patch for adjustable heuristics from Python

2 files changed, 9 insertions, 125 deletions

diff --git a/src/runtime/c/utils/pgf-chunk.c b/src/runtime/c/utils/pgf-chunk.c
deleted file mode 100644
index 5f4b8972a..000000000
--- a/src/runtime/c/utils/pgf-chunk.c
+++ /dev/null
@@ -1,112 +0,0 @@
-// Don't give too much hope to this script. It is doing the wrong thing
-// but let's see how far we can get with it.
-
-#include <gu/variant.h>
-#include <gu/map.h>
-#include <gu/dump.h>
-#include <gu/log.h>
-#include <gu/enum.h>
-#include <gu/file.h>
-#include <pgf/pgf.h>
-#include <pgf/parser.h>
-#include <pgf/lexer.h>
-#include <pgf/literals.h>
-#include <pgf/linearizer.h>
-#include <pgf/expr.h>
-#include <pgf/edsl.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <locale.h>
-#include <time.h>
-
-int main(int argc, char* argv[]) {
-	// Set the character locale, so we can produce proper output.
-	setlocale(LC_CTYPE, "");
-
-	// Create the pool that is used to allocate everything
-	GuPool* pool = gu_new_pool();
-	int status = EXIT_SUCCESS;
-	if (argc != 4) {
-		fprintf(stderr, "usage: %s pgf cat from_lang\n", argv[0]);
-		status = EXIT_FAILURE;
-		goto fail;
-	}
-	char* filename = argv[1];
-
-	GuString cat = gu_str_string(argv[2], pool);
-
-	GuString from_lang = gu_str_string(argv[3], pool);
-	
-	// Create an exception frame that catches all errors.
-	GuExn* err = gu_new_exn(NULL, gu_kind(type), pool);
-
-	// Read the PGF grammar.
-	PgfPGF* pgf = pgf_read(filename, pool, err);
-
-	// If an error occured, it shows in the exception frame
-	if (!gu_ok(err)) {
-		fprintf(stderr, "Reading PGF failed\n");
-		status = EXIT_FAILURE;
-		goto fail;
-	}
-
-	pgf_load_meta_child_probs(pgf, "../../../treebanks/PennTreebank/ParseEngAbs3.probs", pool, err);
-	if (!gu_ok(err)) {
-		fprintf(stderr, "Loading meta child probs failed\n");
-		status = EXIT_FAILURE;
-		goto fail;
-	}
-
-	// Look up the source and destination concrete categories
-	PgfConcr* from_concr = pgf_get_language(pgf, from_lang);
-	if (!from_concr) {
-		fprintf(stderr, "Unknown language\n");
-		status = EXIT_FAILURE;
-		goto fail_concr;
-	}
-	
-	// Register a callback for the literal category Symbol
-	pgf_parser_add_literal(from_concr, gu_str_string("Symb", pool),
-	                       &pgf_nerc_literal_callback);
-
-	// We will keep the latest results in the 'ppool' and
-	// we will iterate over them by using 'result'.
-	GuPool* ppool = NULL;
-
-	// The interactive translation loop.
-	// XXX: This currently reads stdin directly, so it doesn't support
-	// encodings properly. TODO: use a locale reader for input
-	while (true) {
-		char buf[4096];
-		char* line = fgets(buf, sizeof(buf), stdin);
-		if (line == NULL) {
-			if (ferror(stdin)) {
-				fprintf(stderr, "Input error\n");
-				status = EXIT_FAILURE;
-			}
-			break;
-		} else if (strcmp(line, "") == 0) {
-			// End nicely on empty input
-			break;
-		}
-		
-		// We create a temporary pool for translating a single
-		// sentence, so our memory usage doesn't increase over time.
-		ppool = gu_new_pool();
-
-		GuReader *rdr =
-			gu_string_reader(gu_str_string(line, ppool), ppool);
-		PgfLexer *lexer =
-			pgf_new_simple_lexer(rdr, ppool);
-
-		pgf_print_chunks(from_concr, cat, lexer, ppool);
-		
-		// Free all resources allocated during parsing and linearization
-		gu_pool_free(ppool);
-	}
-fail_concr:
-fail:
-	gu_pool_free(pool);
-	return status;
-}
diff --git a/src/runtime/c/utils/pgf-parse.c b/src/runtime/c/utils/pgf-parse.c
index a05d7988b..ba1088890 100644
--- a/src/runtime/c/utils/pgf-parse.c
+++ b/src/runtime/c/utils/pgf-parse.c
@@ -25,8 +25,8 @@ int main(int argc, char* argv[]) {
   // Create the pool that is used to allocate everything
   GuPool* pool = gu_new_pool();
   int status = EXIT_SUCCESS;
-  if (argc != 4) {
-    fprintf(stderr, "usage: %s pgf-file start-cat cnc-lang\n", argv[0]);
+  if (argc < 4 || argc > 5) {
+    fprintf(stderr, "usage: %s pgf-file start-cat cnc-lang [heuristics]\n(0.0 <= heuristics < 1.0, default: 0.95)\n", argv[0]);
     status = EXIT_FAILURE;
     goto fail;
   }
@@ -34,6 +34,11 @@ int main(int argc, char* argv[]) {
   GuString cat = gu_str_string(argv[2], pool);
   GuString lang = gu_str_string(argv[3], pool);
 
+  double heuristics = 0.95;
+  if (argc == 5) {
+      heuristics = atof(argv[4]);
+  }
+
   // Create an exception frame that catches all errors.
   GuExn* err = gu_new_exn(NULL, gu_kind(type), pool);
 
@@ -65,7 +70,7 @@ int main(int argc, char* argv[]) {
   clock_t end = clock();
   double cpu_time_used = ((double) (end - start)) / CLOCKS_PER_SEC;
 
-  fprintf(stderr, "(%.0f ms) Ready to parse!\n", 1000.0 * cpu_time_used);
+  fprintf(stderr, "(%.0f ms) Ready to parse [heuristics=%.2f]!\n", 1000.0 * cpu_time_used, heuristics);
 
   // Create an output stream for stdout
   GuOut* out = gu_file_out(stdout, pool);
@@ -113,18 +118,9 @@ int main(int argc, char* argv[]) {
 
     clock_t start = clock();
 
-    // Begin parsing a sentence of the specified category
-    PgfParseState* state =
-      pgf_parser_init_state(concr, cat, 0, ppool, ppool);
-    if (state == NULL) {
-      fprintf(stderr, "Couldn't begin parsing\n");
-      status = EXIT_FAILURE;
-      break;
-    }
-
     GuReader *rdr = gu_string_reader(gu_str_string(line, ppool), ppool);
     PgfLexer *lexer = pgf_new_simple_lexer(rdr, ppool);
-    GuEnum* result = pgf_parse(concr, cat, lexer, ppool, ppool);
+    GuEnum* result = pgf_parse_with_heuristics(concr, cat, lexer, heuristics, ppool, ppool);
 
     PgfExprProb* ep = NULL;
     if (result != NULL)