explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2024-11-16 06:37:04 +03:00
Code Issues Packages Projects Releases Wiki Activity
11,104 Commits 68 Branches 175 Tags 279 MiB
a48e69d8d1
Commit Graph

868 Commits

Author SHA1 Message Date
Matthew Honnibal
8026c183d0 Add hacky logic to accelerate depth=0 case in parser 2017-05-23 11:06:49 -05:00
Matthew Honnibal
a8b6d11c5b Support optional maxout layer 2017-05-23 05:58:07 -05:00
Matthew Honnibal
c55b8fa7c5 Fix bugs in parse_batch 2017-05-23 05:57:52 -05:00
Matthew Honnibal
964707d795 Restore support for deeper networks in parser 2017-05-23 05:31:13 -05:00
Matthew Honnibal
6b918cc58e Support making updates periodically during training 2017-05-23 04:23:29 -05:00
Matthew Honnibal
3f725ff7b3 Roll back changes to parser update 2017-05-23 04:23:05 -05:00
Matthew Honnibal
3959d778ac Revert "Revert "WIP on improving parser efficiency"" 
This reverts commit 532afef4a8.
 2017-05-23 03:06:53 -05:00
Matthew Honnibal
532afef4a8 Revert "WIP on improving parser efficiency" 
This reverts commit bdaac7ab44.
 2017-05-23 03:05:25 -05:00
Matthew Honnibal
bdaac7ab44 WIP on improving parser efficiency 2017-05-23 02:59:31 -05:00
Matthew Honnibal
8a9e318deb Put the parsing loop in a nogil prange block 2017-05-22 17:58:12 -05:00
Matthew Honnibal
e2136232f9 Exclude states with no matching gold annotations from parsing 2017-05-22 10:30:12 -05:00
Matthew Honnibal
f00f821496 Fix pseudoprojectivity->nonproj 2017-05-22 06:14:42 -05:00
Matthew Honnibal
5d59e74cf6 PseudoProjectivity->nonproj 2017-05-22 05:49:53 -05:00
Matthew Honnibal
b45b4aa392 PseudoProjectivity --> nonproj 2017-05-22 05:17:44 -05:00
Matthew Honnibal
aae97f00e9 Fix nonproj import 2017-05-22 05:15:06 -05:00
Matthew Honnibal
2a5eb9f61e Make nonproj methods top-level functions, instead of class methods 2017-05-22 04:51:08 -05:00
Matthew Honnibal
33e2222839 Remove unused code in deprojectivize 2017-05-22 04:51:08 -05:00
Matthew Honnibal
025d9bbc37 Fix handling of non-projective deps 2017-05-22 04:51:08 -05:00
Matthew Honnibal
1b5fa68996 Do pseudo-projective pre-processing for parser 2017-05-22 04:51:08 -05:00
Matthew Honnibal
1d5d9838a2 Fix action collection for parser 2017-05-22 04:51:08 -05:00
Matthew Honnibal
3b7c108246 Pass tokvecs through as a list, instead of concatenated. Also fix padding 2017-05-20 13:23:32 -05:00
Matthew Honnibal
d52b65aec2 Revert "Move to contiguous buffer for token_ids and d_vectors" 
This reverts commit 3ff8c35a79.
 2017-05-20 11:26:23 -05:00
Matthew Honnibal
b272890a8c Try to move parser to simpler PrecomputedAffine class. Currently broken -- maybe the previous change 2017-05-20 06:40:10 -05:00
Matthew Honnibal
3ff8c35a79 Move to contiguous buffer for token_ids and d_vectors 2017-05-20 04:17:30 -05:00
Matthew Honnibal
8b04b0af9f Remove freqs from transition_system 2017-05-20 02:20:48 -05:00
Matthew Honnibal
a1ba20e2b1 Fix over-run on parse_batch 2017-05-19 18:57:30 -05:00
Matthew Honnibal
e84de028b5 Remove 'rebatch' op, and remove min-batch cap 2017-05-19 18:16:36 -05:00
Matthew Honnibal
c12ab47a56 Remove state argument in pipeline. Other changes 2017-05-19 13:26:36 -05:00
Matthew Honnibal
c2c825127a Fix use_params and pipe methods 2017-05-18 08:30:59 -05:00
Matthew Honnibal
fc8d3a112c Add util.env_opt support: Can set hyper params through environment variables. 2017-05-18 04:36:53 -05:00
Matthew Honnibal
d2626fdb45 Fix name error in nn parser 2017-05-18 04:31:01 -05:00
Matthew Honnibal
793430aa7a Get spaCy train command working with neural network 
* Integrate models into pipeline
* Add basic serialization (maybe incorrect)
* Fix pickle on vocab
 2017-05-17 12:04:50 +02:00
Matthew Honnibal
8cf097ca88 Redesign training to integrate NN components 
* Obsolete .parser, .entity etc names in favour of .pipeline
* Components no longer create models on initialization
* Models created by loading method (from_disk(), from_bytes() etc), or
    .begin_training()
* Add .predict(), .set_annotations() methods in components
* Pass state through pipeline, to allow components to share information
    more flexibly.
 2017-05-16 16:17:30 +02:00
Matthew Honnibal
5211645af3 Get data flowing through pipeline. Needs redesign 2017-05-16 11:21:59 +02:00
Matthew Honnibal
a9edb3aa1d Improve integration of NN parser, to support unified training API 2017-05-15 21:53:27 +02:00
Matthew Honnibal
4b9d69f428 Merge branch 'v2' into develop 
* Move v2 parser into nn_parser.pyx
* New TokenVectorEncoder class in pipeline.pyx
* New spacy/_ml.py module

Currently the two parsers live side-by-side, until we figure out how to
organize them.
 2017-05-14 01:10:23 +02:00
Matthew Honnibal
5cac951a16 Move new parser to nn_parser.pyx, and restore old parser, to make tests pass. 2017-05-14 00:55:01 +02:00
Matthew Honnibal
f8c02b4341 Remove cupy imports from parser, so it can work on CPU 2017-05-14 00:37:53 +02:00
Matthew Honnibal
e6d71e1778 Small fixes to parser 2017-05-13 17:19:04 -05:00
Matthew Honnibal
188c0f6949 Clean up unused import 2017-05-13 17:18:27 -05:00
Matthew Honnibal
f85c8464f7 Draft support of regression loss in parser 2017-05-13 17:17:27 -05:00
Matthew Honnibal
827b5af697 Update draft of parser neural network model 
Model is good, but code is messy. Currently requires Chainer, which may cause the build to fail on machines without a GPU.

Outline of the model:

We first predict context-sensitive vectors for each word in the input:

(embed_lower | embed_prefix | embed_suffix | embed_shape)
>> Maxout(token_width)
>> convolution ** 4

This convolutional layer is shared between the tagger and the parser. This prevents the parser from needing tag features.
To boost the representation, we make a "super tag" with POS, morphology and dependency label. The tagger predicts this
by adding a softmax layer onto the convolutional layer --- so, we're teaching the convolutional layer to give us a
representation that's one affine transform from this informative lexical information. This is obviously good for the
parser (which backprops to the convolutions too).

The parser model makes a state vector by concatenating the vector representations for its context tokens. Current
results suggest few context tokens works well. Maybe this is a bug.

The current context tokens:

* S0, S1, S2: Top three words on the stack
* B0, B1: First two words of the buffer
* S0L1, S0L2: Leftmost and second leftmost children of S0
* S0R1, S0R2: Rightmost and second rightmost children of S0
* S1L1, S1L2, S1R2, S1R, B0L1, B0L2: Likewise for S1 and B0

This makes the state vector quite long: 13*T, where T is the token vector width (128 is working well). Fortunately,
there's a way to structure the computation to save some expense (and make it more GPU friendly).

The parser typically visits 2*N states for a sentence of length N (although it may visit more, if it back-tracks
with a non-monotonic transition). A naive implementation would require 2*N (B, 13*T) @ (13*T, H) matrix multiplications
for a batch of size B. We can instead perform one (B*N, T) @ (T, 13*H) multiplication, to pre-compute the hidden
weights for each positional feature wrt the words in the batch. (Note that our token vectors come from the CNN
-- so we can't play this trick over the vocabulary. That's how Stanford's NN parser works --- and why its model
is so big.)

This pre-computation strategy allows a nice compromise between GPU-friendliness and implementation simplicity.
The CNN and the wide lower layer are computed on the GPU, and then the precomputed hidden weights are moved
to the CPU, before we start the transition-based parsing process. This makes a lot of things much easier.
We don't have to worry about variable-length batch sizes, and we don't have to implement the dynamic oracle
in CUDA to train.

Currently the parser's loss function is multilabel log loss, as the dynamic oracle allows multiple states to
be 0 cost. This is defined as:

(exp(score) / Z) - (exp(score) / gZ)

Where gZ is the sum of the scores assigned to gold classes. I'm very interested in regressing on the cost directly,
but so far this isn't working well.

Machinery is in place for beam-search, which has been working well for the linear model. Beam search should benefit
greatly from the pre-computation trick.
 2017-05-12 16:09:15 -05:00
Matthew Honnibal
b44f7e259c Clean up unused parser code 2017-05-08 15:42:04 +02:00
Matthew Honnibal
17efb1c001 Change width 2017-05-08 08:40:13 -05:00
Matthew Honnibal
bef89ef23d Mergery 2017-05-08 08:29:36 -05:00
Matthew Honnibal
50ddc9fc45 Fix infinite loop bug 2017-05-08 07:54:26 -05:00
Matthew Honnibal
a66a4a4d0f Replace einsums 2017-05-08 14:46:50 +02:00
Matthew Honnibal
8d2eab74da Use PretrainableMaxouts 2017-05-08 14:24:55 +02:00
Matthew Honnibal
2e2268a442 Precomputable hidden now working 2017-05-08 11:36:37 +02:00
Matthew Honnibal
10682d35ab Get pre-computed version working 2017-05-08 00:38:35 +02:00
Matthew Honnibal
35458987e8 Checkpoint -- nearly finished reimpl 2017-05-07 23:05:01 +02:00
Matthew Honnibal
4441866f55 Checkpoint -- nearly finished reimpl 2017-05-07 22:47:06 +02:00
Matthew Honnibal
6782eedf9b Tmp GPU code 2017-05-07 11:04:24 -05:00
Matthew Honnibal
e420e5a809 Tmp 2017-05-07 07:31:09 -05:00
Matthew Honnibal
700979fb3c CPU/GPU compat 2017-05-07 04:01:11 +02:00
Matthew Honnibal
f99f5b75dc working residual net 2017-05-07 03:57:26 +02:00
Matthew Honnibal
bdf2dba9fb WIP on refactor, with hidde pre-computing 2017-05-07 02:02:43 +02:00
Matthew Honnibal
b439e04f8d Learning smoothly 2017-05-06 20:38:12 +02:00
Matthew Honnibal
08bee76790 Learns things 2017-05-06 18:24:38 +02:00
Matthew Honnibal
bcf4cd0a5f Learns things 2017-05-06 17:37:36 +02:00
Matthew Honnibal
8e48b58cd6 Gradients look correct 2017-05-06 16:47:15 +02:00
Matthew Honnibal
7e04260d38 Data running through, likely errors in model 2017-05-06 14:22:20 +02:00
Matthew Honnibal
ef4fa594aa Draft of NN parser, to be tested 2017-05-05 19:20:39 +02:00
Matthew Honnibal
ccaf26206b Pseudocode for parser 2017-05-04 12:17:59 +02:00
Matthew Honnibal
2da16adcc2 Add dropout optin for parser and NER 
Dropout can now be specified in the `Parser.update()` method via
the `drop` keyword argument, e.g.

    nlp.entity.update(doc, gold, drop=0.4)

This will randomly drop 40% of features, and multiply the value of the
others by 1. / 0.4. This may be useful for generalising from small data
sets.

This commit also patches the examples/training/train_new_entity_type.py
example, to use dropout and fix the output (previously it did not output
the learned entity).
 2017-04-27 13:18:39 +02:00
Matthew Honnibal
d2436dc17b Update fix for Issue #999 2017-04-23 18:14:37 +02:00
Matthew Honnibal
60703cede5 Ensure noun chunks can't be nested. Closes #955 2017-04-23 17:56:39 +02:00
Matthew Honnibal
4eef200bab Persist the actions within spacy.parser.cfg 2017-04-20 17:02:44 +02:00
Matthew Honnibal
137b210bcf Restore use of FTRL training 2017-04-16 18:02:42 +02:00
Matthew Honnibal
45464d065e Remove print statement 2017-04-15 16:11:43 +02:00
Matthew Honnibal
c76cb8af35 Fix training for new labels 2017-04-15 16:11:26 +02:00
Matthew Honnibal
4884b2c113 Refix StepwiseState 2017-04-15 16:00:28 +02:00
Matthew Honnibal
1a98e48b8e Fix Stepwisestate' 2017-04-15 13:35:01 +02:00
ines
0739ae7b76 Tidy up and fix formatting and imports 2017-04-15 13:05:15 +02:00
Matthew Honnibal
354458484c WIP on add_label bug during NER training 
Currently when a new label is introduced to NER during training,
it causes the labels to be read in in an unexpected order. This
invalidates the model.
 2017-04-14 23:52:17 +02:00
Matthew Honnibal
49e2de900e Add costs property to StepwiseState, to show which moves are gold. 2017-04-10 11:37:04 +02:00
Matthew Honnibal
cc36c308f4 Fix noun_chunk rules around coordination 
Closes #693.
 2017-04-07 17:06:40 +02:00
Matthew Honnibal
1bb7b4ca71 Add comment 2017-03-31 13:59:19 +02:00
Matthew Honnibal
47a3ef06a6 Unhack deprojetivization, moving it into pipeline 
Previously the deprojectivize() call was attached to the transition
system, and only called for German. Instead it should be a separate
process, called after the parser. This makes it available for any
language. Closes #898.
 2017-03-31 12:31:50 +02:00
Matthew Honnibal
a9b1f23c7d Enable regression loss for parser 2017-03-26 09:26:30 -05:00
Matthew Honnibal
b487b8735a Decrease beam density, and fix Python 3 problem in beam 2017-03-20 12:56:05 +01:00
Matthew Honnibal
c90dc7ac29 Clean up state initiatisation in transition system 2017-03-16 11:59:11 -05:00
Matthew Honnibal
a46933a8fe Clean up FTRL parsing stuff. 2017-03-16 11:58:20 -05:00
Matthew Honnibal
2611ac2a89 Fix scorer bug for NER, related to ambiguity between missing annotations and misaligned tokens 2017-03-16 09:38:28 -05:00
Matthew Honnibal
3d0833c3df Fix off-by-1 in parse features fill_context 2017-03-15 19:55:35 -05:00
Matthew Honnibal
4ef68c413f Approximate cost in Break transition, to speed things up a bit. 2017-03-15 16:40:27 -05:00
Matthew Honnibal
8543db8a5b Use ftrl optimizer in parser 2017-03-15 11:56:37 -05:00
Matthew Honnibal
d719f8e77e Use nogil in parser, and set L1 to 0.0 by default 2017-03-15 09:31:01 -05:00
Matthew Honnibal
c61c501406 Update beam-parser to allow parser to maintain nogil 2017-03-15 09:30:22 -05:00
Matthew Honnibal
c79b3129e3 Fix setting of empty lexeme in initial parse state 2017-03-15 09:26:53 -05:00
Matthew Honnibal
6c4108c073 Add header for beam parser 2017-03-11 12:45:12 -06:00
Matthew Honnibal
931feb3360 Allow beam parsing for NER 2017-03-11 11:12:01 -06:00
Matthew Honnibal
ca9c8c57c0 Add iteration argument to parser.update 2017-03-11 07:00:47 -06:00
Matthew Honnibal
d59c6926c1 I think this fixes the segfault 2017-03-11 06:58:34 -06:00
Matthew Honnibal
318b9e32ff WIP on beam parser. Currently segfaults. 2017-03-11 06:19:52 -06:00
Matthew Honnibal
b0d80dc9ae Update name of 'train' function in BeamParser 2017-03-10 14:35:43 -06:00
Matthew Honnibal
d11f1a4ddf Record negative costs in non-monotonic arc eager oracle 2017-03-10 11:22:04 -06:00
Matthew Honnibal
ecf91a2dbb Support beam parser 2017-03-10 11:21:21 -06:00
Matthew Honnibal
c62da02344 Use ftrl training, to learn compressed model. 2017-03-09 18:43:21 -06:00
Matthew Honnibal
40703988bc Use FTRL training in parser 2017-03-08 01:38:51 +01:00
Roman Inflianskas
66e1109b53 Add support for Universal Dependencies v2.0 2017-03-03 13:17:34 +01:00
Matthew Honnibal
97a1286129 Revert changes to tagger and parser for thinc 6 2017-01-09 10:08:34 -06:00
Matthew Honnibal
af81ac8bb0 Use thinc 6.0 2016-12-29 11:58:42 +01:00
Matthew Honnibal
bc0a202c9c Fix unicode problem in nonproj module 2016-11-25 17:29:17 -06:00
Matthew Honnibal
159e8c46e1 Merge old training fixes with newer state 2016-11-25 09:16:36 -06:00
Matthew Honnibal
39341598bb Fix NER label calculation 2016-11-25 09:02:22 -06:00
Matthew Honnibal
ca773a1f53 Tweak arc_eager n_gold to deal with negative costs, and improve error message. 2016-11-25 09:01:52 -06:00
Matthew Honnibal
608d8f5421 Pass cfg through parser, and have is_valid default to 1, not 0 when resetting state 2016-11-25 09:00:21 -06:00
Matthew Honnibal
b8c4f5ea76 Allow German noun chunks to work on Span 
Update the German noun chunks iterator, so that it also works on Span objects.
 2016-11-24 23:30:15 +11:00
Pokey Rule
3e3bda142d Add noun_chunks to Span 2016-11-24 10:47:20 +00:00
Matthew Honnibal
b86f8af0c1 Fix doc strings 2016-11-01 12:25:36 +01:00
Matthew Honnibal
708ea22208 Infer types in transition_system.pyx 2016-10-27 18:08:13 +02:00
Matthew Honnibal
301f3cc898 Fix Issue #429. Add an initialize_state method to the named entity recogniser that adds missing entity types. This is a messy place to add this, because it's strange to have the method mutate state. A better home for this logic could be found. 2016-10-27 18:01:55 +02:00
Matthew Honnibal
03a520ec4f Change signature of Parser.parseC, so that nr_class is read from the transition system. This allows the transition system to modify the number of actions in initialize_state. 2016-10-27 17:58:56 +02:00
Matthew Honnibal
a209b10579 Improve error message when oracle fails for non-projective trees, re Issue #571. 2016-10-24 20:31:30 +02:00
Matthew Honnibal
3e688e6d4b Fix issue #514 -- serializer fails when new entity type has been added. The fix here is quite ugly. It's best to add the entities ASAP after loading the NLP pipeline, to mitigate the brittleness. 2016-10-23 17:45:44 +02:00
Matthew Honnibal
59038f7efa Restore support for prior data format -- specifically, the labels field of the config. 2016-10-17 00:53:26 +02:00
Matthew Honnibal
7887ab3b36 Fix default use of feature_templates in parser 2016-10-16 21:41:56 +02:00
Matthew Honnibal
f787cd29fe Refactor the pipeline classes to make them more consistent, and remove the redundant blank() constructor. 2016-10-16 21:34:57 +02:00
Matthew Honnibal
274a4d4272 Fix queue Python property in StateClass 2016-10-16 17:04:41 +02:00
Matthew Honnibal
e8c8aa08ce Make action_name optional in StepwiseState 2016-10-16 17:04:16 +02:00
Matthew Honnibal
4fc56d4a31 Rename 'labels' to 'actions' in parser options 2016-10-16 11:42:26 +02:00
Matthew Honnibal
3259a63779 Whitespace 2016-10-16 01:47:28 +02:00
Matthew Honnibal
d9ae2d68af Load features by string-name for backwards compatibility. 2016-10-12 20:15:11 +02:00
Matthew Honnibal
3a03c668c3 Fix message in ParserStateError 2016-10-12 14:44:31 +02:00
Matthew Honnibal
6bf505e865 Fix error on ParserStateError 2016-10-12 14:35:55 +02:00
Matthew Honnibal
ea23b64cc8 Refactor training, with new spacy.train module. Defaults still a little awkward. 2016-10-09 12:24:24 +02:00
Matthew Honnibal
1d70db58aa Revert "Changes to iterators.pyx for new StringStore scheme" 
This reverts commit 4f794b215a.
 2016-09-30 20:19:53 +02:00
Matthew Honnibal
9e09b39b9f Revert "Changes to transition systems for new StringStore scheme" 
This reverts commit 0442e0ab1e.
 2016-09-30 20:11:49 +02:00
Matthew Honnibal
e3285f6f30 Revert "Fix report of ParserStateError" 
This reverts commit 78f19baafa.
 2016-09-30 20:11:33 +02:00
Matthew Honnibal
78f19baafa Fix report of ParserStateError 2016-09-30 19:59:22 +02:00
Matthew Honnibal
0442e0ab1e Changes to transition systems for new StringStore scheme 2016-09-30 19:58:51 +02:00
Matthew Honnibal
4f794b215a Changes to iterators.pyx for new StringStore scheme 2016-09-30 19:57:49 +02:00
Matthew Honnibal
4cbf0d3bb6 Handle errors when no valid actions are available, pointing users to the issue tracker. 2016-09-27 19:19:53 +02:00
Matthew Honnibal
430473bd98 Raise errors when no actions are available, re Issue #429 2016-09-27 19:09:37 +02:00
Matthew Honnibal
8e7df3c4ca Expect the parser data, if parser.load() is called. 2016-09-27 14:02:12 +02:00
Matthew Honnibal
a44763af0e Fix Issue #469: Incorrectly cased root label in noun chunk iterator 2016-09-27 13:13:01 +02:00
Matthew Honnibal
e07b9665f7 Don't expect parser model 2016-09-26 18:09:33 +02:00
Matthew Honnibal
ee6fa106da Fix parser features 2016-09-26 17:57:32 +02:00
Matthew Honnibal
e607e4b598 Fix parser loading 2016-09-26 17:51:11 +02:00
Matthew Honnibal
2debc4e0a2 Add .blank() method to Parser. Start housing default dep labels and entity types within the Defaults class. 2016-09-26 11:57:54 +02:00
Matthew Honnibal
fd65cf6cbb Finish refactoring data loading 2016-09-24 20:26:17 +02:00
Matthew Honnibal
83e364188c Mostly finished loading refactoring. Design is in place, but doesn't work yet. 2016-09-24 15:42:01 +02:00
Matthew Honnibal
60fdf4d5f1 Remove commented out debuggng code 2016-09-24 01:17:18 +02:00
Matthew Honnibal
070af4af9d Revert "* Working neural net, but features hacky. Switching to extractor." 
This reverts commit 7c2f1a673b.
 2016-09-21 12:26:14 +02:00
Matthew Honnibal
7c2f1a673b * Working neural net, but features hacky. Switching to extractor. 2016-05-26 19:06:10 +02:00
Matthew Honnibal
13fad36e49 * Cosmetic change to english noun chunks iterator -- use enumerate instead of range loop 2016-05-20 10:11:05 +02:00
Wolfgang Seeker
7b78239436 add fix for German noun chunk iterator (issue #365) 2016-05-06 01:41:26 +02:00
Matthew Honnibal
bb94022975 * Fix Issue #365: Error introduced during noun phrase chunking, due to use of corrected PRON/PROPN/etc tags. 2016-05-06 00:21:05 +02:00
Wolfgang Seeker
dbf8f5f3ec fix bug in StateC.set_break() 2016-05-05 15:15:34 +02:00
Wolfgang Seeker
3c44b5dc1a call deprojectivization after parsing 2016-05-05 15:10:36 +02:00
Matthew Honnibal
472f576b82 * Deprojectivize German parses 2016-05-05 15:01:10 +02:00
Wolfgang Seeker
e4ea2bea01 fix whitespace 2016-05-04 07:40:38 +02:00
Wolfgang Seeker
5bf2fd1f78 make the code less cryptic 2016-05-03 17:19:05 +02:00
Wolfgang Seeker
a06fca9fdf German noun chunk iterator now doesn't return tokens more than once 2016-05-03 16:58:59 +02:00
Wolfgang Seeker
7b246c13cb reformulate noun chunk tests for English 2016-05-03 14:24:35 +02:00
Matthew Honnibal
1f1532142f * Fix cost calculation on non-monotonic oracle 2016-05-03 00:21:08 +02:00
Matthew Honnibal
508fd1f6dc * Refactor noun chunk iterators, so that they're simple functions. Install the iterator when the Doc is created, but allow users to write to the noun_chunk_iterator attribute. The iterator functions accept an object and yield (int start, int end, int label) triples. 2016-05-02 14:25:10 +02:00
Matthew Honnibal
77609588b6 * Fix assignment of root label to words left as root implicitly, after parsing ends. 2016-04-25 19:41:59 +00:00
Matthew Honnibal
7c2d2deaa7 * Revise transition system so that the Break transition retains sole responsibility for setting sentence boundaries. Re Issue #322 2016-04-25 19:41:59 +00:00
Wolfgang Seeker
12024b0b0a bugfix: introducing multiple roots now updates original head's properties 
adjust tests to rely less on statistical model
 2016-04-20 16:42:41 +02:00
Wolfgang Seeker
b98cc3266d bugfix: iterators now reset properly when called a second time 2016-04-15 17:49:16 +02:00
Wolfgang Seeker
289b10f441 remove some comments 2016-04-14 15:37:51 +02:00
Wolfgang Seeker
d99a9cbce9 different handling of space tokens 
space tokens are now always attached to the previous non-space token
there are two exceptions:
leading space tokens are attached to the first following non-space token
in input that consists exclusively of space tokens, the last space token
is the head of all others.
 2016-04-13 15:28:28 +02:00
Wolfgang Seeker
d328e0b4a8 Merge branch 'master' into space_head_bug 2016-04-11 12:11:01 +02:00
Wolfgang Seeker
80bea62842 bugfix in unit test 2016-04-08 16:46:44 +02:00
Wolfgang Seeker
1fe911cdb0 bigfix 2016-04-07 18:19:51 +02:00
Matthew Honnibal
872695759d Merge pull request #306 from wbwseeker/german_noun_chunks 
add German noun chunk functionality
 2016-04-08 00:54:24 +10:00
Wolfgang Seeker
7195b6742d add restrictions to L-arc and R-arc to prevent space heads 2016-03-28 10:40:52 +02:00
Wolfgang Seeker
5e2e8e951a add baseclass DocIterator for iterators over documents 
add classes for English and German noun chunks

the respective iterators are set for the document when created by the parser
as they depend on the annotation scheme of the parsing model
 2016-03-16 15:53:35 +01:00
Wolfgang Seeker
46e3f979f1 add function for setting head and label to token 
change PseudoProjectivity.deprojectivize to use these functions
 2016-03-11 17:31:06 +01:00
Wolfgang Seeker
7adbd7a785 replace Counter with normal dict 2016-03-03 21:36:27 +01:00
Wolfgang Seeker
1ae487a4f6 add backwards compatibility with python 2.6 2016-03-03 21:18:12 +01:00
Wolfgang Seeker
72b8df0684 turned PseudoProjectivity into a normal python class 2016-03-03 19:05:08 +01:00
Wolfgang Seeker
690c5acabf adjust train.py to train both english and german models 2016-03-03 15:21:00 +01:00
Wolfgang Seeker
3448cb40a4 integrated pseudo-projective parsing into parser 
- nonproj.pyx holds a class PseudoProjectivity which currently holds
  all functionality to implement Nivre & Nilsson 2005's pseudo-projective
  parsing using the HEAD decoration scheme
- changed lefts/rights in Token to account for possible non-projective
  structures
 2016-03-01 10:09:08 +01:00
Wolfgang Seeker
56b7210e82 moved nonproj.py to syntax/nonproj.pyx 2016-02-25 15:08:49 +01:00
Matthew Honnibal
1b83cb9dfa * Fix Issue #251: Incorrect right edge calculation on left-clobber low in the tree 2016-02-07 00:00:42 +01:00
Matthew Honnibal
4412a70dc5 * Initialize StateC._empty_token to 0, to avoid undefined behaviour. 2016-02-06 13:34:38 +01:00
Matthew Honnibal
1b41f868d2 * Check for errors in parser, and parallelise the left-over batch 2016-02-06 10:06:30 +01:00
Matthew Honnibal
165ca28b80 * Set is_parsed flag in Parser.pipe 2016-02-05 19:51:44 +01:00
Matthew Honnibal
bdd579db0a * Set is_parsed flag in Parser.pipe 2016-02-05 19:50:11 +01:00
Matthew Honnibal
b04c9aad71 * Fix off-by-one in Parser.pipe 2016-02-05 19:37:50 +01:00
Matthew Honnibal
048dfe35aa * cimport cython.parallel 2016-02-05 12:20:42 +01:00
Matthew Honnibal
8a13cebdcc * Update for modified thinc interface 2016-02-05 11:44:39 +01:00
Matthew Honnibal
84b247ef83 * Add a .pipe method, that takes a stream of input, operates on it, and streams the output. Internally, the stream may be buffered, to allow multi-threading. 2016-02-03 02:10:58 +01:00
Matthew Honnibal
e3db39dd21 * Fix compiler warning about signed/unsigned comparison 2016-02-01 09:08:07 +01:00
Matthew Honnibal
b3802562d6 Merge branch 'rethinc2' of https://github.com/honnibal/spaCy into rethinc2 2016-02-01 08:59:24 +01:00
Matthew Honnibal
4b08a3fafd * Fix merge conflict 2016-02-01 08:58:18 +01:00
Matthew Honnibal
5188f6d9d8 * Fix parseC function 2016-02-01 08:48:48 +01:00
Matthew Honnibal
bcf8f7ba40 * Add a parse_batch method to Parser, that releases the GIL around a batch of documents. 2016-02-01 08:34:55 +01:00
Matthew Honnibal
d5579cd0d8 Merge branch 'rethinc2' of https://github.com/honnibal/spaCy into rethinc2 2016-02-01 03:08:49 +01:00
Matthew Honnibal
490ba65398 * Use openmp in parser 2016-02-01 03:08:42 +01:00
Matthew Honnibal
cb78d91ec5 * Fix ArcEager.set_valid 2016-02-01 03:07:37 +01:00
Matthew Honnibal
28e5ad62bc * Pass a StateC pointer into the transition and validation methods in the parser, so that the GIL can be released over a batch of documents 2016-02-01 03:00:15 +01:00
Matthew Honnibal
a47f00901b * Pass a StateC pointer into the transition and validation methods in the parser, so that the GIL can be released over a batch of documents 2016-02-01 02:58:14 +01:00
Matthew Honnibal
daaad66448 * Now fully proxied 2016-02-01 02:37:08 +01:00
Matthew Honnibal
7a0e3bb9c1 * Continue proxying. Some problem currently 2016-02-01 02:22:21 +01:00
Matthew Honnibal
2169bbb7ea * Shadow StateClass with StateC, to start proxying 2016-02-01 01:16:14 +01:00
Matthew Honnibal
2fa228458e * Add _state file, which StateClass will proxy to 2016-02-01 01:09:21 +01:00
Matthew Honnibal
9410e74c92 * Switch parser to use nogil functions 2016-01-30 20:27:07 +01:00
Matthew Honnibal
10877a7791 * Update for thinc 5.0, including changing cost from int to weight_t, and updating the tagger and parser 2016-01-30 14:31:36 +01:00
Matthew Honnibal
84c5dfbfc3 * Clean up debugging python list 2016-01-19 20:10:32 +01:00
Matthew Honnibal
04d0686b26 * Make TransitionSystem.add_action idempotent, i.e. ignore duplicate added actions. 2016-01-19 20:10:04 +01:00
Matthew Honnibal
65c5bc4988 * Add add_label method, to allow users to register new entity types and dependency labels. 2016-01-19 19:11:02 +01:00
Matthew Honnibal
151aa0b0e2 * Allow users to add_label, in order to extend the entity recogniser to new classes. Does not by itself add a class to the model 2016-01-19 19:09:33 +01:00
Matthew Honnibal
c8e0011ebc * Add iterators to the NER and parser transition systems, to get the action types 2016-01-19 19:07:43 +01:00
Matthew Honnibal
04177debd0 * Unwind limit to sentence boundary detection that prevents it from inserting boundaries on whitespace. Replace it with a check for whitespace in StateClass.fast_forward, so that whitespace is LeftArced when it's on the stack. This should prevent the previous problem of whitespace-only sentences. Should fix Issue #184, but may cause further problems. Needs testing. 2016-01-19 02:54:15 +01:00
Matthew Honnibal
3dc398b727 * Fix merge conflict in requirements.txt 2016-01-16 16:20:49 +01:00
Matthew Honnibal
c025a0c64b * Check for KeyboardInerrupt in parser.__call__ 2016-01-16 16:18:44 +01:00
Matthew Honnibal
aec130af56 Use util.Package class for io 
Previous Sputnik integration caused API change: Vocab, Tagger, etc
were loaded via a from_package classmethod, that required a
sputnik.Package instance. This forced users to first create a
sputnik.Sputnik() instance, in order to acquire a Package via
sp.pool().

Instead I've created a small file-system shim, util.Package, which
allows classes to have a .load() classmethod, that accepts either
util.Package objects, or strings. We can later gut the internals
of this and make it a proxy for Sputnik if we need more functionality
that should live in the Sputnik library.

Sputnik is now only used to download and install the data, in
spacy.en.download
 2015-12-29 18:00:48 +01:00
Matthew Honnibal
5623242b3e * Adjust NER rules, so that U entries in gazetteer don't become B moves to the model 2015-11-12 04:48:23 +11:00
Matthew Honnibal
44fbdc7260 * Fix bug in NER transition system, that sometimes left no valid moves 2015-11-08 16:19:12 +01:00
Matthew Honnibal
e92371bb54 * Fix rule that made Last action invalid if there was a preset of O, since if the entity is already open, that ship has sailed. 2015-11-08 22:17:51 +11:00
Matthew Honnibal
6f47074214 * Make constructor of ParserModel and TaggerModel the same as AveragedPerceptron, for each pickling. 2015-11-07 18:25:17 +11:00
Matthew Honnibal
1cfa20fb17 * Fix sentence-final whitespace issue 2015-11-07 17:34:46 +11:00
Matthew Honnibal
888c05a7fa * Fix variable naming in StepwiseState, for thinc 4.0 2015-11-07 11:02:44 +11:00
Matthew Honnibal
fc2185bfe3 * Fix variable naming in StepwiseState, for thinc 4.0 2015-11-07 10:48:31 +11:00
Matthew Honnibal
954442a807 * Fix variable naming in StepwiseState, for thinc 4.0 2015-11-07 10:30:45 +11:00
Matthew Honnibal
af70dc166a * Fix Last restriction, that was supposed to prevent conflicts with presets, but was incorrect. 2015-11-07 09:52:00 +11:00
Matthew Honnibal
a06e3c8963 * Fix bone-headed mistake in StateClass.E 2015-11-07 07:35:28 +11:00
Matthew Honnibal
d24b8509e4 * Correct screw ups from the previous commits 2015-11-07 06:51:41 +11:00
Matthew Honnibal
5efad178b5 * Set ent tag when close entity 2015-11-07 06:09:25 +11:00
Matthew Honnibal
9285f01d26 * Fix broken StateClass.E tracking 2015-11-07 06:06:39 +11:00
Matthew Honnibal
19136b0e7d * Add better debug message for illegal move 2015-11-07 05:34:37 +11:00
Matthew Honnibal
2733816b7b * Fix whitespace 2015-11-07 05:31:06 +11:00
Matthew Honnibal
01ab464383 * Prevent Begin and In moves from applying in NER if we're at the last token of a sentence, as this would mean the entity would span over a sentence boundary. Re Issue #169 2015-11-07 05:30:44 +11:00
Matthew Honnibal
b65633f270 * Fix function that returns nth entity in StateClass. Was only returning the first. 2015-11-07 05:29:11 +11:00
Matthew Honnibal
3c162dcac3 * Refactor away from the _ml module, to use thinc 4.0. Still some work needs to be done, e.g. to add __reduce__ to the models, more testing, etc. 2015-11-07 03:24:30 +11:00
Matthew Honnibal
b9991fbd20 * Update to use thinc 3.0 2015-11-06 00:25:59 +11:00
Matthew Honnibal
68f479e821 * Rename Doc.data to Doc.c 2015-11-04 00:15:14 +11:00
Matthew Honnibal
329ae57520 * Fix whitespace attachment thing 2015-10-13 09:46:38 +02:00
Matthew Honnibal
37919eac82 * Fix whitespace attachment in simpler way. Leaves problem with setting left/right children. 2015-10-13 18:23:24 +11:00
Matthew Honnibal
c70eb776ae * Fix whitespace attachment, so that left/right children are consistent with head. 2015-10-13 15:58:22 +11:00
Matthew Honnibal
20fd36a0f7 * Very scrappy, likely buggy first-cut pickle implementation, to work on Issue #125: allow pickle for Apache Spark. The current implementation sends stuff to temp files, and does almost nothing to ensure all modifiable state is actually preserved. The Language() instance is a deep tree of extension objects, and if pickling during training, some of the C-data state is hard to preserve. 2015-10-13 13:44:41 +11:00
Matthew Honnibal
9dd2f25c74 * Fix Issue #131: Force whitespace characters to attach syntactically to previous token, and ensure they cannot serve as stand-alone 'sentence' units. 2015-10-10 15:53:30 +11:00
Matthew Honnibal
8b39feefbe * Add dependency post-process rule to ensure spaces are attached to neighbouring tokens, so that they can't be sentence boundaries 2015-10-10 15:32:13 +11:00
Matthew Honnibal
0e24d099a1 * Fix L/R edge bug, by ensuring l_edge and r_edge are preset, and fixing the way the edge update in del_arc. Bugs keep arising here because the edges are absolute positions, where everything else is relative. I'm also not 100% convinced that del_arc is handled correctly. Do we need to update the parents? 2015-09-09 03:40:44 +02:00
Matthew Honnibal
86c888667f * Merge in changes from de branch 2015-09-06 19:49:28 +02:00
Matthew Honnibal
5edac11225 * Wrap self.parse in nogil, and break if an invalid move is predicted. The invalid break is a work-around that papers over likely bugs, but we can't easily break in the nogil block, and otherwise we'll get an infinite loop. Need to set this as an error flag. 2015-09-06 04:15:00 +02:00
Matthew Honnibal
a3d5e6c0dd * Reform constructor and save/load workflow in parser model 2015-08-26 19:19:01 +02:00
Matthew Honnibal
bf38b3b883 * Hack on l/r reversal bug 2015-08-10 05:58:43 +02:00
Matthew Honnibal
6116413b47 * Fix label prediction in StepwiseState 2015-08-10 05:05:31 +02:00
Matthew Honnibal
2c9753eff2 * Whitespace 2015-08-10 00:09:02 +02:00
Matthew Honnibal
9de98f5a6f * Add Parser.stepthrough method, with context manager 2015-08-10 00:08:46 +02:00
Matthew Honnibal
fe43f8cf39 * Whitespace 2015-08-09 02:31:53 +02:00
Matthew Honnibal
9c090945e0 * Add Parser.predict method, and clean up Parser.get_state 2015-08-09 02:29:58 +02:00
Matthew Honnibal
04fccfb984 * Fix get_state for parser prediction 2015-08-09 02:11:22 +02:00
Matthew Honnibal
55fde0e240 * Fix get_state 2015-08-09 01:45:30 +02:00
Matthew Honnibal
f0f4fa9838 * Fix Parser.get_state 2015-08-09 01:40:13 +02:00
Matthew Honnibal
18331dca89 * Add continue_for argument to parser 'partial' function, which is now renamed to get_state 2015-08-09 01:31:54 +02:00
Matthew Honnibal
0653288fa5 * Fix stateclass.queue 2015-08-09 00:39:02 +02:00
Matthew Honnibal
9de218b7ba * Fix Parser.partial function 2015-08-08 23:45:18 +02:00
Matthew Honnibal
cc9deae960 * Add is_valid method to transition_system 2015-08-08 23:36:18 +02:00
Matthew Honnibal
2a46c77324 * Whitespace 2015-08-08 23:35:59 +02:00
Matthew Honnibal
7bafc789e7 * Add stack and queue properties to stateclass, for python access 2015-08-08 23:32:42 +02:00
Matthew Honnibal
3af938365f * Add function partial to Parser 2015-08-08 23:32:15 +02:00
Matthew Honnibal
76a1f0481a * Whitespace 2015-08-08 23:31:54 +02:00
Matthew Honnibal
59c3bf60a6 * Ensure entity recognizer doesn't over-write preset types 2015-08-06 16:09:08 +02:00
Matthew Honnibal
9c1724ecae * Gazetteer stuff working, now need to wire up to API 2015-08-06 00:35:40 +02:00
Matthew Honnibal
a8bbd7312c * Hackishly patch long dependencies problem 2015-07-28 00:14:29 +02:00
Matthew Honnibal
bb583f7f09 * Hackishly patch long dependencies problem 2015-07-27 23:14:33 +02:00
Matthew Honnibal
823ef4a00b * Remove profile declarations 2015-07-25 18:13:06 +02:00
Matthew Honnibal
aa28e2e01d * Release the GIL around parse function 2015-07-24 04:53:27 +02:00
Matthew Honnibal
d5255aad77 * Update freqs for missing tags in ner, for serializer 2015-07-23 01:17:11 +02:00
Matthew Honnibal
12699a1152 * Set initial freqs, to avoid missing values in serializer 2015-07-23 01:16:27 +02:00
Matthew Honnibal
317cbbc015 * Serialization round trip now working with decent API, but with rough spots in the organisation and requiring vocabulary to be fixed ahead of time. 2015-07-19 15:18:17 +02:00
Matthew Honnibal
b1d74ce60d * Remove unused joint.pyx and joint.pxd files 2015-07-17 23:31:44 +02:00
Matthew Honnibal
fb0a641a2d * Don't release the gil around Parser.parse. Does this indicate thread problems? 2015-07-17 23:07:37 +02:00
Matthew Honnibal
e29daea85f * Fix bint/int typing problem in TransitionSystem. In C++ bint* means bool*, but in C it means int*. So, type-casting to bint* is unsafe. 2015-07-17 22:37:24 +02:00
Matthew Honnibal
45ae1ce428 * Remove unused declaration in parser 2015-07-16 01:27:11 +02:00
Matthew Honnibal
9a8db9743c * Remove gil from parser.call 2015-07-14 23:47:33 +02:00
Matthew Honnibal
38ca0c33f5 Merge branch 'neuralnet' into refactor 
Mostly refactors parser, to use new thinc3.2 Example class.
Aim is to remove use of shared memory, so that we can parallelize
over documents easily.

Conflicts:
	setup.py
	spacy/syntax/parser.pxd
	spacy/syntax/parser.pyx
	spacy/syntax/stateclass.pyx
 2015-07-14 14:13:47 +02:00
Matthew Honnibal
6eef0bf9ab * Break up tokens.pyx into tokens/doc.pyx, tokens/token.pyx, tokens/spans.pyx 2015-07-13 20:20:58 +02:00
Matthew Honnibal
55f1042443 * Improve efficiency of L and R features, correcting the non-linear-in-length problem. 2015-07-09 12:17:26 +02:00
Matthew Honnibal
70d2acb579 * Fix edge features 2015-07-09 12:15:01 +02:00
Matthew Honnibal
adb868bdad * Add warning for models not found in parser 2015-07-08 20:04:55 +02:00
Matthew Honnibal
05b28ec9eb * Add warning for models not found in parser 2015-07-08 20:02:13 +02:00
Matthew Honnibal
ef700401a6 * Add warning for models not found in parser 2015-07-08 20:00:46 +02:00
Matthew Honnibal
6218d8b389 * Add warning for models not found in parser 2015-07-08 19:59:16 +02:00
Matthew Honnibal
f6a6c39ce8 * Add warning for models not found in parser 2015-07-08 19:52:30 +02:00
Matthew Honnibal
0ceb1f71c2 * Update parse features 2015-07-08 19:11:36 +02:00
Matthew Honnibal
bb522496dd * Rename Tokens to Doc 2015-07-08 18:53:00 +02:00
Matthew Honnibal
ff885e8511 * Add ParserFactory convenience function 2015-07-08 12:35:46 +02:00
Matthew Honnibal
52fd80c6c6 * Add experimental supersense features for parsing, based on lookup into wordnet. 2015-07-01 20:12:44 +02:00
Matthew Honnibal
e20106fdff * Begin reorganizing neuralnet work 2015-06-30 14:26:32 +02:00
Matthew Honnibal
3bb5876c5a * Inline methods in StateClass 2015-06-29 01:10:14 +02:00
Matthew Honnibal
313a7f87b3 * Inline methods in StateClass 2015-06-29 01:06:28 +02:00
Matthew Honnibal
a02fd3af5d * Check valency in L and R feature methods, to make feaure calculation faster 2015-06-29 00:27:56 +02:00
Matthew Honnibal
5d870720bc * Check valency in L and R feature methods, to make feaure calculation faster 2015-06-29 00:17:29 +02:00
Matthew Honnibal
f4986d5d3c * Use new Example class 2015-06-28 22:36:03 +02:00
Matthew Honnibal
735f1af91f * Fix neural net stuff 2015-06-28 11:44:58 +02:00
Matthew Honnibal
e7003f1cf3 * Remove hard-coding of vector lengths 2015-06-28 11:37:17 +02:00
Matthew Honnibal
897dd0dd0b * Merge changes, and adjust Example to use memoryview 2015-06-28 11:36:11 +02:00
Matthew Honnibal
9282a8e72c * Prepare for new models to be plugged in by using Example class 2015-06-28 11:02:35 +02:00
Matthew Honnibal
75aeccc064 * Rejig parser interface to use new thinc.api.Example class, in prep of theano model. Comment out beam search 2015-06-28 11:02:34 +02:00
Matthew Honnibal
bbef71f213 * Fix min function in fill_context 2015-06-28 10:46:39 +02:00
Matthew Honnibal
142b6f9510 * Revert last changes 2015-06-28 10:44:28 +02:00
Matthew Honnibal
b06962f18b * Pad buffers in state 2015-06-28 10:36:14 +02:00
Matthew Honnibal
53be72387c * Hack at fill_context to investigate performance loss 2015-06-28 10:34:28 +02:00
Matthew Honnibal
71a4e876a9 * Fix parse features 2015-06-28 09:27:33 +02:00
Matthew Honnibal
5af500909c * Remove unused directve from parser.pyx 2015-06-28 06:20:21 +02:00
Matthew Honnibal
d5b4090705 * Add profile directive 2015-06-28 06:19:33 +02:00
Matthew Honnibal
2b5421e60c * Add profile directive 2015-06-28 06:07:04 +02:00
Matthew Honnibal
8b5de4a411 * Add word / tag / label sets, for use in neural net 2015-06-28 05:46:53 +02:00
Matthew Honnibal
ed40a8380e * Remove hard-coding of vector lengths 2015-06-27 04:18:47 +02:00
Matthew Honnibal
ebe630cc8d * Enable more features for NN 2015-06-27 04:17:29 +02:00
Matthew Honnibal
f8bb43475e * Bridge to Theano working. Very disorganised. Using thinc adb60aba966ed2 2015-06-27 02:39:18 +02:00
Matthew Honnibal
2fe98b8a9a * Prepare for new models to be plugged in by using Example class 2015-06-26 13:51:39 +02:00
Matthew Honnibal
6896455884 * Rejig parser interface to use new thinc.api.Example class, in prep of theano model. Comment out beam search 2015-06-26 06:25:36 +02:00
Matthew Honnibal
02b171ee67 * Bug fixes to edge calculation 2015-06-24 04:28:02 +02:00
Matthew Honnibal
7f9384f53c * Remove deprecated _state module 2015-06-23 17:28:24 +02:00
Matthew Honnibal
6dbe182491 * Fix merge conflicts 2015-06-23 17:28:00 +02:00
Matthew Honnibal
579735a095 * Remove import of _state module 2015-06-23 17:25:08 +02:00
Matthew Honnibal
88f55d136b * Remove deprecated _state module 2015-06-23 17:19:51 +02:00
Matthew Honnibal
9ab9dd2bf7 * Clean up unused orig_arc_eager and tree_arc_eager modules, which were only added for EMNLP experiments 2015-06-23 17:17:33 +02:00
Matthew Honnibal
7ebfe4b983 * Fixes to edge features 2015-06-23 16:32:54 +02:00
Matthew Honnibal
7b125f5a86 * Fixes to edge features 2015-06-23 16:31:01 +02:00
Matthew Honnibal
35c290bee4 * Fix edge features 2015-06-23 15:50:56 +02:00
Matthew Honnibal
221e2e485f * Assign 'ROOT' as label, not 'root' 2015-06-23 15:09:54 +02:00
Matthew Honnibal
a7bf7b0626 * Rename sent_start to sent_end, to reflect its new usage in the Break transition 2015-06-23 05:39:43 +02:00
Matthew Honnibal
ee3e56f27b * Fix bounds checking on entities 2015-06-23 04:35:08 +02:00
Matthew Honnibal
43ef5ddea5 * Ensure root albel is spelled ROOT, for backwards compatibility 2015-06-23 04:14:03 +02:00
Matthew Honnibal
065c2e1d2d * Add some bounds checking around state arrays 2015-06-23 04:13:09 +02:00
Matthew Honnibal
f01b3d043e * Add padding to arrays in stateclass. May be papering over a deeper bug. 2015-06-23 03:03:41 +02:00
Matthew Honnibal
69507bc729 * Re-enable Break transition in arc_eager.pyx 2015-06-23 00:03:30 +02:00
Matthew Honnibal
ab110be125 * Remove debugging in parser.pyx 2015-06-16 23:37:25 +02:00
Matthew Honnibal
9b13d11ab3 * Fix handling of entities in StateClass 2015-06-16 23:35:21 +02:00
Matthew Honnibal
c40a2c661c * Add tree_arc_eager 2015-06-15 08:23:24 +02:00
Matthew Honnibal
5da5cf7084 * Add some more features for S1/S0 2015-06-15 04:07:13 +02:00
Matthew Honnibal
8156a01bca * Fix root label for orig_arc_eager 2015-06-15 02:54:55 +02:00
Matthew Honnibal
21930ede15 * Switch toggle on USE_ROOT_ARC_SEGMENT 2015-06-15 02:54:32 +02:00
Matthew Honnibal
38a6afa484 * Make possibly dubious correction to the unshift oracle 2015-06-15 02:50:00 +02:00
Matthew Honnibal
f66228f253 * Add some more features, esp for labels 2015-06-14 21:18:02 +02:00
Matthew Honnibal
3da8e0f317 * Add orig_arc_eager 2015-06-14 20:31:44 +02:00
Matthew Honnibal
ea8a103007 * Fix import of TransitionSystem in parser.pyx 2015-06-14 19:01:26 +02:00
Matthew Honnibal
e0984ca139 * Fix valency features in StateClass 2015-06-14 17:50:26 +02:00
Matthew Honnibal
763cbd23d5 * Upd stateclass.print_state 2015-06-14 17:44:29 +02:00
Matthew Honnibal
bdd07bf000 * Fix Break oracle, but disable the Break transition for now, while we finalize the gold-standard experiments 2015-06-14 17:44:03 +02:00
Matthew Honnibal
399f15fbdf * Add flag to toggle handling of multi-root inputs without the Break transition. Clear up now unused best_valid stuff. 2015-06-14 00:28:37 +02:00
Matthew Honnibal
75289b4761 * Don't refuse to parse single token sentences, incase some transition system needs them, e.g. single word entity. Instead fix error in _init_state. 2015-06-13 22:55:55 +02:00
Matthew Honnibal
77d7e79c7e * Fix r/l and distance features. 2015-06-12 13:06:15 +02:00
Matthew Honnibal
15e177d7a1 * Fixes to unshift/fast-forward strategy. Getting 91.55 greedy on NW dev, gold preproc 2015-06-12 01:50:23 +02:00
Matthew Honnibal
afd77a529b * Prepare for break transition, with fast-forwarding. 86.5 on 1k nw gold preproc 2015-06-10 14:08:30 +02:00
Matthew Honnibal
495f528709 * Add support for sentence breaks in stateclass 2015-06-10 12:34:28 +02:00
Matthew Honnibal
b7b18c279d * Fix Reduce oracle. Getting 86.35 2015-06-10 11:33:39 +02:00
Matthew Honnibal
bb09b5d91a * Fix shifted bit vector in stateclass --- should reflect whether the word has been *unshifted*. 2015-06-10 11:33:09 +02:00
Matthew Honnibal
aa9625f688 * Do non-monotonic Unshift. Every word can be shifted at most 1 time. When the Reduce move is used, if S0 has no head, we put the word back on the buffer. Gets 86.4 on nw 1k with gold pre-proc. Break transition not yet implemented for this. 2015-06-10 10:15:56 +02:00
Matthew Honnibal
7bf6b7de3e * Add unshift action to StateClass, and track which moves have been shifted 2015-06-10 10:13:03 +02:00
Matthew Honnibal
f7c8069e65 * Fix bug in distance feature 2015-06-10 10:12:17 +02:00
Matthew Honnibal
abd07c067a * Inline B and S methods on stateclass 2015-06-10 07:22:33 +02:00
Matthew Honnibal
e2f9a80713 * Remove old _state imports 2015-06-10 07:09:17 +02:00
Matthew Honnibal
e9aaecc619 * Remove from_struct method from StateClass 2015-06-10 06:58:27 +02:00
Matthew Honnibal
18cc326dc0 * Bug fixes to ner.pyx 2015-06-10 06:57:41 +02:00
Matthew Honnibal
e5570c9700 * Set nogil for oracle functions 2015-06-10 06:56:56 +02:00
Matthew Honnibal
4575e7a60f * Fix beam search with new StateClass 2015-06-10 06:33:39 +02:00
Matthew Honnibal
04b1cd9b8c * Greedy parsing working with new StateClass. Beam parsing broken 2015-06-10 04:20:23 +02:00
Matthew Honnibal
6a94b64eca * Remove State* from parser.pyx entirely, switching over to StateClass. Beam parsing still untested. 2015-06-10 02:03:38 +02:00
Matthew Honnibal
f14a1526aa * Remove version of fill_context that takes State* 2015-06-10 01:39:07 +02:00
Matthew Honnibal
d68c686ec1 * Move StateClass into interface of transition functions 2015-06-10 01:35:28 +02:00
Matthew Honnibal
4b98b3e9c8 * Cost functions now take StateClass argument, instead of State*. 2015-06-10 00:40:43 +02:00
Matthew Honnibal
e0cf61f591 * Move StateClass into the interface for is_valid 2015-06-09 23:23:28 +02:00
Matthew Honnibal
0895d454fb * Prepare to switch to using state class, instead of state struct 2015-06-09 21:20:14 +02:00
Matthew Honnibal
2b9629ed62 * Begin adding stateclass to ArcEager 2015-06-09 01:41:09 +02:00
Matthew Honnibal
ba10fd8af5 * Add StateClass, to replace/refactor the mess in _state 2015-06-09 01:39:54 +02:00
Matthew Honnibal
c7e3dfc1dc * Don't automatically push words when stack is empty, as it messes up beam parsing. Add hash method to beam state. 2015-06-08 14:49:04 +02:00
Matthew Honnibal
6e2564239d * Bug fixes to beam parser. Search still broken on non-gold sentences 2015-06-07 19:12:59 +02:00
Matthew Honnibal
731e5f1e46 * Add get() function in spacy/syntax/Config 2015-06-07 19:09:15 +02:00
Matthew Honnibal
8f142c1838 * Refactor transition system oracles, to split out move and label cost. Preparing to add Unshift move. Will exclude non-monotonic. 2015-06-07 03:21:29 +02:00
Matthew Honnibal
1fee7ade61 * Tweak to ner 2015-06-05 23:48:43 +02:00
Matthew Honnibal
33e70b167f * Remove dead code from ner.pyx 2015-06-05 17:12:47 +02:00
Matthew Honnibal
88ac5c6e98 * Send beam_width < 0 to greedy parser 2015-06-05 17:12:06 +02:00
Matthew Honnibal
0114e7600d * Fix NER oracle 2015-06-05 17:11:26 +02:00
Matthew Honnibal
6bf35cecc3 * Refactor transition system to use classes with staticmethods. 2015-06-05 02:27:17 +02:00
Matthew Honnibal
36a34d544b * Refactoring arc_eager, grouping oracle functions into transitions 2015-06-04 22:43:03 +02:00
Matthew Honnibal
4433396005 * Impove efficiency of dynamic oracle, making beam training faster 2015-06-04 21:15:14 +02:00
Matthew Honnibal
079dad28a7 * Update for faster beam training 2015-06-04 19:32:32 +02:00
Matthew Honnibal
a2627b6102 * Fix bug in refactored init_transition 2015-06-03 06:01:26 +02:00
Matthew Honnibal
dd0867645d * Remove stray const from State header 2015-06-03 00:10:04 +02:00
Matthew Honnibal
6c47b10a6e * Make optimization to children_in_buffer: stop searching when we would cross a bracket. 2015-06-02 21:05:24 +02:00
Matthew Honnibal
a513ec500f * Have oracle functions take a struct instead of a Python object 2015-06-02 20:01:06 +02:00
Matthew Honnibal
d1b55310a1 * Refactor _advance_beam function 2015-06-02 18:38:41 +02:00
Matthew Honnibal
0786d9b3c7 * Refactor TransitionSystem, adding set_valid method 2015-06-02 18:38:07 +02:00
Matthew Honnibal
a3964957f6 * Add profiling for _state.pyx 2015-06-02 18:36:27 +02:00
Matthew Honnibal
e822df0867 * Fix bugs in new greedy/beam parser 2015-06-02 02:01:33 +02:00
Matthew Honnibal
66dfa95847 * Revise greedy_parse/beam_parse ownership goof 2015-06-02 01:34:19 +02:00
Matthew Honnibal
75658b2ed3 * Remove use of new beam.loss property, to maintain compatibility with older versions of thinc for now. 2015-06-02 00:57:09 +02:00
Matthew Honnibal
7c29362d60 * Rename parser class in parser.pxd, now that beam parsing is supported 2015-06-02 00:53:49 +02:00
Matthew Honnibal
58d5ac0944 * Add beam search capabilities to Parser. Rename GreedyParser to Parser. 2015-06-02 00:28:02 +02:00
Matthew Honnibal
e09a08bd00 * Add copy_state function 2015-06-01 23:06:30 +02:00
Matthew Honnibal
c7876aa8b6 * Add get_valid method 2015-06-01 23:06:00 +02:00
Matthew Honnibal
5e99ff94c8 * Edits to arc eager oracle. Couldn't figure out how the non-monotonic lines made sense. They seem covered by children_in_stack 2015-05-31 15:14:37 +02:00
Matthew Honnibal
6c5632b71c * Roll back proposed change to Break transition while investigate effect 2015-05-31 06:49:52 +02:00
Matthew Honnibal
e77940565d * Add length cap to distance feature 2015-05-31 05:25:30 +02:00
Matthew Honnibal
fd596351ba * Fix valency features 2015-05-31 05:24:33 +02:00
Matthew Honnibal
76300bbb1b * Use updated JSON format, with sentences below paragraphs. Allows use of gold preprocessing flag. 2015-05-30 01:25:46 +02:00
Matthew Honnibal
8f31d3b864 * Relax constraint on Break transition for non-monotonic parsing. 2015-05-28 23:39:52 +02:00
Matthew Honnibal
4010b9b6d9 * Pass parameter for regularization in parser.pyx 2015-05-27 03:18:50 +02:00
Matthew Honnibal
fc75210941 * Move spacy.syntax.conll to spacy.gold 2015-05-24 21:35:02 +02:00
Matthew Honnibal
efe7a7d7d6 * Clean unused functions from spacy.syntax.conll 2015-05-24 20:06:46 +02:00