From: Martin Mares <mj@ucw.cz>
Date: Sun, 18 Nov 2007 13:52:52 +0000 (+0100)
Subject: Just backing up a part of judge-shuff, will continue later.
X-Git-Tag: python-dummy-working~277
X-Git-Url: http://mj.ucw.cz/gitweb/?a=commitdiff_plain;h=ed92a914fa9d0db91fc7efcf70552076c46a89b5;p=eval.git

Just backing up a part of judge-shuff, will continue later.
---

diff --git a/judge/Makefile b/judge/Makefile
index 448940d..c9ddb07 100644
--- a/judge/Makefile
+++ b/judge/Makefile
@@ -2,7 +2,7 @@ CC=gcc-4.1.1
 CFLAGS=-O2 -Wall -W -Wno-parentheses -Wstrict-prototypes -Wmissing-prototypes -Wundef -Wredundant-decls -Winline $(DEBUG) -std=gnu99
 CFLAGS+=-Wno-pointer-sign -Wdisabled-optimization -Wno-missing-field-initializers
 
-all: test-io test-tok judge-tok filter-cmt
+all: test-io test-tok judge-tok judge-shuff filter-cmt
 
 JLIB=utils.o io.o token.o
 
@@ -10,6 +10,7 @@ test-io: test-io.o $(JLIB)
 test-tok: test-tok.o $(JLIB)
 judge-tok: judge-tok.o $(JLIB)
 judge-tok: LDLIBS+=-lm
+judge-shuff: judge-shuff.o $(JLIB)
 filter-cmt: filter-cmt.o $(JLIB)
 
 clean::
diff --git a/judge/judge-shuff.c b/judge/judge-shuff.c
new file mode 100644
index 0000000..d5150e1
--- /dev/null
+++ b/judge/judge-shuff.c
@@ -0,0 +1,193 @@
+/*
+ *	A judge comparing shuffled sequences of tokens
+ *
+ *	(c) 2007 Martin Krulis <bobrik@matfyz.cz>
+ *	(c) 2007 Martin Mares <mj@ucw.cz>
+ *
+ *	FIXME: INCOMPLETE!
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <getopt.h>
+#include <math.h>
+
+#include "judge.h"
+
+static int ignore_nl, ignore_empty, ignore_case;
+static int shuffle_lines, shuffle_words;
+
+/*** Token buffer ***/
+
+struct tokpage {
+  struct tokpage *next;
+  char *pos, *end;
+  char buf[];
+};
+
+struct tokbuf {
+  // For writing:
+  struct tokpage *first_page, *last_page;
+  unsigned int num_tokens;
+  // For reading:
+  struct tokpage *this_page;
+  char *read_pos;
+};
+
+#define TOKBUF_PAGE 256
+
+static void init_tokbuf(struct tokbuf *tb)
+{
+  memset(tb, 0, sizeof(*tb));
+}
+
+static void add_token(struct tokbuf *tb, char *token, int l)
+{
+  l++;
+  struct tokpage *pg = tb->last_page;
+  if (!pg || pg->end - pg->pos < l)
+    {
+      int size = TOKBUF_PAGE - sizeof(struct tokbuf);
+      if (l > size/5)
+	size = l;
+      pg = xmalloc(sizeof(struct tokbuf) + size);
+      if (tb->last_page)
+	tb->last_page->next = pg;
+      else
+	tb->first_page = pg;
+      tb->last_page = pg;
+      pg->next = NULL;
+      pg->pos = pg->buf;
+      pg->end = pg->buf + size;
+    }
+  memcpy(pg->pos, token, l);
+  pg->pos += l;
+  tb->num_tokens++;
+}
+
+static char *get_next(struct tokbuf *tb)
+{
+  struct tokpage *pg = tb->this_page;
+  tb->read_pos += strlen(tb->read_pos) + 1;
+  if (tb->read_pos >= pg->end)
+    {
+      tb->this_page = pg = pg->next;
+      if (!pg)
+	return NULL;
+      tb->read_pos = pg->buf;
+    }
+  return tb->read_pos;
+}
+
+static char *get_first(struct tokbuf *tb)
+{
+  struct tokpage *pg = tb->first_page;
+  if (!pg)
+    return NULL;
+  tb->this_page = pg;
+  tb->read_pos = pg->buf;
+  return pg->buf;
+}
+
+/*** Reading of input ***/
+
+struct tok {
+  char *token;
+  unsigned int hash;
+};
+
+struct line {
+  struct tok *toks;
+  unsigned int num_toks;
+  unsigned int hash;
+};
+
+struct shouffle {
+  struct tokbuf tb;
+  struct tok *tok_array;
+  struct line *line_array;
+  unsigned int num_lines;
+};
+
+static void read_input(struct tokenizer *t, struct tokbuf *tb)
+{
+  char *tok;
+  int nl = 1;
+
+  init_tokbuf(tb);
+  while (tok = get_token(t))
+    {
+      if (tok[0])
+	nl = 0;
+      else
+	{
+	  if (nl && ignore_nl)
+	    continue;
+	  nl = 1;
+	}
+      add_token(tb, tok, t->toksize);
+    }
+  if (!nl)
+    add_token(tb, "", 0);
+}
+
+static void slurp(struct tokenizer *t, struct shouffle *s)
+{
+}
+
+/*** Main ***/
+
+static void usage(void)
+{
+  fprintf(stderr, "Usage: judge-shuff [<options>] <output> <correct>\n\
+\n\
+Options:\n\
+-n\t\tIgnore newlines and match the whole input as a single line\n\
+-e\t\tIgnore empty lines\n\
+-l\t\tShuffle lines (i.e., ignore their order)\n\
+-w\t\tShuffle words in each line\n\
+-i\t\tIgnore case\n\
+");
+  exit(2);
+}
+
+int main(int argc, char **argv)
+{
+  struct tokenizer t1, t2;
+  int opt;
+
+  while ((opt = getopt(argc, argv, "nelwi")) >= 0)
+    switch (opt)
+      {
+      case 'n':
+	ignore_nl++;
+	break;
+      case 'e':
+	ignore_empty++;
+	break;
+      case 'l':
+	shuffle_lines++;
+	break;
+      case 'w':
+	shuffle_words++;
+	break;
+      case 'i':
+	ignore_case++;
+	break;
+      default:
+	usage();
+      }
+  if (optind + 2 != argc)
+    usage();
+
+  tok_init(&t1, sopen_read(argv[optind]));
+  tok_init(&t2, sopen_read(argv[optind+1]));
+  if (!ignore_nl)
+    t1.flags = t2.flags = TF_REPORT_LINES;
+
+  struct tokbuf b1;
+  read_input(&t1, &b1);
+
+  return 0;
+}