kr.angelov
|
2893397fbb
|
bugfix in the statistical parser
|
2013-03-11 14:47:43 +00:00 |
|
kr.angelov
|
5a54596fe8
|
the parser in the C runtime should not crash if the start category is not defined
|
2013-02-19 12:08:48 +00:00 |
|
kr.angelov
|
f86dcb6572
|
bugfix in the grammar reader in the C runtime
|
2013-02-19 12:04:10 +00:00 |
|
kr.angelov
|
ffb17bd26a
|
bugfix in the linearizer for the C runtime
|
2013-02-13 15:39:01 +00:00 |
|
kr.angelov
|
55203110bb
|
now the beam size for the statistical parser can be configured by using the flag beam_size in the top-level concrete module
|
2013-02-12 10:53:13 +00:00 |
|
kr.angelov
|
1f77afcfce
|
the statistical parser now uses a baseline lexical estimation of the beam size
|
2013-02-12 09:41:32 +00:00 |
|
kr.angelov
|
a6b35a9053
|
the class PgfConcr from the Python binding now has a property name which returns the name of the concrete syntax
|
2013-02-11 15:51:26 +00:00 |
|
kr.angelov
|
0b7b939aca
|
refactoring: now all named objects in the C runtime have an explicit name field
|
2013-02-11 14:10:54 +00:00 |
|
kr.angelov
|
56c8f91d19
|
remove the pgf2yaml tool which was both broken and redundant. The declarations for generic programming from data.c are removed as well
|
2013-02-11 13:51:12 +00:00 |
|
kr.angelov
|
ff25ba8f90
|
the grammar reader in the C runtime is completely rewritten and it doesn't use the generic programming API
|
2013-02-11 10:16:58 +00:00 |
|
kr.angelov
|
e9b5557c6c
|
This patch removes Gregoire's parse_tokens function in the python binding and adds another implementation which builds on the existing API for lexers in the C runtime. Now it is possible to write incremental Lexers in Python
|
2013-02-01 09:29:43 +00:00 |
|
gregoire.detrez
|
459a9a855b
|
Python binding: add a parsing function that accepts directly a list of tokens.
Is allows to define a tokenizer in python (or use an existing one, from nltk for instance.)
|
2013-01-24 13:31:34 +00:00 |
|
kr.angelov
|
3be31c62e9
|
a new reasoner in the C runtime. It supports tabling which makes it decideable for propositional logic. dependent types and high-order types are not supported yet. The generation is still in decreasing probability order
|
2013-01-07 12:50:32 +00:00 |
|
kr.angelov
|
bb077b8330
|
bugfix: the linearizer should not generate extra space at the end of the sentence
|
2012-12-19 11:18:34 +00:00 |
|
kr.angelov
|
f7eaa8a89a
|
bugfix for linearization of metavariables at the root of a tree
|
2012-12-19 10:03:05 +00:00 |
|
kr.angelov
|
6201640d7b
|
rename linearize.{h/c} to linearizer.{h/c} which follows the convention used in parser.c and reasoner.c
|
2012-12-19 09:17:24 +00:00 |
|
kr.angelov
|
5c9ee467a9
|
a major reimplementation of the linearizer in the C runtime
|
2012-12-19 09:07:05 +00:00 |
|
kr.angelov
|
dc809da91f
|
the C runtime now can read abstract expressions with literals and meta variables
|
2012-12-18 12:29:30 +00:00 |
|
kr.angelov
|
32905c8363
|
debugging infrastructure in the reasoner
|
2012-12-14 21:25:00 +00:00 |
|
kr.angelov
|
5cec2d5a50
|
bugfix for the reasoner in the C runtime
|
2012-12-14 21:24:17 +00:00 |
|
kr.angelov
|
8aefd1e072
|
The first prototype for exhaustive generation in the C runtime. The trees are always listed in decreasing probability order. There is also an API for generation from Python
|
2012-12-14 15:32:49 +00:00 |
|
kr.angelov
|
e1bab39458
|
bugfix in the lexer from the C runtime. the input sentence doesn't have to terminate with whitespace
|
2012-12-13 16:45:44 +00:00 |
|
kr.angelov
|
2ba632dc9f
|
a top-level API for parsing in the C runtime
|
2012-12-13 14:44:33 +00:00 |
|
kr.angelov
|
60942c440a
|
bugfix: the outside probability of a PgfItemConts must always be initialized to zero
|
2012-12-13 11:11:45 +00:00 |
|
kr.angelov
|
fe51a7fb98
|
bugfix: pgf_read_expr no longer requires a semicolon at the end of an abstract expression
|
2012-12-13 11:09:26 +00:00 |
|
kr.angelov
|
1376df457d
|
started an official API to the C runtime
|
2012-12-12 11:25:58 +00:00 |
|
kr.angelov
|
3182e382dc
|
bugfix for robust parsing with multi-word units
|
2012-12-11 12:57:22 +00:00 |
|
kr.angelov
|
1863e4c3d6
|
added experimental script for chunking in the C runtime
|
2012-12-03 10:07:54 +00:00 |
|
kr.angelov
|
f8c302f9ef
|
remove the duplicated definition of PgfProductionIdx in parser.c
|
2012-11-19 14:16:31 +00:00 |
|
kr.angelov
|
71b7c09ffe
|
bugfix for the building of bottom-up filter in the C runtime
|
2012-11-16 13:27:15 +00:00 |
|
kr.angelov
|
a3ba1991f4
|
revised heuristic in the statistical parser
|
2012-11-14 12:34:22 +00:00 |
|
kr.angelov
|
70c68f0527
|
bugfix in the statistical parser
|
2012-11-13 09:48:23 +00:00 |
|
kr.angelov
|
08ee662944
|
two simple heuristics which speed up the statistical parser more than seven times.
|
2012-11-12 22:17:40 +00:00 |
|
kr.angelov
|
68170d5b08
|
a simple refactoring in the statistical parser
|
2012-11-12 21:48:22 +00:00 |
|
kr.angelov
|
a2771552d6
|
more counters in the profiler for the statistical parser
|
2012-11-12 15:36:21 +00:00 |
|
kr.angelov
|
46de62c452
|
now we store the state instead of the offset for every continuation in the chart for the statistical parser
|
2012-11-12 14:04:52 +00:00 |
|
kr.angelov
|
9967c3ad04
|
in the statistical parser: move the outside probability from the parse items to their continuation. this makes the value slot shared between many items
|
2012-11-12 13:43:43 +00:00 |
|
kr.angelov
|
9d23093492
|
small refactoring in the C runtime
|
2012-11-12 13:05:35 +00:00 |
|
kr.angelov
|
a50c7c24b8
|
use size_t consistently as the type for constituent indices in the C runtime
|
2012-11-12 12:51:27 +00:00 |
|
kr.angelov
|
52255664be
|
use prob_t instead of float in a few places
|
2012-10-29 08:52:56 +00:00 |
|
kr.angelov
|
0ad2405d69
|
forgot to add one #ifdef
|
2012-10-25 18:37:22 +00:00 |
|
kr.angelov
|
9721833680
|
a major refactoring in the robust parser: bottom-up filtering and garbage collection for the chart
|
2012-10-25 14:42:53 +00:00 |
|
kr.angelov
|
18fe8af964
|
now the meta probability for a category is explicitly specified in the statistical model instead of computed internally. this avoids rounding errors while computing the sum of a large number of small values.
|
2012-09-24 09:37:21 +00:00 |
|
kr.angelov
|
bb15542a85
|
in the robust parser we don't have to care about trees which yeld empty strings. this makes the parser a lot faster
|
2012-09-24 09:30:20 +00:00 |
|
kr.angelov
|
44df7a33cf
|
the C runtime now has a type prob_t which is used only for probability values
|
2012-09-18 09:18:48 +00:00 |
|
kr.angelov
|
cd3cca4aa2
|
bugfix in the C parser
|
2012-09-06 14:52:19 +00:00 |
|
kr.angelov
|
3a352a953f
|
Use a separated tag for meta productions in the robust parser. This cleans up the code a lot
|
2012-06-13 05:49:30 +00:00 |
|
kr.angelov
|
7549a4876d
|
now there is a limit of 2000000 items in the chart of the robust parser. This prevents from explosion in the memory size but it will also prevent us from parsing some sentences.
|
2012-06-12 11:30:01 +00:00 |
|
kr.angelov
|
b765b0c054
|
now the robust parser is purely top-down and the meta rules compete on a fair basis with the grammar rules
|
2012-06-12 09:29:51 +00:00 |
|
kr.angelov
|
cab4602b62
|
the viterbi probability for the epsilon categories is now updated properly
|
2012-05-25 07:30:35 +00:00 |
|