X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=src%2Frecord_conv.c;h=1595faf2ceb49a5601bc27cbf8b70bc2652b3c40;hp=c6ce964638364ac9f05468cc9d5b6755c2d07376;hb=80d0b86c41a5934878cb51657e92060909f320f4;hpb=95b69dabd857fa3342a71df2be123333b42c8da2 diff --git a/src/record_conv.c b/src/record_conv.c index c6ce964..1595faf 100644 --- a/src/record_conv.c +++ b/src/record_conv.c @@ -1,5 +1,5 @@ /* This file is part of the YAZ toolkit. - * Copyright (C) 1995-2012 Index Data + * Copyright (C) Index Data * See the file LICENSE for details. */ /** @@ -25,6 +25,8 @@ #include #include #include +#include +#include #if YAZ_HAVE_XSLT #include #include @@ -49,9 +51,6 @@ struct yaz_record_conv_struct { /** \brief path for opening files */ char *path; - - /** \brief handlers */ - struct yaz_record_conv_type *types; }; struct marc_info { @@ -60,6 +59,7 @@ struct marc_info { const char *output_charset; int input_format_mode; int output_format_mode; + const char *leader_spec; }; /** \brief tranformation info (rule info) */ @@ -86,44 +86,35 @@ static void yaz_record_conv_reset(yaz_record_conv_t p) p->rules_p = &p->rules; } -void yaz_record_conv_add_type(yaz_record_conv_t p, - struct yaz_record_conv_type *type) -{ - struct yaz_record_conv_type **tp = &p->types; - while (*tp) - tp = &(*tp)->next; - *tp = xmalloc(sizeof(*type)); - memcpy(*tp, type, sizeof(*type)); - (*tp)->next = 0; -} - void yaz_record_conv_destroy(yaz_record_conv_t p) { if (p) { - struct yaz_record_conv_type *t = p->types; - yaz_record_conv_reset(p); nmem_destroy(p->nmem); wrbuf_destroy(p->wr_error); - while (t) - { - struct yaz_record_conv_type *t_next = t->next; - xfree(t); - t = t_next; - } xfree(p->path); xfree(p); } } #if YAZ_HAVE_XSLT -static void *construct_xslt(yaz_record_conv_t p, const xmlNode *ptr, +struct xslt_info { + NMEM nmem; + xmlDocPtr xsp_doc; + const char **xsl_parms; +}; + +static void *construct_xslt(const xmlNode *ptr, const char *path, WRBUF wr_error) { struct _xmlAttr *attr; const char *stylesheet = 0; + struct xslt_info *info = 0; + NMEM nmem = 0; + int max_parms = 10; + int no_parms = 0; if (strcmp((const char *) ptr->name, "xslt")) return 0; @@ -140,17 +131,77 @@ static void *construct_xslt(yaz_record_conv_t p, const xmlNode *ptr, return 0; } } + nmem = nmem_create(); + info = nmem_malloc(nmem, sizeof(*info)); + info->nmem = nmem; + info->xsl_parms = nmem_malloc( + nmem, (2 * max_parms + 1) * sizeof(*info->xsl_parms)); + + for (ptr = ptr->children; ptr; ptr = ptr->next) + { + const char *name = 0; + const char *value = 0; + char *qvalue = 0; + if (ptr->type != XML_ELEMENT_NODE) + continue; + if (strcmp((const char *) ptr->name, "param")) + { + wrbuf_printf(wr_error, "Bad element '%s'" + "Expected param.", ptr->name); + nmem_destroy(nmem); + return 0; + } + for (attr = ptr->properties; attr; attr = attr->next) + { + if (!xmlStrcmp(attr->name, BAD_CAST "name") && + attr->children && attr->children->type == XML_TEXT_NODE) + name = (const char *) attr->children->content; + else if (!xmlStrcmp(attr->name, BAD_CAST "value") && + attr->children && attr->children->type == XML_TEXT_NODE) + value = (const char *) attr->children->content; + else + { + wrbuf_printf(wr_error, "Bad attribute '%s'" + "Expected name or value.", attr->name); + nmem_destroy(nmem); + return 0; + } + } + if (!name || !value) + { + wrbuf_printf(wr_error, "Missing attributes name or value"); + nmem_destroy(nmem); + return 0; + } + if (no_parms >= max_parms) + { + wrbuf_printf(wr_error, "Too many parameters given"); + nmem_destroy(nmem); + return 0; + } + + qvalue = nmem_malloc(nmem, strlen(value) + 3); + strcpy(qvalue, "\'"); + strcat(qvalue, value); + strcat(qvalue, "\'"); + + info->xsl_parms[2 * no_parms] = nmem_strdup(nmem, name); + info->xsl_parms[2 * no_parms + 1] = qvalue; + no_parms++; + } + + info->xsl_parms[2 * no_parms] = '\0'; + if (!stylesheet) { wrbuf_printf(wr_error, "Element : " "attribute 'stylesheet' expected"); - return 0; + nmem_destroy(nmem); } else { char fullpath[1024]; xsltStylesheetPtr xsp; - xmlDocPtr xsp_doc; if (!yaz_filepath_resolve(stylesheet, path, 0, fullpath)) { wrbuf_printf(wr_error, "Element :" @@ -158,21 +209,23 @@ static void *construct_xslt(yaz_record_conv_t p, const xmlNode *ptr, stylesheet, stylesheet); if (path) wrbuf_printf(wr_error, " with path '%s'", path); - + + nmem_destroy(nmem); return 0; } - xsp_doc = xmlParseFile(fullpath); - if (!xsp_doc) + info->xsp_doc = xmlParseFile(fullpath); + if (!info->xsp_doc) { wrbuf_printf(wr_error, "Element: :" " xml parse failed: %s", stylesheet, fullpath); if (path) wrbuf_printf(wr_error, " with path '%s'", path); + nmem_destroy(nmem); return 0; } /* need to copy this before passing it to the processor. It will be encapsulated in the xsp and destroyed by xsltFreeStylesheet */ - xsp = xsltParseStylesheetDoc(xmlCopyDoc(xsp_doc, 1)); + xsp = xsltParseStylesheetDoc(xmlCopyDoc(info->xsp_doc, 1)); if (!xsp) { wrbuf_printf(wr_error, "Element: :" @@ -181,27 +234,29 @@ static void *construct_xslt(yaz_record_conv_t p, const xmlNode *ptr, wrbuf_printf(wr_error, " with path '%s'", path); wrbuf_printf(wr_error, " (" #if YAZ_HAVE_EXSLT - + "EXSLT enabled" #else "EXSLT not supported" #endif ")"); - xmlFreeDoc(xsp_doc); - return 0; + xmlFreeDoc(info->xsp_doc); + nmem_destroy(info->nmem); } else { xsltFreeStylesheet(xsp); - return xsp_doc; + return info; } } return 0; } -static int convert_xslt(void *info, WRBUF record, WRBUF wr_error) +static int convert_xslt(void *vinfo, WRBUF record, WRBUF wr_error) { int ret = 0; + struct xslt_info *info = vinfo; + xmlDocPtr doc = xmlParseMemory(wrbuf_buf(record), wrbuf_len(record)); if (!doc) @@ -211,14 +266,14 @@ static int convert_xslt(void *info, WRBUF record, WRBUF wr_error) } else { - xmlDocPtr xsp_doc = xmlCopyDoc((xmlDocPtr) info, 1); + xmlDocPtr xsp_doc = xmlCopyDoc(info->xsp_doc, 1); xsltStylesheetPtr xsp = xsltParseStylesheetDoc(xsp_doc); - xmlDocPtr res = xsltApplyStylesheet(xsp, doc, 0); + xmlDocPtr res = xsltApplyStylesheet(xsp, doc, info->xsl_parms); if (res) { xmlChar *out_buf = 0; int out_len; - + #if HAVE_XSLTSAVERESULTTOSTRING xsltSaveResultToString(&out_buf, &out_len, res, xsp); #else @@ -234,7 +289,7 @@ static int convert_xslt(void *info, WRBUF record, WRBUF wr_error) { wrbuf_rewind(record); wrbuf_write(record, (const char *) out_buf, out_len); - + xmlFree(out_buf); } xmlFreeDoc(res); @@ -250,20 +305,143 @@ static int convert_xslt(void *info, WRBUF record, WRBUF wr_error) return ret; } -static void destroy_xslt(void *info) +static void destroy_xslt(void *vinfo) { + struct xslt_info *info = vinfo; + if (info) { - xmlDocPtr xsp_doc = info; - xmlFreeDoc(xsp_doc); + xmlFreeDoc(info->xsp_doc); + nmem_destroy(info->nmem); } } /* YAZ_HAVE_XSLT */ #endif +struct select_info { + NMEM nmem; + char *xpath_expr; +}; + +static void *construct_select(const xmlNode *ptr, + const char *path, WRBUF wr_error) +{ + if (strcmp((const char *) ptr->name, "select")) + return 0; + else + { + struct _xmlAttr *attr; + NMEM nmem = nmem_create(); + struct select_info *info = nmem_malloc(nmem, sizeof(*info)); + + info->nmem = nmem; + info->xpath_expr = 0; + for (attr = ptr->properties; attr; attr = attr->next) + { + if (!xmlStrcmp(attr->name, BAD_CAST "path") && + attr->children && attr->children->type == XML_TEXT_NODE) + info->xpath_expr = + nmem_strdup(nmem, (const char *) attr->children->content); + else + { + wrbuf_printf(wr_error, "Bad attribute '%s'" + "Expected xpath.", attr->name); + nmem_destroy(nmem); + return 0; + } + } + return info; + } +} + +static int convert_select(void *vinfo, WRBUF record, WRBUF wr_error) +{ + int ret = 0; + struct select_info *info = vinfo; + + xmlDocPtr doc = xmlParseMemory(wrbuf_buf(record), + wrbuf_len(record)); + if (!doc) + { + wrbuf_printf(wr_error, "xmlParseMemory failed"); + ret = -1; + } + else + { + xmlXPathContextPtr xpathCtx = xmlXPathNewContext(doc); + if (xpathCtx && info->xpath_expr) + { + xmlXPathObjectPtr xpathObj = + xmlXPathEvalExpression((const xmlChar *) info->xpath_expr, + xpathCtx); + if (xpathObj) + { + xmlNodeSetPtr nodes = xpathObj->nodesetval; + wrbuf_rewind(record); + if (nodes) + { + int i; + for (i = 0; i < nodes->nodeNr; i++) + { + xmlNode *ptr = nodes->nodeTab[i]; + fprintf(stderr, "xpath result %d type=%d\n", i, + ptr->type); + if (ptr->type == XML_ELEMENT_NODE) + ptr = ptr->children; + if (ptr->type == XML_TEXT_NODE) + for (; ptr; ptr = ptr->next) + wrbuf_puts(record, (const char *) ptr->content); + } + } + xmlXPathFreeObject(xpathObj); + } + xmlXPathFreeContext(xpathCtx); + } + xmlFreeDoc(doc); + } + return ret; +} + +static void destroy_select(void *info) +{ +} + + +static void *construct_solrmarc(const xmlNode *ptr, + const char *path, WRBUF wr_error) +{ + if (strcmp((const char *) ptr->name, "solrmarc")) + return 0; + return wr_error; /* any non-null ptr will do; we don't use it later*/ +} + +static int convert_solrmarc(void *info, WRBUF record, WRBUF wr_error) +{ + WRBUF w = wrbuf_alloc(); + const char *buf = wrbuf_buf(record); + size_t i, sz = wrbuf_len(record); + for (i = 0; i < sz; i++) + { + int ch; + if (buf[i] == '#' && i < sz - 3 && buf[i+3] == ';' + && atoi_n_check(buf+i+1, 2, &ch)) + i += 3; + else + ch = buf[i]; + wrbuf_putc(w, ch); + } + wrbuf_rewind(record); + wrbuf_write(record, wrbuf_buf(w), wrbuf_len(w)); + wrbuf_destroy(w); + return 0; +} + +static void destroy_solrmarc(void *info) +{ +} -static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, +static void *construct_marc(const xmlNode *ptr, const char *path, WRBUF wr_error) { NMEM nmem = nmem_create(); @@ -277,12 +455,12 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, nmem_destroy(nmem); return 0; } - info->nmem = nmem; info->input_charset = 0; info->output_charset = 0; info->input_format_mode = 0; info->output_format_mode = 0; + info->leader_spec = 0; for (attr = ptr->properties; attr; attr = attr->next) { @@ -298,11 +476,15 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, else if (!xmlStrcmp(attr->name, BAD_CAST "outputformat") && attr->children && attr->children->type == XML_TEXT_NODE) output_format = (const char *) attr->children->content; + else if (!xmlStrcmp(attr->name, BAD_CAST "leaderspec") && + attr->children && attr->children->type == XML_TEXT_NODE) + info->leader_spec = + nmem_strdup(info->nmem,(const char *) attr->children->content); else { wrbuf_printf(wr_error, "Element : expected attributes" "'inputformat', 'inputcharset', 'outputformat' or" - " 'outputcharset', got attribute '%s'", + " 'outputcharset', got attribute '%s'", attr->name); nmem_destroy(info->nmem); return 0; @@ -310,7 +492,7 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, } if (!input_format) { - wrbuf_printf(p->wr_error, "Element : " + wrbuf_printf(wr_error, "Element : " "attribute 'inputformat' required"); nmem_destroy(info->nmem); return 0; @@ -323,24 +505,28 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, { info->input_format_mode = YAZ_MARC_MARCXML; /** Libxml2 generates UTF-8 encoding by default . - So we convert from UTF-8 to outputcharset (if defined) + So we convert from UTF-8 to outputcharset (if defined) */ if (!info->input_charset && info->output_charset) info->input_charset = "utf-8"; } + else if (!strcmp(input_format, "json")) + { + info->input_format_mode = YAZ_MARC_JSON; + } else { wrbuf_printf(wr_error, "Element : " " Unsupported input format" - " defined by attribute value", + " defined by attribute value", input_format); nmem_destroy(info->nmem); return 0; } - + if (!output_format) { - wrbuf_printf(wr_error, + wrbuf_printf(wr_error, "Element : attribute 'outputformat' required"); nmem_destroy(info->nmem); return 0; @@ -371,11 +557,17 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, if (info->input_charset && !info->output_charset) info->output_charset = "utf-8"; } + else if (!strcmp(output_format, "json")) + { + info->output_format_mode = YAZ_MARC_JSON; + if (info->input_charset && !info->output_charset) + info->output_charset = "utf-8"; + } else { wrbuf_printf(wr_error, "Element : " " Unsupported output format" - " defined by attribute value", + " defined by attribute value", output_format); nmem_destroy(info->nmem); return 0; @@ -386,7 +578,7 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, info->input_charset); if (!cd) { - wrbuf_printf(p->wr_error, + wrbuf_printf(wr_error, "Element :" " Unsupported character set mapping" " defined by attribute values", @@ -396,43 +588,47 @@ static void *construct_marc(yaz_record_conv_t p, const xmlNode *ptr, } yaz_iconv_close(cd); } - else if (info->input_charset) + else if (!info->output_charset) { wrbuf_printf(wr_error, "Element : " "attribute 'outputcharset' missing"); nmem_destroy(info->nmem); return 0; } - else if (info->output_charset) + else if (!info->input_charset) { wrbuf_printf(wr_error, "Element : " "attribute 'inputcharset' missing"); nmem_destroy(info->nmem); return 0; } - info->input_charset = nmem_strdup(p->nmem, info->input_charset); - info->output_charset = nmem_strdup(p->nmem, info->output_charset); + info->input_charset = nmem_strdup(info->nmem, info->input_charset); + info->output_charset = nmem_strdup(info->nmem, info->output_charset); return info; } static int convert_marc(void *info, WRBUF record, WRBUF wr_error) { struct marc_info *mi = info; + const char *input_charset = mi->input_charset; int ret = 0; - - yaz_iconv_t cd = yaz_iconv_open(mi->output_charset, mi->input_charset); yaz_marc_t mt = yaz_marc_create(); - + yaz_marc_xml(mt, mi->output_format_mode); - - if (cd) - yaz_marc_iconv(mt, cd); + if (mi->leader_spec) + yaz_marc_leader_spec(mt, mi->leader_spec); + if (mi->input_format_mode == YAZ_MARC_ISO2709) { int sz = yaz_marc_read_iso2709(mt, wrbuf_buf(record), wrbuf_len(record)); if (sz > 0) + { + if (yaz_marc_check_marc21_coding(input_charset, wrbuf_buf(record), + wrbuf_len(record))) + input_charset = "utf-8"; ret = 0; + } else ret = -1; } @@ -461,13 +657,18 @@ static int convert_marc(void *info, WRBUF record, WRBUF wr_error) } if (ret == 0) { + yaz_iconv_t cd = yaz_iconv_open(mi->output_charset, input_charset); + + if (cd) + yaz_marc_iconv(mt, cd); + wrbuf_rewind(record); ret = yaz_marc_write_mode(mt, record); if (ret) wrbuf_printf(wr_error, "yaz_marc_write_mode failed"); + if (cd) + yaz_iconv_close(cd); } - if (cd) - yaz_iconv_close(cd); yaz_marc_destroy(mt); return ret; } @@ -475,12 +676,40 @@ static int convert_marc(void *info, WRBUF record, WRBUF wr_error) static void destroy_marc(void *info) { struct marc_info *mi = info; - + nmem_destroy(mi->nmem); } -int yaz_record_conv_configure(yaz_record_conv_t p, const xmlNode *ptr) +int yaz_record_conv_configure_t(yaz_record_conv_t p, const xmlNode *ptr, + struct yaz_record_conv_type *types) { + struct yaz_record_conv_type bt[4]; + size_t i = 0; + + /* register marc */ + bt[i].construct = construct_marc; + bt[i].convert = convert_marc; + bt[i++].destroy = destroy_marc; + + bt[i-1].next = &bt[i]; + bt[i].construct = construct_solrmarc; + bt[i].convert = convert_solrmarc; + bt[i++].destroy = destroy_solrmarc; + + bt[i-1].next = &bt[i]; + bt[i].construct = construct_select; + bt[i].convert = convert_select; + bt[i++].destroy = destroy_select; + +#if YAZ_HAVE_XSLT + /* register xslt */ + bt[i-1].next = &bt[i]; + bt[i].construct = construct_xslt; + bt[i].convert = convert_xslt; + bt[i++].destroy = destroy_xslt; +#endif + + bt[i-1].next = types; yaz_record_conv_reset(p); /* parsing element children */ @@ -491,30 +720,39 @@ int yaz_record_conv_configure(yaz_record_conv_t p, const xmlNode *ptr) void *info = 0; if (ptr->type != XML_ELEMENT_NODE) continue; - for (t = p->types; t; t = t->next) + for (t = &bt[0]; t; t = t->next) { wrbuf_rewind(p->wr_error); - info = t->construct(p, ptr, p->path, p->wr_error); - if (info) + info = t->construct(ptr, p->path, p->wr_error); + + if (info || wrbuf_len(p->wr_error)) break; + /* info== 0 and no error reported , ie not handled by it */ } if (!info) { - wrbuf_printf(p->wr_error, "Element : expected " - " or element, got <%s>" - , ptr->name); + if (wrbuf_len(p->wr_error) == 0) + wrbuf_printf(p->wr_error, "Element : expected " + " or element, got <%s>" + , ptr->name); return -1; } r = (struct yaz_record_conv_rule *) nmem_malloc(p->nmem, sizeof(*r)); r->next = 0; r->info = info; - r->type = t; + r->type = nmem_malloc(p->nmem, sizeof(*t)); + memcpy(r->type, t, sizeof(*t)); *p->rules_p = r; p->rules_p = &r->next; } return 0; } +int yaz_record_conv_configure(yaz_record_conv_t p, const xmlNode *ptr) +{ + return yaz_record_conv_configure_t(p, ptr, 0); +} + static int yaz_record_conv_record_rule(yaz_record_conv_t p, struct yaz_record_conv_rule *r, const char *input_record_buf, @@ -524,7 +762,7 @@ static int yaz_record_conv_record_rule(yaz_record_conv_t p, int ret = 0; WRBUF record = output_record; /* pointer transfer */ wrbuf_rewind(p->wr_error); - + wrbuf_write(record, input_record_buf, input_record_len); for (; ret == 0 && r; r = r->next) ret = r->type->convert(r->info, record, p->wr_error); @@ -538,25 +776,32 @@ int yaz_record_conv_opac_record(yaz_record_conv_t p, int ret = 0; struct yaz_record_conv_rule *r = p->rules; if (!r || r->type->construct != construct_marc) + { + wrbuf_puts(p->wr_error, "Expecting MARC rule as first rule for OPAC"); ret = -1; /* no marc rule so we can't do OPAC */ + } else { struct marc_info *mi = r->info; + const char *input_charset = mi->input_charset; + yaz_iconv_t cd; WRBUF res = wrbuf_alloc(); yaz_marc_t mt = yaz_marc_create(); - yaz_iconv_t cd = yaz_iconv_open(mi->output_charset, - mi->input_charset); - + + if (yaz_opac_check_marc21_coding(input_charset, input_record)) + input_charset = "utf-8"; + cd = yaz_iconv_open(mi->output_charset, input_charset); + wrbuf_rewind(p->wr_error); yaz_marc_xml(mt, mi->output_format_mode); - + yaz_marc_iconv(mt, cd); - + yaz_opac_decode_wrbuf(mt, input_record, res); if (ret != -1) { - ret = yaz_record_conv_record_rule(p, + ret = yaz_record_conv_record_rule(p, r->next, wrbuf_buf(res), wrbuf_len(res), output_record); @@ -599,31 +844,6 @@ yaz_record_conv_t yaz_record_conv_create() p->wr_error = wrbuf_alloc(); p->rules = 0; p->path = 0; - p->types = 0; - -#if YAZ_HAVE_EXSLT - exsltRegisterAll(); -#endif - { /* register marc */ - struct yaz_record_conv_type t; - - t.construct = construct_marc; - t.convert = convert_marc; - t.destroy = destroy_marc; - - yaz_record_conv_add_type(p, &t); - } -#if YAZ_HAVE_XSLT - { /* register xslt */ - struct yaz_record_conv_type t; - - t.construct = construct_xslt; - t.convert = convert_xslt; - t.destroy = destroy_xslt; - - yaz_record_conv_add_type(p, &t); - } -#endif return p; }