than Dave wants". The Rw link is used to link question words
to the relative clauses that follow them.
- Dictionary Fixes for "Espresso is a coffee drink", "Teach me fetch",
"I am pooped" as synonym for "I am tired", "Mother likes her",
"Mommy loves me" and related. Also, directives involving "go":
"Go play ball", "Go take a walk", "You and Rover go play with the
ball."
- Dictionary support for external entity markup. This includes the
recognition of personID0..personID60, dateID0..dateID60,
organizationID0..organizationID60 and locationID0..locationID60
as appropriate words.
- Fixes of numerous compile-time warnings.
- Simple Java (JNI) bindings.
Version 4.2.5 (11 November 2007) includes the following changes:
- Fix for a security problem, involving a buffer overflow:
CVE-2007-5395.
Version 4.1 (30 August 2000)
Version 4.0 (4 August 2000) changed significantly from version 3.0.
These changes are described in the
improvements page.
Version 3.0 (15 April 1998)
Version 2.0 (17 October 1995)