Al eff7a7a27a [optimization] moving regularization methods to their own module

2017-04-03 00:16:30 -04:00

..

cmp

[geodisambig] C msgpack serialization dependency

2015-04-12 15:14:01 -04:00

geohash

[fix] Removing C++ checks from all but the main API functions

2015-08-07 17:15:39 -04:00

klib

[fix] warnings in klib/ksort.h

2015-10-26 18:50:22 -04:00

linenoise

[cli] Adding antirez's linenoise for command-line interfaces

2015-11-29 11:28:31 -05:00

log

[logging] Wrapping logging statements in a do while (0) so the compiler always at least sees debug code

2015-10-12 13:42:10 -05:00

murmur

[geodisambig] adding MurmurHash3, used by the Bloom filter

2015-03-11 17:47:57 -04:00

sparkey

[build] pulling in change from parser-data that allows user to pass CFLAGS

2016-12-21 14:39:27 -05:00

utf8proc

[strings] reverting to utf8proc v1.3.1, as 2.0 and above can chop off certain sequences

2017-01-01 20:03:23 -05:00

address_dictionary_builder.c

[expansion] using easier-to-access data structure for address dictionaries

2016-11-27 00:56:48 -08:00

address_dictionary.c

[merge] merging master

2017-01-13 19:58:49 -05:00

address_dictionary.h

[merge] merging master

2017-01-13 19:58:49 -05:00

address_expansion_data.c

[dictionaries] regenerating address_expansion_data.c from upstream changes

2017-01-01 14:26:54 -05:00

address_expansion_rule.h

[fix] Adding stdint.h include to most of the header files for portability

2015-08-07 02:43:44 -04:00

address_parser_cli.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

address_parser_io.c

[fix] handle multiple separators (like parens used in https://www.openstreetmap.org/node/244081449). Creates bad trie entries otherwise, which affect more than just that toponym

2017-03-18 06:09:52 -04:00

address_parser_io.h

[parser] header changes for the data set struct

2016-12-09 13:37:45 -05:00

address_parser_test.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

address_parser_train.c

[utils] removing default chunk size from address_parser_train

2017-04-02 23:26:51 -04:00

address_parser.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

address_parser.h

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

averaged_perceptron_tagger.c

[parser] moving feature printing to averaged perceptron tagger, taking advantage of trie prefix-sharing in feature incorporating previous tags

2017-03-06 20:32:50 -05:00

averaged_perceptron_tagger.h

[parser] moving feature printing to averaged perceptron tagger, taking advantage of trie prefix-sharing in feature incorporating previous tags

2017-03-06 20:32:50 -05:00

averaged_perceptron_trainer.c

[parser] for the min updates method to work, the feature that have not yet reached the min_updates threshold also need to be ignored when scoring, that way the model has to perform without those features, and should make more updates if they're relevant

2017-03-08 15:40:12 -05:00

averaged_perceptron_trainer.h

[parser] learning a sparser averaged perceptron model for the parser using the following method:

2017-03-06 22:28:33 -05:00

averaged_perceptron.c

[parser] using new zeros API in averaged_perceptron.c

2017-02-19 14:02:54 -08:00

averaged_perceptron.h

[parser] Averaged perceptron training

2015-09-10 10:26:24 -07:00

bench.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

bloom.c

[fix] calloc instead of malloc when performing initialization on structs that may fail halfway and need to clean up while partially initialized (calloc will set all the bytes to zero so the member pointers are NULL instead of garbage memory)

2017-01-13 18:30:04 -05:00

bloom.h

[fix] Adding stdint.h include to most of the header files for portability

2015-08-07 02:43:44 -04:00

cartesian_product.c

[utils] cartesian product iterator for grid search during model selection

2017-04-03 00:15:31 -04:00

cartesian_product.h

[utils] cartesian product iterator for grid search during model selection

2017-04-03 00:15:31 -04:00

chunked_shuffle

[utils] adding a chunked shuffle as the concatenated file sizes may get larger than memory

2016-11-21 14:04:34 -05:00

collections.h

[utils] hash_get is no longer a string-only function, can be used for generic hashtables

2017-04-02 23:28:17 -04:00

constants.h

[constants] max country code length

2015-09-19 01:39:58 -04:00

crf_context.c

[openaddresses] adding Chesterfield, SC

2017-03-19 16:10:29 -04:00

crf_context.h

[parser/crf] adding the beginnings of a linear-chain Conditional Random Field

2017-03-10 01:10:22 -05:00

crf_trainer_averaged_perceptron.c

[fix] handle case of T = 0 in Viterbi decoding

2017-03-12 22:55:52 -04:00

crf_trainer_averaged_perceptron.h

[parser/crf] adding an initial training algorithm for CRFs, the averaged

2017-03-10 01:28:31 -05:00

crf_trainer.c

[parser/crf] adding crf_trainer, which can be thought of as a "base class" as much as that's possible in C, for creating trainers for the CRF. It doesn't deal with the weights or their representation, just provides an interface for keeping track of string features and label names, and holds the crf_context

2017-03-10 01:25:20 -05:00

crf_trainer.h

[utils] adding a function for checking if files exists (yay C), or at least the closest agreed-upon method for it (may return false if the user doesn't have permissions, but that's ok for our purposes here)

2017-03-10 13:39:52 -05:00

crf.c

[fix] formatting for print features in CRF model

2017-03-11 01:18:35 -05:00

crf.h

[parser/crf] adding runtime CRF tagger, which can be loaded/used once trained. Currently only does Viterbi inference, can add top-N and/or sequence probabilities later

2017-03-10 02:06:45 -05:00

features.c

[utils] char_array_add_vjoined to stay consistent (add_* methods NUL termiante)

2017-01-09 16:10:07 -05:00

features.h

[features] No copy versions of feature counts functions

2016-01-05 16:39:50 -05:00

file_utils.c

[utils] adding a function for checking if files exists (yay C), or at least the closest agreed-upon method for it (may return false if the user doesn't have permissions, but that's ok for our purposes here)

2017-03-10 13:39:52 -05:00

file_utils.h

[utils] adding file_exists to header

2017-03-10 19:30:46 -05:00

float_utils.c

[utils] adding non-branching sign functions

2017-04-02 13:41:57 -04:00

float_utils.h

[utils] adding non-branching sign functions

2017-04-02 13:41:57 -04:00

ftrl.c

[optimization] implemented Google's FTRL-Proximal, adapted for the multiclass/multinomial case. It is L1 and L2 regularized, and should both encourage sparsity with the L1 penalty while being robust to collinearity of features due to the L2 penalty. Ref: https://research.google.com/pubs/archive/41159.pdf

2017-04-02 14:28:25 -04:00

ftrl.h

[optimization] implemented Google's FTRL-Proximal, adapted for the multiclass/multinomial case. It is L1 and L2 regularized, and should both encourage sparsity with the L1 penalty while being robust to collinearity of features due to the L2 penalty. Ref: https://research.google.com/pubs/archive/41159.pdf

2017-04-02 14:28:25 -04:00

gazetteer_data.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

gazetteers.h

[dictionaries] Making new component for near/nearby prepositions

2016-07-21 17:04:57 -04:00

geo_disambiguation.c

[geodisambig] Shaving a few hundred more megabytes off of the geodb by only adding a single geohash prefix and not indexing the neighbors (query can use its neighbors)

2015-10-11 00:45:26 -05:00

geo_disambiguation.h

[geodisambig] Shaving a few hundred more megabytes off of the geodb by only adding a single geohash prefix and not indexing the neighbors (query can use its neighbors)

2015-10-11 00:45:26 -05:00

geodb_builder.c

[merge] merging master

2017-01-13 19:58:49 -05:00

geodb.c

[fix] calloc instead of malloc when performing initialization on structs that may fail halfway and need to clean up while partially initialized (calloc will set all the bytes to zero so the member pointers are NULL instead of garbage memory)

2017-01-13 18:30:04 -05:00

geodb.h

[geodb] Additional filenames and struct members in geodb.h

2015-10-09 15:37:10 -04:00

geonames_fields.h

[geonames] Re-generating postal/geonames fields headers

2015-07-08 17:02:59 -04:00

geonames.c

[fix] Check the return of malloc() in geonames.c

2016-02-25 14:53:31 -06:00

geonames.h

[geodb] Adding separate bitset for geonames place types and using NFC normalization instead of NFD (requires retraining)

2016-07-21 17:04:57 -04:00

graph_builder.c

[parser] using a bipartite graph (indptr + indices) to represent postal code<=>admin relationships instead of a set of 64-bit ints. Requires |V(postal codes)| + |E| 32 bit ints instead of |E| 64 bit ints. Saves several hundred MB in file size and even more space in memory because of the hashtable overhead

2017-03-18 06:05:28 -04:00

graph_builder.h

[geodb] Additional filenames and struct members in geodb.h

2015-10-09 15:37:10 -04:00

graph.c

[parser] using a bipartite graph (indptr + indices) to represent postal code<=>admin relationships instead of a set of 64-bit ints. Requires |V(postal codes)| + |E| 32 bit ints instead of |E| 64 bit ints. Saves several hundred MB in file size and even more space in memory because of the hashtable overhead

2017-03-18 06:05:28 -04:00

graph.h

[parser] using a bipartite graph (indptr + indices) to represent postal code<=>admin relationships instead of a set of 64-bit ints. Requires |V(postal codes)| + |E| 32 bit ints instead of |E| 64 bit ints. Saves several hundred MB in file size and even more space in memory because of the hashtable overhead

2017-03-18 06:05:28 -04:00

json_encode.c

[api] Simple JSON encoding for strings, UTF-8 rather than Unicode

2015-12-17 12:25:05 -05:00

json_encode.h

[api] Simple JSON encoding for strings, UTF-8 rather than Unicode

2015-12-17 12:25:05 -05:00

language_classifier_cli.c

[language_classification] Making directory optional on language_classifier client/test program

2016-01-27 03:18:53 -05:00

language_classifier_io.c

[log] log the offending line if token count does not match in language_classifier_io

2017-04-02 23:47:07 -04:00

language_classifier_io.h

[language_classification] Automatic hyperparameter optimization using either the cross-validation set or two distinct subsets of the training set

2016-01-17 21:11:37 -05:00

language_classifier_test.c

[language_classification] Making directory optional on language_classifier client/test program

2016-01-27 03:18:53 -05:00

language_classifier_train.c

[language_classification] adding options to language_classifier_train for using SGD with {L2, L1} regularization or FTRL-Proximal using both.

2017-04-02 23:46:14 -04:00

language_classifier.c

[language_classification] Runtime language classifier can now use dense or sparse weights, with a different header signature for the sparse version (using old signature for the dense version, so backward-compatible)

2017-04-02 23:51:54 -04:00

language_classifier.h

[language_classification] Runtime language classifier can now use dense or sparse weights, with a different header signature for the sparse version (using old signature for the dense version, so backward-compatible)

2017-04-02 23:51:54 -04:00

language_features.c

[languages] adding replace_hyphens and split_alpha_from_numeric in language classifier input normalization

2017-04-02 23:32:24 -04:00

language_features.h

[language_classifier] Features for address languages classification, quadgrams for most languages, unigrams for ideographic characters, script for single-script languages like Thai, Hebrew, etc.

2016-01-09 03:42:57 -05:00

libpostal_config.h

[build] adding libpostal_setup_datadir, libpostal_setup_parser_datadir, libpostal_setup_language_classifier_datadir functions for configuring the datadir at runtime

2017-01-09 16:11:26 -05:00

libpostal_data

[data] 12 worker pool in data download instead of 10 to download the new parser in one shot

2017-03-31 15:52:17 -04:00

libpostal.c

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

libpostal.h

[api] doing this now since we're bumping a major version. Using a libpostal prefixes for all public header functions and definitions

2017-03-31 03:35:51 -04:00

logistic_regression_trainer.c

[classification] flexible logistic regression trainer that can handle either SGD (with either L1 or L2) or FTRL as optimiers

2017-04-02 14:30:14 -04:00

logistic_regression_trainer.h

[classification] flexible logistic regression trainer that can handle either SGD (with either L1 or L2) or FTRL as optimiers

2017-04-02 14:30:14 -04:00

logistic_regression.c

[classification] removing regularization update from gradient computation in logistic regression, as that's now handled by the optimizer

2017-04-02 14:32:14 -04:00

logistic_regression.h

[classification] removing regularization update from gradient computation in logistic regression, as that's now handled by the optimizer

2017-04-02 14:32:14 -04:00

logistic.c

[math] using new matrix methods in softmax

2017-04-02 23:29:52 -04:00

logistic.h

[math] Generic dense matrix implementation using BLAS calls for matrix-matrix multiplication if available

2016-08-06 00:40:01 -04:00

main.c

[fix] expansion array destroy API in libpostal expand program

2017-04-02 23:55:04 -04:00

Makefile.am

[build] Makefile changes for new language_classifier_train

2017-04-02 23:55:31 -04:00

matrix.h

[optimization] implemented Google's FTRL-Proximal, adapted for the multiclass/multinomial case. It is L1 and L2 regularized, and should both encourage sparsity with the L1 penalty while being robust to collinearity of features due to the L2 penalty. Ref: https://research.google.com/pubs/archive/41159.pdf

2017-04-02 14:28:25 -04:00

minibatch.c

[fix] NULL check

2016-01-09 01:43:25 -05:00

minibatch.h

[features] Functions for dealing with minibatches

2016-01-08 00:48:11 -05:00

msgpack_utils.c

[utils] cstring_array_terminate, moving msgpack_utils to separate file

2015-07-25 18:41:02 -04:00

msgpack_utils.h

[fix] Adding stdint.h include to most of the header files for portability

2015-08-07 02:43:44 -04:00

ngrams.c

[ngrams] changing args to ngrams

2016-12-21 18:09:45 -05:00

ngrams.h

[ngrams] changing args to ngrams

2016-12-21 18:09:45 -05:00

normalize.c

[utils] string_replace_char does single byte/character replacement, new string_replace to do full string replacement, again using char_array for safety, string_replace_with_array function for memory reuse

2017-02-17 13:58:51 -05:00

normalize.h

[normalization] Adding NORMALIZE_STRING_SIMPLE_LATIN_ASCII option so parser can normalize punctuation and HTML entities, etc. without touching the alphanumeric parts of the original input

2016-08-21 19:45:32 -04:00

numex_data.c

[numex] regenerated numex data file

2016-11-30 15:58:55 -08:00

numex_rule.h

[numex] Separating rules from keys for Linux gcc compilation

2015-08-09 01:00:57 -04:00

numex_table_builder.c

[fix] numex_table builder program using new API (heap-allocated strings)

2016-03-02 16:28:28 -05:00

numex.c

[fix] calloc instead of malloc when performing initialization on structs that may fail halfway and need to clean up while partially initialized (calloc will set all the bytes to zero so the member pointers are NULL instead of garbage memory)

2017-01-13 18:30:04 -05:00

numex.h

[build] adding libpostal_setup_datadir, libpostal_setup_parser_datadir, libpostal_setup_language_classifier_datadir functions for configuring the datadir at runtime

2017-01-09 16:11:26 -05:00

postal_fields.h

[geonames] Re-generating postal/geonames fields headers

2015-07-08 17:02:59 -04:00

regularization.c

[optimization] moving regularization methods to their own module

2017-04-03 00:16:30 -04:00

regularization.h

[optimization] moving regularization methods to their own module

2017-04-03 00:16:30 -04:00

scanner.c

[tokenization] Re-generating scanner

2016-07-21 17:04:57 -04:00

scanner.h

[fix] stdint include

2015-09-20 20:10:47 -04:00

scanner.re

[tokenization] Hyphens, etc. between non-ASCII digits (e.g. Unicode full-width numbers) should be single tokens

2016-07-21 17:04:57 -04:00

shuffle.c

[utils] adding a function for checking if files exists (yay C), or at least the closest agreed-upon method for it (may return false if the user doesn't have permissions, but that's ok for our purposes here)

2017-03-10 13:39:52 -05:00

shuffle.h

[utils] adding default chunk size to shuffle.h

2017-04-02 13:51:45 -04:00

sparse_matrix_utils.c

[utils] sparse_matrix_add_unique_columns_alias, adds the actual column indices to hashtable/array and aliases those in the table from 1 to N (where N is the number of unique columns in this batch). This way it's compatible with smaller matrices of batch weights.

2017-04-02 13:48:46 -04:00

sparse_matrix_utils.h

[utils] sparse_matrix_add_unique_columns_alias, adds the actual column indices to hashtable/array and aliases those in the table from 1 to N (where N is the number of unique columns in this batch). This way it's compatible with smaller matrices of batch weights.

2017-04-02 13:48:46 -04:00

sparse_matrix.c

[merge] merging master

2017-01-13 19:58:49 -05:00

sparse_matrix.h

[math] Generic dense matrix implementation using BLAS calls for matrix-matrix multiplication if available

2016-08-06 00:40:01 -04:00

stochastic_gradient_descent.c

[optimization] new sgd_trainer struct to manage weights in stochastic gradient descent, allows L1 or L2 regularization, cumulative penalties instead of exponential decay, SGD using L1 regularization encouraged sparsity and can produce a sparse matrix after training rather than a dense one

2017-04-02 13:44:59 -04:00

stochastic_gradient_descent.h

[optimization] new sgd_trainer struct to manage weights in stochastic gradient descent, allows L1 or L2 regularization, cumulative penalties instead of exponential decay, SGD using L1 regularization encouraged sparsity and can produce a sparse matrix after training rather than a dense one

2017-04-02 13:44:59 -04:00

string_utils.c

[utils] adding string_is_digit function, similar to Python\'s (i.e. counts if it's in the Nd unicode category)

2017-03-15 13:04:39 -04:00

string_utils.h

[utils] adding string_is_digit function, similar to Python\'s (i.e. counts if it's in the Nd unicode category)

2017-03-15 13:04:39 -04:00

tagger.h

[parser] moving tagger function pointer definition to a separate header so it can be used for other models

2017-03-06 21:42:06 -05:00

token_types.h

[fix] is_numeric_token includes IDEOGRAPHIC_NUMBER

2017-02-12 15:11:56 -05:00

tokens.c

[fix] calls and NULL checks

2016-12-12 11:37:27 -05:00

tokens.h

[tokenization] tokenized string with source

2015-10-04 18:27:04 -04:00

transliterate.c

[merge] merging master

2017-01-13 19:58:49 -05:00

transliterate.h

[merge] merging master

2017-01-13 19:58:49 -05:00

transliteration_data.c

[transliteration] regenerate transliteration data

2017-03-17 18:28:41 -04:00

transliteration_rule.h

[transliteration] Using revisit strings instead of keeping a backtrack count so we don't have to later map logical characters to the actual string, removing any duplicate keys in the table builder so that if any rules happen to overlap within a step, the first will take precedence

2015-05-29 16:54:05 -04:00

transliteration_scripts_data.c

[transliteration] regenerated transliteration_scripts_data.c

2017-01-02 13:52:48 -05:00

transliteration_table_builder.c

[logging] converting most of the steps in building the transliteration table to use debug logging

2017-01-02 00:41:11 -05:00

trie_search.c

[phrases] set node data only when we're sure we have a correct match, otherwise the longer phrase may actually be matched

2017-03-17 03:40:29 -04:00

trie_search.h

[phrases] adding token_phrase_memberships to trie_search for reuse

2017-02-08 01:59:39 -05:00

trie_utils.c

[fix] trie_new_from_hash

2017-03-06 15:57:42 -05:00

trie_utils.h

[mv] Moving trie_new_from_hash to a module

2016-01-05 16:43:17 -05:00

trie.c

[fix] calloc instead of malloc when performing initialization on structs that may fail halfway and need to clean up while partially initialized (calloc will set all the bytes to zero so the member pointers are NULL instead of garbage memory)

2017-01-13 18:30:04 -05:00

trie.h

[fix] All file_read_uint64 calls that use stack variables read into a uint64_t not a size_t so as not to smash the stack under a 32-bit arch (issue #18)

2016-02-29 22:36:00 -05:00

unicode_script_types.h

[unicode] Regenerating unicode_script_data.c

2016-01-17 20:53:44 -05:00

unicode_scripts_data.c

[unicode] Regenerating unicode_script_data.c

2016-01-17 20:53:44 -05:00

unicode_scripts.c

[fix] declaring is_common_script function as static

2017-04-02 23:53:21 -04:00

unicode_scripts.h

[transliteration] Regenerating transliteration data files

2015-09-23 04:04:38 -04:00

vector_math.h

[math] adding mean, variance and standard deviation to generic vector functions

2017-04-02 23:29:15 -04:00

vector.h

[utils] new_fixed and resize_fixed in vector.h

2017-03-10 19:31:34 -05:00