copied xslt.c filter over on alvis.c filter, to give alvis folks their own version
authorMarc Cromme <marc@indexdata.dk>
Fri, 19 Aug 2005 14:06:31 +0000 (14:06 +0000)
committerMarc Cromme <marc@indexdata.dk>
Fri, 19 Aug 2005 14:06:31 +0000 (14:06 +0000)
recctrl/alvis.c

index 4fbbbda..57e2ad2 100644 (file)
@@ -1,4 +1,4 @@
-/* $Id: alvis.c,v 1.3 2005-06-23 06:45:47 adam Exp $
+/* $Id: alvis.c,v 1.4 2005-08-19 14:06:31 marc Exp $
    Copyright (C) 1995-2005
    Index Data ApS
 
@@ -20,217 +20,635 @@ Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
 02111-1307, USA.
 */
 
-
 #include <stdio.h>
 #include <assert.h>
 #include <ctype.h>
 
+#include <yaz/diagbib1.h>
+#include <libxml/xmlversion.h>
+#include <libxml/parser.h>
+#include <libxml/tree.h>
+#include <libxml/xmlIO.h>
+#include <libxml/xmlreader.h>
+#include <libxslt/transform.h>
+
 #include <idzebra/util.h>
 #include <idzebra/recctrl.h>
 
+struct filter_schema {
+    const char *name;
+    const char *identifier;
+    const char *stylesheet;
+    struct filter_schema *next;
+    const char *default_schema;
+    const char *include_snippet;
+    xsltStylesheetPtr stylesheet_xsp;
+};
+
 struct filter_info {
-    char *sep;
+    xmlDocPtr doc;
+    char *fname;
+    const char *split_level;
+    const char *split_path;
+    ODR odr;
+    struct filter_schema *schemas;
+    xmlTextReaderPtr reader;
 };
 
-static void *filter_init (Res res, RecType recType)
+#define ZEBRA_SCHEMA_XSLT_NS "http://indexdata.dk/zebra/xslt/1"
+
+static const char *zebra_xslt_ns = ZEBRA_SCHEMA_XSLT_NS;
+
+static void set_param_xml(const char **params, const char *name,
+                         const char *value, ODR odr)
+{
+    while (*params)
+       params++;
+    params[0] = name;
+    params[1] = value;
+    params[2] = 0;
+}
+
+static void set_param_str(const char **params, const char *name,
+                         const char *value, ODR odr)
+{
+    char *quoted = odr_malloc(odr, 3 + strlen(value));
+    sprintf(quoted, "'%s'", value);
+    while (*params)
+       params++;
+    params[0] = name;
+    params[1] = quoted;
+    params[2] = 0;
+}
+
+static void set_param_int(const char **params, const char *name,
+                         zint value, ODR odr)
+{
+    char *quoted = odr_malloc(odr, 30); /* 25 digits enough for 2^64 */
+    while (*params)
+       params++;
+    sprintf(quoted, "'" ZINT_FORMAT "'", value);
+    params[0] = name;
+    params[1] = quoted;
+    params[2] = 0;
+}
+
+
+int zebra_xmlInputMatchCallback (char const *filename)
+{
+  /* yaz_log(YLOG_LOG, "match %s", filename); */
+    return 0;
+}
+
+
+void * zebra_xmlInputOpenCallback (char const *filename)
+{
+    return 0;
+}
+
+int zebra_xmlInputReadCallback (void * context, char * buffer, int len)
+{
+    return 0;
+}
+
+int zebra_xmlInputCloseCallback (void * context)
+{
+    return 0;
+}
+
+static void *filter_init_xslt(Res res, RecType recType)
 {
     struct filter_info *tinfo = (struct filter_info *) xmalloc(sizeof(*tinfo));
-    tinfo->sep = 0;
+    tinfo->reader = 0;
+    tinfo->fname = 0;
+    tinfo->split_level = 0;
+    tinfo->split_path = 0;
+    tinfo->odr = odr_createmem(ODR_ENCODE);
+    tinfo->doc = 0;
+    tinfo->schemas = 0;
+
+#if 0
+    xmlRegisterDefaultInputCallbacks();
+    xmlRegisterInputCallbacks(zebra_xmlInputMatchCallback,
+                             zebra_xmlInputOpenCallback,
+                             zebra_xmlInputReadCallback,
+                             zebra_xmlInputCloseCallback);
+#endif
     return tinfo;
 }
 
-static void filter_config(void *clientData, Res res, const char *args)
+static int attr_content(struct _xmlAttr *attr, const char *name,
+                       const char **dst_content)
 {
+    if (!strcmp(attr->name, name) && attr->children &&
+       attr->children->type == XML_TEXT_NODE)
+    {
+       *dst_content = attr->children->content;
+       return 1;
+    }
+    return 0;
+}
 
+static void destroy_schemas(struct filter_info *tinfo)
+{
+    struct filter_schema *schema = tinfo->schemas;
+    while (schema)
+    {
+       struct filter_schema *schema_next = schema->next;
+       if (schema->stylesheet_xsp)
+           xsltFreeStylesheet(schema->stylesheet_xsp);
+       xfree(schema);
+       schema = schema_next;
+    }
+    tinfo->schemas = 0;
+    xfree(tinfo->fname);
+    if (tinfo->doc)
+       xmlFreeDoc(tinfo->doc);    
+    tinfo->doc = 0;
 }
 
-static void filter_destroy (void *clientData)
+static ZEBRA_RES create_schemas(struct filter_info *tinfo, const char *fname)
+{
+    xmlNodePtr ptr;
+    tinfo->fname = xstrdup(fname);
+    tinfo->doc = xmlParseFile(tinfo->fname);
+    if (!tinfo->doc)
+       return ZEBRA_FAIL;
+    ptr = xmlDocGetRootElement(tinfo->doc);
+    if (!ptr || ptr->type != XML_ELEMENT_NODE ||
+       strcmp(ptr->name, "schemaInfo"))
+       return ZEBRA_FAIL;
+    for (ptr = ptr->children; ptr; ptr = ptr->next)
+    {
+       if (ptr->type != XML_ELEMENT_NODE)
+           continue;
+       if (!strcmp(ptr->name, "schema"))
+       {
+           struct _xmlAttr *attr;
+           struct filter_schema *schema = xmalloc(sizeof(*schema));
+           schema->name = 0;
+           schema->identifier = 0;
+           schema->stylesheet = 0;
+           schema->default_schema = 0;
+           schema->next = tinfo->schemas;
+           schema->stylesheet_xsp = 0;
+           schema->include_snippet = 0;
+           tinfo->schemas = schema;
+           for (attr = ptr->properties; attr; attr = attr->next)
+           {
+               attr_content(attr, "identifier", &schema->identifier);
+               attr_content(attr, "name", &schema->name);
+               attr_content(attr, "stylesheet", &schema->stylesheet);
+               attr_content(attr, "default", &schema->default_schema);
+               attr_content(attr, "snippet", &schema->include_snippet);
+           }
+           if (schema->stylesheet)
+               schema->stylesheet_xsp =
+                   xsltParseStylesheetFile(
+                       (const xmlChar*) schema->stylesheet);
+       }
+       else if (!strcmp(ptr->name, "split"))
+       {
+           struct _xmlAttr *attr;
+           for (attr = ptr->properties; attr; attr = attr->next)
+           {
+               attr_content(attr, "level", &tinfo->split_level);
+               attr_content(attr, "path", &tinfo->split_path);
+           }
+       }
+       else
+       {
+           yaz_log(YLOG_WARN, "Bad element %s in %s", ptr->name, fname);
+           return ZEBRA_FAIL;
+       }
+    }
+    return ZEBRA_OK;
+}
+
+static struct filter_schema *lookup_schema(struct filter_info *tinfo,
+                                          const char *est)
+{
+    struct filter_schema *schema;
+    for (schema = tinfo->schemas; schema; schema = schema->next)
+    {
+       if (est)
+       {
+           if (schema->identifier && !strcmp(schema->identifier, est))
+               return schema;
+           if (schema->name && !strcmp(schema->name, est))
+               return schema;
+       }
+       if (schema->default_schema)
+           return schema;
+    }
+    return 0;
+}
+
+static void filter_config(void *clientData, Res res, const char *args)
 {
     struct filter_info *tinfo = clientData;
-    xfree (tinfo->sep);
-    xfree (tinfo);
+    if (!args || !*args)
+       args = "xsltfilter.xml";
+    if (tinfo->fname && !strcmp(args, tinfo->fname))
+       return;
+    destroy_schemas(tinfo);
+    create_schemas(tinfo, args);
 }
 
-struct buf_info {
-    struct recExtractCtrl *p;
-    char *buf;
-    int offset;
-    int max;
-};
+static void filter_destroy(void *clientData)
+{
+    struct filter_info *tinfo = clientData;
+    destroy_schemas(tinfo);
+    if (tinfo->reader)
+       xmlFreeTextReader(tinfo->reader);
+    odr_destroy(tinfo->odr);
+    xfree(tinfo);
+}
 
-static struct buf_info *buf_open (struct recExtractCtrl *p)
+static int ioread_ex(void *context, char *buffer, int len)
 {
-    struct buf_info *fi = (struct buf_info *) xmalloc (sizeof(*fi));
+    struct recExtractCtrl *p = context;
+    return (*p->readf)(p->fh, buffer, len);
+}
 
-    fi->p = p;
-    fi->buf = (char *) xmalloc (4096);
-    fi->offset = 1;
-    fi->max = 1;
-    return fi;
+static int ioclose_ex(void *context)
+{
+    return 0;
 }
 
-static int buf_read (struct filter_info *tinfo, struct buf_info *fi, char *dst)
+static void index_cdata(struct filter_info *tinfo, struct recExtractCtrl *ctrl,
+                       xmlNodePtr ptr, RecWord *recWord)
 {
-    if (fi->offset >= fi->max)
+    for(; ptr; ptr = ptr->next)
     {
-        if (fi->max <= 0)
-            return 0;
-        fi->max = (*fi->p->readf)(fi->p->fh, fi->buf, 4096);
-        fi->offset = 0;
-        if (fi->max <= 0)
-            return 0;
+       index_cdata(tinfo, ctrl, ptr->children, recWord);
+       if (ptr->type != XML_TEXT_NODE)
+           continue;
+       recWord->term_buf = ptr->content;
+       recWord->term_len = strlen(ptr->content);
+       (*ctrl->tokenAdd)(recWord);
     }
-    *dst = fi->buf[(fi->offset)++];
-    if (tinfo->sep && *dst == *tinfo->sep)
+}
+
+static void index_node(struct filter_info *tinfo,  struct recExtractCtrl *ctrl,
+                      xmlNodePtr ptr, RecWord *recWord)
+{
+    for(; ptr; ptr = ptr->next)
     {
-       off_t off = (*fi->p->tellf)(fi->p->fh);
-       (*fi->p->endf)(fi->p->fh, off - (fi->max - fi->offset));
-       return 0;
+       index_node(tinfo, ctrl, ptr->children, recWord);
+       if (ptr->type != XML_ELEMENT_NODE || !ptr->ns ||
+           strcmp(ptr->ns->href, zebra_xslt_ns))
+           continue;
+       if (!strcmp(ptr->name, "index"))
+       {
+           const char *name_str = 0;
+           const char *type_str = 0;
+           const char *xpath_str = 0;
+           struct _xmlAttr *attr;
+           for (attr = ptr->properties; attr; attr = attr->next)
+           {
+               attr_content(attr, "name", &name_str);
+               attr_content(attr, "xpath", &xpath_str);
+               attr_content(attr, "type", &type_str);
+           }
+           if (name_str)
+           {
+               int prev_type = recWord->index_type; /* save default type */
+
+               if (type_str && *type_str)
+                   recWord->index_type = *type_str; /* type was given */
+               recWord->index_name = name_str;
+               index_cdata(tinfo, ctrl, ptr->children, recWord);
+
+               recWord->index_type = prev_type;     /* restore it again */
+           }
+       }
     }
-    return 1;
 }
 
-static void buf_close (struct buf_info *fi)
+static void index_record(struct filter_info *tinfo,struct recExtractCtrl *ctrl,
+                        xmlNodePtr ptr, RecWord *recWord)
 {
-    xfree (fi->buf);
-    xfree (fi);
+    if (ptr && ptr->type == XML_ELEMENT_NODE && ptr->ns &&
+       !strcmp(ptr->ns->href, zebra_xslt_ns)
+       && !strcmp(ptr->name, "record"))
+    {
+       const char *type_str = "update";
+       const char *id_str = 0;
+       const char *rank_str = 0;
+       struct _xmlAttr *attr;
+       for (attr = ptr->properties; attr; attr = attr->next)
+       {
+           attr_content(attr, "type", &type_str);
+           attr_content(attr, "id", &id_str);
+           attr_content(attr, "rank", &rank_str);
+       }
+       if (id_str)
+           sscanf(id_str, "%255s", ctrl->match_criteria);
+       if (rank_str)
+       {
+           ctrl->staticrank = atoi(rank_str);
+           yaz_log(YLOG_LOG, "rank=%d",ctrl->staticrank);
+       }
+       else
+           yaz_log(YLOG_LOG, "no rank");
+       
+       ptr = ptr->children;
+    }
+    index_node(tinfo, ctrl, ptr, recWord);
 }
-
-static int filter_extract (void *clientData, struct recExtractCtrl *p)
+    
+static int extract_doc(struct filter_info *tinfo, struct recExtractCtrl *p,
+                      xmlDocPtr doc)
 {
-    struct filter_info *tinfo = clientData;
-    char w[512];
     RecWord recWord;
-    int r;
-    struct buf_info *fi = buf_open (p);
+    const char *params[10];
+    xmlChar *buf_out;
+    int len_out;
+
+    struct filter_schema *schema = lookup_schema(tinfo, zebra_xslt_ns);
+
+    params[0] = 0;
+    set_param_str(params, "schema", zebra_xslt_ns, tinfo->odr);
 
-#if 0
-    yaz_log(YLOG_LOG, "filter_extract off=%ld",
-           (long) (*fi->p->tellf)(fi->p->fh));
-#endif
-    xfree(tinfo->sep);
-    tinfo->sep = 0;
     (*p->init)(p, &recWord);
-    do
-    {
-        int i = 0;
-            
-        r = buf_read (tinfo, fi, w);
-        while (r > 0 && i < 511 && w[i] != '\n' && w[i] != '\r')
-        {
-            i++;
-            r = buf_read (tinfo, fi, w + i); 
+
+    if (schema && schema->stylesheet_xsp)
+    {
+       xmlNodePtr root_ptr;
+       xmlDocPtr resDoc = 
+           xsltApplyStylesheet(schema->stylesheet_xsp,
+                               doc, params);
+       if (p->flagShowRecords)
+       {
+           xmlDocDumpMemory(resDoc, &buf_out, &len_out);
+           fwrite(buf_out, len_out, 1, stdout);
+           xmlFree(buf_out);
+       }
+       root_ptr = xmlDocGetRootElement(resDoc);
+       if (root_ptr)
+           index_record(tinfo, p, root_ptr, &recWord);
+       else
+       {
+           yaz_log(YLOG_WARN, "No root for index XML record."
+                   " split_level=%s stylesheet=%s",
+                   tinfo->split_level, schema->stylesheet);
        }
-        if (i)
-        {
-            recWord.term_buf = w;
-           recWord.term_len = i;
-            (*p->tokenAdd)(&recWord);
-        }
-    } while (r > 0);
-    buf_close (fi);
+       xmlFreeDoc(resDoc);
+    }
+    xmlDocDumpMemory(doc, &buf_out, &len_out);
+    if (p->flagShowRecords)
+       fwrite(buf_out, len_out, 1, stdout);
+    (*p->setStoreData)(p, buf_out, len_out);
+    xmlFree(buf_out);
+    
+    xmlFreeDoc(doc);
     return RECCTRL_EXTRACT_OK;
 }
 
-static int filter_retrieve (void *clientData, struct recRetrieveCtrl *p)
+static int extract_split(struct filter_info *tinfo, struct recExtractCtrl *p)
 {
-    int r, filter_ptr = 0;
-    static char *filter_buf = NULL;
-    static int filter_size = 0;
-    int make_header = 1;
-    int make_body = 1;
-    const char *elementSetName = NULL;
-    int no_lines = 0;
+    int ret;
+    int split_depth = 0;
+    if (p->first_record)
+    {
+       if (tinfo->reader)
+           xmlFreeTextReader(tinfo->reader);
+       tinfo->reader = xmlReaderForIO(ioread_ex, ioclose_ex,
+                                      p /* I/O handler */,
+                                      0 /* URL */, 
+                                      0 /* encoding */,
+                                      XML_PARSE_XINCLUDE);
+    }
+    if (!tinfo->reader)
+       return RECCTRL_EXTRACT_ERROR_GENERIC;
+
+    if (tinfo->split_level)
+       split_depth = atoi(tinfo->split_level);
+    ret = xmlTextReaderRead(tinfo->reader);
+    while (ret == 1) {
+       int type = xmlTextReaderNodeType(tinfo->reader);
+       int depth = xmlTextReaderDepth(tinfo->reader);
+       if (split_depth == 0 ||
+           (split_depth > 0 &&
+            type == XML_READER_TYPE_ELEMENT && split_depth == depth))
+       {
+           xmlNodePtr ptr = xmlTextReaderExpand(tinfo->reader);
+           xmlNodePtr ptr2 = xmlCopyNode(ptr, 1);
+           xmlDocPtr doc = xmlNewDoc("1.0");
 
-    if (p->comp && p->comp->which == Z_RecordComp_simple &&
-        p->comp->u.simple->which == Z_ElementSetNames_generic)
-        elementSetName = p->comp->u.simple->u.generic;
+           xmlDocSetRootElement(doc, ptr2);
 
-    if (elementSetName)
+           return extract_doc(tinfo, p, doc);      
+       }
+       ret = xmlTextReaderRead(tinfo->reader);
+    }
+    xmlFreeTextReader(tinfo->reader);
+    tinfo->reader = 0;
+    return RECCTRL_EXTRACT_EOF;
+}
+
+static int extract_full(struct filter_info *tinfo, struct recExtractCtrl *p)
+{
+    if (p->first_record) /* only one record per stream */
     {
-       /* don't make header for the R(aw) element set name */
-       if (!strcmp(elementSetName, "R"))
+       xmlDocPtr doc = xmlReadIO(ioread_ex, ioclose_ex, p /* I/O handler */,
+                                 0 /* URL */,
+                                 0 /* encoding */,
+                                 XML_PARSE_XINCLUDE);
+       if (!doc)
        {
-           make_header = 0;
-           make_body = 1;
+           return RECCTRL_EXTRACT_ERROR_GENERIC;
        }
-       /* only make header for the H(eader) element set name */
-       else if (!strcmp(elementSetName, "H"))
+       return extract_doc(tinfo, p, doc);
+    }
+    else
+       return RECCTRL_EXTRACT_EOF;
+}
+
+static int filter_extract(void *clientData, struct recExtractCtrl *p)
+{
+    struct filter_info *tinfo = clientData;
+
+    odr_reset(tinfo->odr);
+
+    if (tinfo->split_level == 0 && tinfo->split_path == 0)
+       return extract_full(tinfo, p);
+    else
+    {
+       return extract_split(tinfo, p);
+    }
+}
+
+static int ioread_ret(void *context, char *buffer, int len)
+{
+    struct recRetrieveCtrl *p = context;
+    return (*p->readf)(p->fh, buffer, len);
+}
+
+static int ioclose_ret(void *context)
+{
+    return 0;
+}
+
+
+static const char *snippet_doc(struct recRetrieveCtrl *p, int text_mode,
+                              int window_size)
+{
+    const char *xml_doc_str;
+    int ord = 0;
+    WRBUF wrbuf = wrbuf_alloc();
+    zebra_snippets *res = 
+       zebra_snippets_window(p->doc_snippet, p->hit_snippet, window_size);
+    zebra_snippet_word *w = zebra_snippets_list(res);
+
+    if (text_mode)
+       wrbuf_printf(wrbuf, "\'");
+    else
+       wrbuf_printf(wrbuf, "<snippet xmlns='%s'>\n", zebra_xslt_ns);
+    for (; w; w = w->next)
+    {
+       if (ord == 0)
+           ord = w->ord;
+       else if (ord != w->ord)
+
+           break;
+       if (text_mode)
+           wrbuf_printf(wrbuf, "%s%s%s ", 
+                        w->match ? "*" : "",
+                        w->term,
+                        w->match ? "*" : "");
+       else
        {
-           make_header = 1;
-           make_body = 0;
+           wrbuf_printf(wrbuf, " <term ord='%d' seqno='" ZINT_FORMAT "' %s>", 
+                        w->ord, w->seqno,
+                        (w->match ? "match='1'" : ""));
+           wrbuf_xmlputs(wrbuf, w->term);
+           wrbuf_printf(wrbuf, "</term>\n");
        }
     }
-    while (1)
-    {
-        if (filter_ptr + 4096 >= filter_size)
-        {
-            char *nb;
-
-            filter_size = 2*filter_size + 8192;
-            nb = (char *) xmalloc (filter_size);
-            if (filter_buf)
-            {
-                memcpy (nb, filter_buf, filter_ptr);
-                xfree (filter_buf);
-            }
-            filter_buf = nb;
-        }
-        if (make_header && filter_ptr == 0)
-        {
-            if (p->score >= 0)
-            {
-                sprintf (filter_buf, "Rank: %d\n", p->score);
-                filter_ptr = strlen(filter_buf);
-            }
-            sprintf (filter_buf + filter_ptr, "Local Number: " ZINT_FORMAT "\n",
-                    p->localno);
-            filter_ptr = strlen(filter_buf);
-           if (p->fname)
-           {
-               sprintf (filter_buf + filter_ptr, "Filename: %s\n", p->fname);
-               filter_ptr = strlen(filter_buf);
-           }
-           strcpy(filter_buf+filter_ptr++, "\n");
-        }
-       if (!make_body)
-           break;
-        r = (*p->readf)(p->fh, filter_buf + filter_ptr, 4096);
-        if (r <= 0)
-            break;
-        filter_ptr += r;
-    }
-    filter_buf[filter_ptr] = '\0';
-    if (elementSetName)
-    {
-        if (!strcmp (elementSetName, "B"))
-            no_lines = 4;
-        if (!strcmp (elementSetName, "M"))
-            no_lines = 20;
-    }
-    if (no_lines)
-    {
-        char *p = filter_buf;
-        int i = 0;
-
-        while (++i <= no_lines && (p = strchr (p, '\n')))
-            p++;
-        if (p)
-        {
-            p[1] = '\0';
-            filter_ptr = p-filter_buf;
-        }
-    }
-    p->output_format = VAL_SUTRS;
-    p->rec_buf = filter_buf;
-    p->rec_len = filter_ptr; 
+    if (text_mode)
+       wrbuf_printf(wrbuf, "\'");
+    else
+       wrbuf_printf(wrbuf, "</snippet>\n");
+
+    xml_doc_str = odr_strdup(p->odr, wrbuf_buf(wrbuf));
+
+    zebra_snippets_destroy(res);
+    wrbuf_free(wrbuf, 1);
+    return xml_doc_str;
+}
+
+static int filter_retrieve (void *clientData, struct recRetrieveCtrl *p)
+{
+    const char *esn = zebra_xslt_ns;
+    const char *params[10];
+    struct filter_info *tinfo = clientData;
+    xmlDocPtr resDoc;
+    xmlDocPtr doc;
+    struct filter_schema *schema;
+    int window_size = -1;
+
+    if (p->comp)
+    {
+       if (p->comp->which != Z_RecordComp_simple
+           || p->comp->u.simple->which != Z_ElementSetNames_generic)
+       {
+           p->diagnostic = YAZ_BIB1_PRESENT_COMP_SPEC_PARAMETER_UNSUPP;
+           return 0;
+       }
+       esn = p->comp->u.simple->u.generic;
+    }
+    schema = lookup_schema(tinfo, esn);
+    if (!schema)
+    {
+       p->diagnostic =
+           YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
+       return 0;
+    }
+
+    if (schema->include_snippet)
+       window_size = atoi(schema->include_snippet);
+
+    params[0] = 0;
+    set_param_str(params, "schema", esn, p->odr);
+    if (p->fname)
+       set_param_str(params, "filename", p->fname, p->odr);
+    if (p->score >= 0)
+       set_param_int(params, "score", p->score, p->odr);
+    set_param_int(params, "size", p->recordSize, p->odr);
+    set_param_int(params, "id", p->localno, p->odr);
+
+    if (window_size >= 0)
+       set_param_xml(params, "snippet", snippet_doc(p, 1, window_size),
+                     p->odr);
+    doc = xmlReadIO(ioread_ret, ioclose_ret, p /* I/O handler */,
+                   0 /* URL */,
+                   0 /* encoding */,
+                   XML_PARSE_XINCLUDE);
+    if (!doc)
+    {
+       p->diagnostic = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
+       return 0;
+    }
+
+    if (window_size >= 0)
+    {
+       xmlNodePtr node = xmlDocGetRootElement(doc);
+       const char *snippet_str = snippet_doc(p, 0, window_size);
+       xmlDocPtr snippet_doc = xmlParseMemory(snippet_str, strlen(snippet_str));
+       xmlAddChild(node, xmlDocGetRootElement(snippet_doc));
+    }
+    if (!schema->stylesheet_xsp)
+       resDoc = doc;
+    else
+    {
+       resDoc = xsltApplyStylesheet(schema->stylesheet_xsp,
+                                    doc, params);
+       xmlFreeDoc(doc);
+    }
+    if (!resDoc)
+    {
+       p->diagnostic = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
+    }
+    else if (p->input_format == VAL_NONE || p->input_format == VAL_TEXT_XML)
+    {
+       xmlChar *buf_out;
+       int len_out;
+       xmlDocDumpMemory(resDoc, &buf_out, &len_out);
+
+       p->output_format = VAL_TEXT_XML;
+       p->rec_len = len_out;
+       p->rec_buf = odr_malloc(p->odr, p->rec_len);
+       memcpy(p->rec_buf, buf_out, p->rec_len);
+       
+       xmlFree(buf_out);
+    }
+    else if (p->output_format == VAL_SUTRS)
+    {
+       xmlChar *buf_out;
+       int len_out;
+       xmlDocDumpMemory(resDoc, &buf_out, &len_out);
+
+       p->output_format = VAL_SUTRS;
+       p->rec_len = len_out;
+       p->rec_buf = odr_malloc(p->odr, p->rec_len);
+       memcpy(p->rec_buf, buf_out, p->rec_len);
+       
+       xmlFree(buf_out);
+    }
+    else
+    {
+       p->diagnostic = YAZ_BIB1_RECORD_SYNTAX_UNSUPP;
+    }
+    xmlFreeDoc(resDoc);
     return 0;
 }
 
-static struct recType filter_type = {
+static struct recType filter_type_alvis = {
     0,
     "alvis",
-    filter_init,
+    filter_init_xslt,
     filter_config,
     filter_destroy,
     filter_extract,
@@ -245,6 +663,6 @@ idzebra_filter
 #endif
 
 [] = {
-    &filter_type,
+    &filter_type_alvis,
     0,
 };