Fix --summary=tag highlighting bugs, make pyzet use INDEX_SUMMARISE_TAG

diff --git a/src/pyzet/zetmodule.c b/src/pyzet/zetmodule.c
index 122c364..b8fabc9 100644 (file)
--- a/src/pyzet/zetmodule.c
+++ b/src/pyzet/zetmodule.c
@@ -1076,7 +1076,7 @@ static PyObject * Index_search(PyObject * self, PyObject * args,
     }
 #if 1 /* Nick */
  opts |= INDEX_SEARCH_SUMMARY_TYPE;
- opt.summary_type = INDEX_SUMMARISE_CAPITALISE;
+ opt.summary_type = INDEX_SUMMARISE_TAG;
 #endif
     if (!index_search(Index->idx, query, startdoc, len,
           result, &results, &total_results, &est, opts, &opt)) {
@@ -1428,7 +1428,7 @@ zet_search(PyObject *self, PyObject *args) {
     }
 #if 1 /* Nick */
  opts |= INDEX_SEARCH_SUMMARY_TYPE;
- opt.summary_type = INDEX_SUMMARISE_CAPITALISE;
+ opt.summary_type = INDEX_SUMMARISE_TAG;
 #endif
     if (!index_search(idx, query, startdoc, len, result, 
           &results, &total_results, &est, opts, &opt)) {

diff --git a/src/summarise.c b/src/summarise.c
index 8d5bc2c..03f3484 100644 (file)
--- a/src/summarise.c
+++ b/src/summarise.c
@@ -147,6 +147,13 @@ static int ensure_space(struct sentence **sent, unsigned int space) {
 /* internal function to finish up extraction of a sentence */
 static struct sentence *extract_finish(struct sentence *sent, struct persum *ps,
   enum index_summary_type type, int highlight) {
+#if 1 // Nick, do this first
+    /* remove superfluous whitespace from the end of the sentence */
+    while (sent->buf[sent->buflen - 1] == ' ') {
+        sent->buflen--;
+    }
+#endif
+
     if (highlight && type == INDEX_SUMMARISE_TAG) {
         /* need to close the tag */
         if (sent->buflen + str_len("</b>") >= sent->bufsize 
@@ -162,13 +169,14 @@ static struct sentence *extract_finish(struct sentence *sent, struct persum *ps,
               && (space < str_len("</b>"))) {
                 sent->buflen--;
             }
-
-            /* end highlighting */
-            str_cpy(sent->buf + sent->buflen, "</b>");
-            sent->buflen += str_len("</b>");
         }
+
+        /* end highlighting */
+        str_cpy(sent->buf + sent->buflen, "</b>");
+        sent->buflen += str_len("</b>");
     }
 
+#if 0 // Nick, don't want to remove markup by mistake, revisit this later
     /* trim overly-long sentence term-by-term */
     while (sent->buflen > ps->summary_len) {
         sent->buflen--;
@@ -181,8 +189,10 @@ static struct sentence *extract_finish(struct sentence *sent, struct persum *ps,
     while (sent->buf[sent->buflen - 1] == ' ') {
         sent->buflen--;
     }
+#endif
 
     sent->buf[sent->buflen] = '\0';
+ //printf("sent->buf %s\n", sent->buf);
     return sent;
 }