[merge] merging master

README.md

@@ -148,6 +148,20 @@ int main(int argc, char **argv) {
 }
 ```
 
+Parser labels
+-------------
+
+The address parser can use any string labels that are defined in the training data, but these are the default labels, based on the fields defined in [OpenCage's address-formatting library](https://github.com/OpenCageData/address-formatting):
+
+- **house**: venue name e.g. "Brooklyn Academy of Music", and building names e.g. "Empire State Building"
+- **house_number**: usually refers to the external (street-facing) building number. In some countries this may be a compount, hyphenated number which also includes an apartment number, or a block number (a la Japan), but libpostal will just call it the house_number for simplicity.
+- **road**: street name(s)
+- **suburb**: usually an unofficial neighborhood name like "Harlem", "South Bronx", or "Crown Heights"
+- **city_district**: these are usually boroughs or districts within a city that serve some official purpose e.g. "Brooklyn" or "Hackney" or "Bratislava IV"
+- **city**: any human settlement including cities, towns, villages, hamlets, localities, etc.
+- **state_district**: usually a second-level administrative division or county.
+- **state**: a first-level administrative division. Scotland, Northern Ireland, Wales, and England in the UK are mapped to "state" as well (convention used in OSM, GeoPlanet, etc.)
+- **country**: sovereign nations and their dependent territories, anything with an [ISO-3166 code](https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2).
 
 Examples of normalization
 -------------------------

configure.ac

@@ -1,7 +1,7 @@
 #                                               -*- Autoconf -*-
 # Process this file with autoconf to produce a configure script.
 
-AC_INIT([libpostal], [0.3])
+AC_INIT([libpostal], [0.3.3])
 
 AC_CONFIG_MACRO_DIRS([m4])
 
@@ -47,10 +47,8 @@ AC_TYPE_UINT8_T
 AC_CHECK_TYPES([ptrdiff_t])
 
 # Checks for library functions.
-AC_FUNC_MALLOC
 AC_FUNC_MMAP
-AC_FUNC_REALLOC
-AC_CHECK_FUNCS([getcwd gettimeofday memmove memset munmap regcomp setlocale sqrt strdup strndup])
+AC_CHECK_FUNCS([malloc realloc getcwd gettimeofday memmove memset munmap regcomp setlocale sqrt strdup strndup])
 
 AC_CONFIG_FILES([Makefile
                  libpostal.pc
@@ -88,4 +86,18 @@ AC_ARG_ENABLE([data-download],
 
 AM_CONDITIONAL([DOWNLOAD_DATA], [test "x$DOWNLOAD_DATA" = "xtrue"])
 
+AC_ARG_WITH(cflags-scanner-extra, [AS_HELP_STRING([--with-cflags-scanner-extra@<:@=VALUE@:>@], [Extra compilation options for scanner.c])],
+[
+ if test "x$withval" = "xno"; then
+    CFLAGS_SCANNER_EXTRA=""
+ else
+   CFLAGS_SCANNER_EXTRA="$withval"
+ fi
+],
+[ CFLAGS_SCANNER_EXTRA="" ]
+)
+
+AC_MSG_NOTICE([extra cflags for scanner.c: $CFLAGS_SCANNER_EXTRA])
+AC_SUBST(CFLAGS_SCANNER_EXTRA)
+
 AC_OUTPUT

src/Makefile.am

@@ -21,9 +21,11 @@ libpostal_la_CFLAGS = $(CFLAGS_O2)
 dist_bin_SCRIPTS = libpostal_data
 
 # Scanner can take a very long time to compile with higher optimization levels, so always use -O0, scanner is fast enough
+# On cross-compilation for ARM using gcc-4.7, there are "out of range" errors during compilation that can be fixed by adding
+# -marm option. For that, CFLAGS_SCANNER_EXTRA is provided that can be filled during configuration stage (see ./configure --help). 
 noinst_LTLIBRARIES = libscanner.la
 libscanner_la_SOURCES = scanner.c
-libscanner_la_CFLAGS = $(CFLAGS_O0)
+libscanner_la_CFLAGS = $(CFLAGS_O0) $(CFLAGS_SCANNER_EXTRA)
 
 noinst_PROGRAMS = libpostal bench build_address_dictionary build_geodb build_numex_table build_trans_table address_parser_train address_parser_test address_parser language_classifier_train language_classifier language_classifier_test
 libpostal_SOURCES = main.c json_encode.c

src/address_dictionary.c

@@ -288,7 +288,7 @@ phrase_t search_address_dictionaries_suffix(char *str, size_t len, char *lang) {
 bool address_dictionary_init(void) {
     if (address_dict != NULL) return false;
 
-    address_dict = malloc(sizeof(address_dictionary_t));
+    address_dict = calloc(1, sizeof(address_dictionary_t));
     if (address_dict == NULL) return false;
 
     address_dict->canonical = cstring_array_new();

src/address_dictionary.h

@@ -21,7 +21,8 @@
 
 #define ALL_LANGUAGES "all"
 
-#define DEFAULT_ADDRESS_EXPANSION_PATH LIBPOSTAL_DATA_DIR PATH_SEPARATOR "address_expansions" PATH_SEPARATOR "address_dictionary.dat"
+#define ADDRESS_DICTIONARY_DATA_FILE "address_dictionary.dat"
+#define DEFAULT_ADDRESS_EXPANSION_PATH LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_ADDRESS_EXPANSIONS_SUBDIR PATH_SEPARATOR ADDRESS_DICTIONARY_DATA_FILE
 
 #define NULL_CANONICAL_INDEX -1
 

src/address_parser.c

@@ -25,14 +25,13 @@ typedef enum {
     ADDRESS_PARSER_SUFFIX_PHRASE
 } address_parser_phrase_type_t;
 
-
 static parser_options_t PARSER_DEFAULT_OPTIONS = {
     .rare_word_threshold = DEFAULT_RARE_WORD_THRESHOLD,
     .print_features = false
 };
 
 address_parser_t *address_parser_new_options(parser_options_t options) {
-    address_parser_t *parser = malloc(sizeof(address_parser_t));
+    address_parser_t *parser = calloc(1, sizeof(address_parser_t));
     parser->options = options;
     return parser;
 }

src/averaged_perceptron.c

@@ -93,7 +93,7 @@ averaged_perceptron_t *averaged_perceptron_read(FILE *f) {
         return NULL;
     }
 
-    averaged_perceptron_t *perceptron = malloc(sizeof(averaged_perceptron_t));
+    averaged_perceptron_t *perceptron = calloc(1, sizeof(averaged_perceptron_t));
 
     if (!file_read_uint32(f, &perceptron->num_features) ||
         !file_read_uint32(f, &perceptron->num_classes) ||

src/averaged_perceptron_trainer.c

@@ -386,7 +386,7 @@ bool averaged_perceptron_trainer_train_example(averaged_perceptron_trainer_t *se
 }
 
 averaged_perceptron_trainer_t *averaged_perceptron_trainer_new(void) {
-    averaged_perceptron_trainer_t *self = malloc(sizeof(averaged_perceptron_trainer_t));
+    averaged_perceptron_trainer_t *self = calloc(1, sizeof(averaged_perceptron_trainer_t));
 
     if (self == NULL) return NULL;
 

src/bloom.c

@@ -56,7 +56,7 @@ int bloom_filter_add(bloom_filter_t *self, const char *key, size_t len) {
 }
 
 bloom_filter_t *bloom_filter_new(uint64_t capacity, double error) {
-    bloom_filter_t *bloom = malloc(sizeof(bloom_filter_t));
+    bloom_filter_t *bloom = calloc(1, sizeof(bloom_filter_t));
 
     if (bloom == NULL) {
         return NULL;

src/file_utils.c

@@ -36,6 +36,21 @@ bool is_relative_path(struct dirent *ent) {
     return strcmp(ent->d_name, ".") == 0 || strcmp(ent->d_name, "..") == 0;
 }
 
+char *path_vjoin(int n, va_list args) {
+    char_array *path = char_array_new();
+    if (path == NULL) return NULL;
+    char_array_add_vjoined(path, PATH_SEPARATOR, true, n, args);
+    return char_array_to_string(path);
+}
+
+char *path_join(int n, ...) {
+    va_list args;
+    va_start(args, n);
+    char *path = path_vjoin(n, args);
+    va_end(args);
+    return path;
+}
+
 inline uint64_t file_deserialize_uint64(unsigned char *buf) {
     return ((uint64_t)buf[0] << 56) | 
            ((uint64_t)buf[1] << 48) |

src/file_utils.h

@@ -9,6 +9,7 @@
 #include <sys/types.h>
 
 #include "libpostal_config.h"
+#include "string_utils.h"
 
 #ifdef HAVE_DIRENT_H
 #include <dirent.h>
@@ -55,6 +56,9 @@ char *file_getline(FILE * f);
 
 bool is_relative_path(struct dirent *ent);
 
+char *path_join(int n, ...);
+char *path_vjoin(int n, va_list args);
+
 uint64_t file_deserialize_uint64(unsigned char *buf);
 bool file_read_uint64(FILE *file, uint64_t *value);
 bool file_write_uint64(FILE *file, uint64_t value);

src/geodb.c

@@ -47,7 +47,7 @@ void geodb_destroy(geodb_t *self) {
 geodb_t *geodb_init(char *dir) {
     if (dir == NULL) return NULL;
 
-    geodb_t *gdb = malloc(sizeof(geodb_t));
+    geodb_t *gdb = calloc(1, sizeof(geodb_t));
 
     if (gdb == NULL) return NULL;
 

src/geodb_builder.c

@@ -338,7 +338,7 @@ void geodb_builder_destroy(geodb_builder_t *self) {
 }
 
 geodb_builder_t *geodb_builder_new(char *log_filename) {
-    geodb_builder_t *builder = malloc(sizeof(geodb_builder_t));
+    geodb_builder_t *builder = calloc(1, sizeof(geodb_builder_t));
 
     if (builder == NULL) return NULL;
 

src/graph.c

@@ -1,7 +1,7 @@
 #include "graph.h"
 
 graph_t *graph_new_dims(graph_type_t type, uint32_t m, uint32_t n, size_t nnz, bool fixed_rows) {
-    graph_t *graph = malloc(sizeof(graph_t));
+    graph_t *graph = calloc(1, sizeof(graph_t));
     graph->m = m;
     graph->fixed_rows = fixed_rows;
     graph->n = n;

src/language_classifier.c

@@ -35,7 +35,7 @@ void language_classifier_destroy(language_classifier_t *self) {
 }
 
 language_classifier_t *language_classifier_new(void) {
-    language_classifier_t *language_classifier = malloc(sizeof(language_classifier_t));
+    language_classifier_t *language_classifier = calloc(1, sizeof(language_classifier_t));
     return language_classifier;
 }
 

src/libpostal.c

@@ -1036,40 +1036,106 @@ address_parser_response_t *parse_address(char *address, address_parser_options_t
     return parsed;
 }
 
+bool libpostal_setup_datadir(char *datadir) {
+    char *transliteration_path = NULL;
+    char *numex_path = NULL;
+    char *address_dictionary_path = NULL;
+
+    if (datadir != NULL) {
+        transliteration_path = path_join(3, datadir, LIBPOSTAL_TRANSLITERATION_SUBDIR, TRANSLITERATION_DATA_FILE);
+        numex_path = path_join(3, datadir, LIBPOSTAL_NUMEX_SUBDIR, NUMEX_DATA_FILE);
+        address_dictionary_path = path_join(3, datadir, LIBPOSTAL_ADDRESS_EXPANSIONS_SUBDIR, ADDRESS_DICTIONARY_DATA_FILE);
+    }
+
+    if (!transliteration_module_setup(transliteration_path)) {
+        log_error("Error loading transliteration module, dir=%s\n", transliteration_path);
+        return false;
+    }
+
+    if (!numex_module_setup(numex_path)) {
+        log_error("Error loading numex module, dir=%s\n", numex_path);
+        return false;
+    }
+
+    if (!address_dictionary_module_setup(address_dictionary_path)) {
+        log_error("Error loading dictionary module, dir=%s\n", address_dictionary_path);
+        return false;
+    }
+
+    if (transliteration_path != NULL) {
+        free(transliteration_path);
+    }
+
+    if (numex_path != NULL) {
+        free(numex_path);
+    }
+
+    if (address_dictionary_path != NULL) {
+        free(address_dictionary_path);
+    }
+
+    return true;
+}
+
 bool libpostal_setup(void) {
-    if (!transliteration_module_setup(NULL)) {
-        log_error("Error loading transliteration module\n");
+    return libpostal_setup_datadir(NULL);
+}
+
+bool libpostal_setup_language_classifier_datadir(char *datadir) {
+    char *language_classifier_dir = NULL;
+
+    if (datadir != NULL) {
+        language_classifier_dir = path_join(2, datadir, LIBPOSTAL_LANGUAGE_CLASSIFIER_SUBDIR);
+    }
+
+    if (!language_classifier_module_setup(language_classifier_dir)) {
+        log_error("Error loading language classifier, dir=%s\n", language_classifier_dir);
         return false;
     }
 
-    if (!numex_module_setup(NULL)) {
-        log_error("Error loading numex module\n");
-        return false;
-    }
-
-    if (!address_dictionary_module_setup(NULL)) {
-        log_error("Error loading dictionary module\n");
-        return false;
+    if (language_classifier_dir != NULL) {
+        free(language_classifier_dir);
     }
 
     return true;
 }
 
 bool libpostal_setup_language_classifier(void) {
-    if (!language_classifier_module_setup(NULL)) {
-        log_error("Error loading language classifier\n");
+    return libpostal_setup_language_classifier_datadir(NULL);
+}
+
+bool libpostal_setup_parser_datadir(char *datadir) {
+    char *parser_dir = NULL;
+    char *geodb_dir = NULL;
+
+    if (datadir != NULL) {
+        parser_dir = path_join(2, datadir, LIBPOSTAL_ADDRESS_PARSER_SUBDIR);
+        geodb_dir = path_join(2, datadir, LIBPOSTAL_GEODB_SUBDIR);
+    }
+
+    if (!geodb_module_setup(geodb_dir)) {
+        log_error("Error loading geodb module, dir=%s\n", geodb_dir);
         return false;
     }
+
+    if (!address_parser_module_setup(parser_dir)) {
+        log_error("Error loading address parser module, dir=%s\n", parser_dir);
+        return false;
+    }
+
+    if (parser_dir != NULL) {
+        free(parser_dir);
+    }
+
+    if (geodb_dir != NULL) {
+        free(geodb_dir);
+    }
+
     return true;
 }
 
 bool libpostal_setup_parser(void) {
-    if (!address_parser_module_setup(NULL)) {
-        log_error("Error loading address parser module\n");
-        return false;
-    }
-
-    return true;
+    return libpostal_setup_parser_datadir(NULL);
 }
 
 void libpostal_teardown(void) {

src/libpostal.h

@@ -92,12 +92,15 @@ address_parser_response_t *parse_address(char *address, address_parser_options_t
 // Setup/teardown methods
 
 bool libpostal_setup(void);
+bool libpostal_setup_datadir(char *datadir);
 void libpostal_teardown(void);
 
 bool libpostal_setup_parser(void);
+bool libpostal_setup_parser_datadir(char *datadir);
 void libpostal_teardown_parser(void);
 
 bool libpostal_setup_language_classifier(void);
+bool libpostal_setup_language_classifier_datadir(char *datadir);
 void libpostal_teardown_language_classifier(void);
 
 #ifdef __cplusplus

src/libpostal_config.h

@@ -12,12 +12,20 @@
 #error LIBPOSTAL_DATA_DIR not defined!
 #endif
 
-#define LIBPOSTAL_ADDRESS_PARSER_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "address_parser"
-#define LIBPOSTAL_DICTIONARIES_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "dictionaries"
-#define LIBPOSTAL_GEONAMES_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "geonames"
-#define LIBPOSTAL_GEODB_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "geodb"
-#define LIBPOSTAL_LANGUAGE_CLASSIFIER_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "language_classifier"
-#define LIBPOSTAL_TRANSLITERATION_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR "transliteration"
+#define LIBPOSTAL_ADDRESS_PARSER_SUBDIR "address_parser"
+#define LIBPOSTAL_ADDRESS_PARSER_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_ADDRESS_PARSER_SUBDIR
+#define LIBPOSTAL_ADDRESS_EXPANSIONS_SUBDIR "address_expansions"
+#define LIBPOSTAL_ADDRESS_EXPANSIONS_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_ADDRESS_EXPANSIONS_SUBDIR
+#define LIBPOSTAL_GEONAMES_SUBDIR "geonames"
+#define LIBPOSTAL_GEONAMES_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_GEONAMES_SUBDIR
+#define LIBPOSTAL_GEODB_SUBDIR "geodb"
+#define LIBPOSTAL_GEODB_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_GEODB_SUBDIR
+#define LIBPOSTAL_LANGUAGE_CLASSIFIER_SUBDIR "language_classifier"
+#define LIBPOSTAL_LANGUAGE_CLASSIFIER_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_LANGUAGE_CLASSIFIER_SUBDIR
+#define LIBPOSTAL_NUMEX_SUBDIR "numex"
+#define LIBPOSTAL_NUMEX_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_NUMEX_SUBDIR
+#define LIBPOSTAL_TRANSLITERATION_SUBDIR "transliteration"
+#define LIBPOSTAL_TRANSLITERATION_DIR LIBPOSTAL_DATA_DIR PATH_SEPARATOR LIBPOSTAL_TRANSLITERATION_SUBDIR
 
 #define GEODB_BLOOM_FILTER_SIZE 100000000
 #define GEODB_BLOOM_FILTER_ERROR 0.001

src/numex.c

@@ -51,7 +51,7 @@ numex_table_t *numex_table_init(void) {
     numex_table_t *numex_table = get_numex_table();
 
     if (numex_table == NULL) {
-        numex_table = malloc(sizeof(numex_table_t));
+        numex_table = calloc(1, sizeof(numex_table_t));
 
         if (numex_table == NULL) return NULL;
 

src/numex.h

@@ -20,7 +20,8 @@
 #include "trie.h"
 #include "trie_search.h"
 
-#define DEFAULT_NUMEX_PATH LIBPOSTAL_DATA_DIR PATH_SEPARATOR "numex" PATH_SEPARATOR "numex.dat"
+#define NUMEX_DATA_FILE "numex.dat"
+#define DEFAULT_NUMEX_PATH LIBPOSTAL_DATA_DIR PATH_SEPARATOR "numex" PATH_SEPARATOR NUMEX_DATA_FILE
 
 #define LATIN_LANGUAGE_CODE "la"
 

src/sparse_matrix.c

@@ -2,7 +2,7 @@
 #include "klib/ksort.h"
 
 sparse_matrix_t *sparse_matrix_new_shape(size_t m, size_t n) {
-    sparse_matrix_t *matrix = malloc(sizeof(sparse_matrix_t));
+    sparse_matrix_t *matrix = calloc(1, sizeof(sparse_matrix_t));
     if (matrix == NULL) return NULL;
     matrix->m = m;
     matrix->n = n;

src/string_utils.c

@@ -782,18 +782,19 @@ cstring_array *cstring_array_new_size(size_t size) {
 
 cstring_array *cstring_array_from_char_array(char_array *str) {
     cstring_array *array = malloc(sizeof(cstring_array));
-    if (array == NULL) return NULL;
+    if (array == NULL || str == NULL) return NULL;
 
     array->str = str;
     array->indices = uint32_array_new_size(1);
     uint32_array_push(array->indices, 0);
     char *ptr = str->a;
-    uint32_t i = 0;
-    for (i = 0; i < str->n - 1; i++, ptr++) {
+    if (str->n > 0) {
+        for (uint32_t i = 0; i < str->n - 1; i++, ptr++) {
             if (*ptr == '\0') {
                 uint32_array_push(array->indices, i + 1);
             }
         }
+    }
     return array;
 }
 

src/string_utils.h

@@ -141,7 +141,6 @@ void char_array_cat_vprintf(char_array *array, char *format, va_list args);
 void char_array_cat_printf(char_array *array, char *format, ...);
 
 // Mainly for paths or delimited strings
-void char_array_append_vjoined(char_array *array, char *separator, bool strip_separator, int count, va_list args);
 void char_array_add_vjoined(char_array *array, char *separator, bool strip_separator, int count, va_list args);
 void char_array_add_joined(char_array *array, char *separator, bool strip_separator, int count, ...);
 void char_array_cat_joined(char_array *array, char *separator, bool strip_separator, int count, ...);

src/transliterate.c

@@ -1087,7 +1087,7 @@ transliteration_table_t *transliteration_table_init(void) {
     transliteration_table_t *trans_table = get_transliteration_table();
 
     if (trans_table == NULL) {
-        trans_table = malloc(sizeof(transliteration_table_t));
+        trans_table = calloc(1, sizeof(transliteration_table_t));
 
         trans_table->trie = trie_new();
         if (trans_table->trie == NULL) {

src/transliterate.h

@@ -17,7 +17,8 @@
 #define LATIN_ASCII_SIMPLE "latin-ascii-simple"
 #define HTML_ESCAPE "html-escape"
 
-#define DEFAULT_TRANSLITERATION_PATH LIBPOSTAL_TRANSLITERATION_DIR PATH_SEPARATOR "transliteration.dat"
+#define TRANSLITERATION_DATA_FILE "transliteration.dat"
+#define DEFAULT_TRANSLITERATION_PATH LIBPOSTAL_TRANSLITERATION_DIR PATH_SEPARATOR TRANSLITERATION_DATA_FILE
 
 #define MAX_TRANS_NAME_LEN 100
 

src/trie.c

@@ -32,7 +32,7 @@ Constructors
 */
 
 static trie_t *trie_new_empty(uint8_t *alphabet, uint32_t alphabet_size) {
-    trie_t *self = malloc(sizeof(trie_t));
+    trie_t *self = calloc(1, sizeof(trie_t));
     if (!self)
         goto exit_no_malloc;
 

src/vector.h

@@ -33,7 +33,7 @@ static inline void _aligned_free(void *p)
         name *array = malloc(sizeof(name));                                         \
         if (array == NULL) return NULL;                                             \
         array->n = array->m = 0;                                                    \
-        array->a = malloc(size * sizeof(type));                                     \
+        array->a = malloc((size > 0 ? size : 1) * sizeof(type));                    \
         if (array->a == NULL) return NULL;                                          \
         array->m = size;                                                            \
         return array;                                                               \