Archives/gnosygnu_xowa
1
0
Fork 0
mirror of https://github.com/gnosygnu/xowa.git synced 2024-10-27 20:34:16 +00:00
Code Issues Projects Releases Wiki Activity

Parser.Wikibase: Retrieve by label if pid is not present; also, return pid, not label [#354]

Browse Source
This commit is contained in:
gnosygnu 2019-04-07 12:27:50 -04:00
parent 623c7a129e
commit 09d9f93d20
5 changed files with 117 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

69
400_xowa/src/gplx/xowa/xtns/scribunto/libs/Scrib_lib_wikibase.java
View File

@ -163,11 +163,11 @@ public class Scrib_lib_wikibase implements Scrib_lib {
return rslt.Init_obj(wkr.Get_referenced_entity());
}
public boolean EntityExists(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, false);
Wdata_doc wdoc = Get_wdoc_or_null(args, core, "EntityExists", false);
return rslt.Init_obj(wdoc != null);
}
public boolean GetEntity(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, true);
Wdata_doc wdoc = Get_wdoc_or_null(args, core, "GetEntity", true);
if (wdoc == null) {
return rslt.Init_ary_empty();
// return rslt.Init_obj(Keyval_.Ary(Keyval_.new_("schemaVersion", 2))); // NOTE: was "rslt.Init_ary_empty()" PAGE:de.w:Critérium_International_2016 DATE:2017-12-30
@ -218,8 +218,46 @@ public function formatValues( $snaksSerialization ) {
throw Err_.new_unimplemented();
}
public boolean ResolvePropertyId(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, true); if (wdoc == null) return rslt.Init_ary_empty(); // NOTE: prop should be of form "P123"; do not add "P"; PAGE:no.w:Anne_Enger; DATE:2015-10-27
return rslt.Init_obj(wdoc.Label_list__get_or_fallback(core.Lang()));
byte[] rv = null;
byte[] pid = Get_xid_from_args(args);
if (pid != null) {
// if pid is "P####", look-up in db by title
byte b0 = pid[0];
if (b0 == Byte_ascii.Ltr_P || b0 == Byte_ascii.Ltr_p) {
// check if rest is numeric
boolean numeric = true;
for (int i = 1; i < pid.length; i++) {
if (!Byte_ascii.Is_num(pid[i])) {
numeric = false;
break;
}
}
if (numeric) {
byte[] key = pid;
if (b0 == Byte_ascii.Ltr_p) key[0] = Byte_ascii.Ltr_P; // uppercase key
Wdata_doc wdoc = entity_mgr.Get_by_xid_or_null(key); // NOTE: by_xid b/c Module passes just "p1" not "Property:P1"
if (wdoc != null) rv = key; // found wdoc; set rv
}
}
// pid is non-numeric or does not exist; look-up by label
if (rv == null) {
Wdata_wiki_mgr wdata_mgr = core.Wiki().Appe().Wiki_mgr().Wdata_mgr();
int pid_int = wdata_mgr.Pid_mgr.Get_pid_or_neg1(core.Wiki().Wdata_wiki_lang(), pid);
if (pid_int != gplx.xowa.xtns.wbases.core.Wbase_pid.Id_null) {
Bry_bfr tmp_bfr = Bry_bfr_.New();
tmp_bfr.Add_byte(Byte_ascii.Ltr_P);
tmp_bfr.Add_long_variable(pid_int);
rv = tmp_bfr.To_bry_and_clear();
}
}
}
if (rv == null) {
Wdata_wiki_mgr.Log_missing_qid(core.Ctx(), "ResolvePropertyId", pid);
return rslt.Init_null(); // NOTE: should be null, not empty; verified with "=mw.wikibase.resolvePropertyId('')" -> nil; DATE:2019-04-07
}
else
return rslt.Init_obj(rv);
}
public boolean GetPropertyOrder(Scrib_proc_args args, Scrib_proc_rslt rslt) {
throw Err_.new_("wbase", "getPropertyOrder not implemented", "url", core.Page().Url().To_str());
@ -228,7 +266,7 @@ public function formatValues( $snaksSerialization ) {
throw Err_.new_("wbase", "orderProperties not implemented", "url", core.Page().Url().To_str());
}
public boolean GetLabel(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, true);
Wdata_doc wdoc = Get_wdoc_or_null(args, core, "GetLabel", true);
if (wdoc == null)
return rslt.Init_ary_empty();
else
@ -240,12 +278,12 @@ public function formatValues( $snaksSerialization ) {
return rslt.Init_obj(wikibaseLanguageIndependentLuaBindings.getLabelByLanguage(prefixedEntityId, languageCode));
}
public boolean GetSiteLinkPageName(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, true); if (wdoc == null) return rslt.Init_ary_empty(); // NOTE: prop should be of form "P123"; do not add "P"; PAGE:no.w:Anne_Enger; DATE:2015-10-27
Wdata_doc wdoc = Get_wdoc_or_null(args, core, "GetSiteLinkPageName", true); if (wdoc == null) return rslt.Init_ary_empty(); // NOTE: prop should be of form "P123"; do not add "P"; PAGE:no.w:Anne_Enger; DATE:2015-10-27
Xow_domain_itm domain_itm = core.Wiki().Domain_itm();
return rslt.Init_obj(wdoc.Slink_list__get_or_fallback(domain_itm.Abrv_wm()));
}
public boolean GetDescription(Scrib_proc_args args, Scrib_proc_rslt rslt) {
Wdata_doc wdoc = Get_wdoc_or_null(args, core, true); if (wdoc == null) return rslt.Init_ary_empty();
Wdata_doc wdoc = Get_wdoc_or_null(args, core, "GetDescription", true); if (wdoc == null) return rslt.Init_ary_empty();
return rslt.Init_obj(wdoc.Descr_list__get_or_fallback(core.Lang()));
}
public boolean GetUserLang(Scrib_proc_args args, Scrib_proc_rslt rslt) {
@ -264,14 +302,21 @@ public function formatValues( $snaksSerialization ) {
public boolean IncrementExpensiveFunctionCount(Scrib_proc_args args, Scrib_proc_rslt rslt) {
return rslt.Init_obj(Keyval_.Ary_empty); // NOTE: for now, always return null (XOWA does not care about expensive parser functions)
}
public Wdata_doc Get_wdoc_or_null(Scrib_proc_args args, Scrib_core core, boolean logMissing) {
// get qid / pid from scrib_arg[0]; if none, return null;
byte[] xid_bry = args.Pull_bry(0); if (Bry_.Len_eq_0(xid_bry)) return null; // NOTE: some Modules do not pass in an argument; return early, else spurious warning "invalid qid for ttl" (since ttl is blank); EX:w:Module:Authority_control; DATE:2013-10-27
xid_bry = Bry_.Trim(xid_bry); // trim, b/c some pages will literally pass in "Property:P5\n"; PAGE:de.w:Mailand–Sanremo_2016 ISSUE#:363; DATE:2019-02-12
private byte[] Get_xid_from_args(Scrib_proc_args args) {
// get qid / pid from scrib_arg[0]
byte[] xid_bry = args.Pull_bry(0);
// NOTE: some Modules do not pass in an argument; return early, else spurious warning "invalid qid for ttl" (since ttl is blank); EX:w:Module:Authority_control; DATE:2013-10-27
return Bry_.Len_eq_0(xid_bry)
? null
: Bry_.Trim(xid_bry); // trim, b/c some pages will literally pass in "Property:P5\n"; PAGE:de.w:Mailand–Sanremo_2016 ISSUE#:363; DATE:2019-02-12
}
private Wdata_doc Get_wdoc_or_null(Scrib_proc_args args, Scrib_core core, String type, boolean logMissing) {
byte[] xid_bry = Get_xid_from_args(args);
if (xid_bry == null) return null;
// get wdoc
Wdata_doc wdoc = entity_mgr.Get_by_xid_or_null(xid_bry); // NOTE: by_xid b/c Module passes just "p1" not "Property:P1"
if (wdoc == null && logMissing) Wdata_wiki_mgr.Log_missing_qid(core.Ctx(), xid_bry);
if (wdoc == null && logMissing) Wdata_wiki_mgr.Log_missing_qid(core.Ctx(), type, xid_bry);
return wdoc;
}
}

2
400_xowa/src/gplx/xowa/xtns/scribunto/libs/Scrib_lib_wikibase_entity.java
View File

@ -63,7 +63,7 @@ public class Scrib_lib_wikibase_entity implements Scrib_lib {
// get wdoc
Wdata_doc wdoc = wdata_mgr.Doc_mgr.Get_by_loose_id_or_null(qid);
if (wdoc == null) {
Wdata_wiki_mgr.Log_missing_qid(core.Ctx(), qid);
Wdata_wiki_mgr.Log_missing_qid(core.Ctx(), "FormatPropertyValues", qid);
return rslt.Init_str_empty(); // NOTE: return empty String, not nil; PAGE:fr.s:Henri_Bergson; DATE:2014-08-13
}

4
400_xowa/src/gplx/xowa/xtns/scribunto/libs/wikibases/Basic__tst.java
View File

@ -140,10 +140,6 @@ public class Basic__tst {
// ( "1=" // not ""
// ));
// }
@Test public void ResolvePropertyId() {
wdata_fxt.Init__docs__add(wdata_fxt.Wdoc_bldr("Property:p2").Add_label("zh-hans", "prop_a").Xto_wdoc());
fxt.Test_scrib_proc_str(lib, Scrib_lib_wikibase.Invk_resolvePropertyId, Object_.Ary("p2"), "prop_a");
}
@Test public void RenderSnaks() {
Keyval[] args = Wbase_snak_utl_.Get_snaks_ary(wdata_fxt, wdata_fxt.Make_claim_monolingual(3, "en", "P3_en"), wdata_fxt.Make_claim_monolingual(3, "de", "P3_de"));
fxt.Test__proc__kvps__flat(lib, Scrib_lib_wikibase.Invk_renderSnaks, args, "P3_en, P3_de");

56
400_xowa/src/gplx/xowa/xtns/scribunto/libs/wikibases/ResolvePropertyId__tst.java Normal file
View File

@ -0,0 +1,56 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012-2017 gnosygnu@gmail.com
XOWA is licensed under the terms of the General Public License (GPL) Version 3,
or alternatively under the terms of the Apache License Version 2.0.
You may use XOWA according to either of these licenses as is most appropriate
for your project on a case-by-case basis.
The terms of each license can be found in the source code repository:
GPLv3 License: https://github.com/gnosygnu/xowa/blob/master/LICENSE-GPLv3.txt
Apache License: https://github.com/gnosygnu/xowa/blob/master/LICENSE-APACHE2.txt
*/
package gplx.xowa.xtns.scribunto.libs.wikibases; import gplx.*; import gplx.xowa.*; import gplx.xowa.xtns.*; import gplx.xowa.xtns.scribunto.*; import gplx.xowa.xtns.scribunto.libs.*;
import org.junit.*; import gplx.xowa.xtns.wbases.*; import gplx.xowa.xtns.wbases.core.*; import gplx.xowa.xtns.wbases.claims.*; import gplx.xowa.xtns.wbases.claims.itms.*;
import gplx.xowa.xtns.wbases.claims.enums.*;
public class ResolvePropertyId__tst {
private final ResolvePropertyId__fxt fxt = new ResolvePropertyId__fxt();
@Before public void init() {
fxt.Init();
}
@Test public void Basic() {
fxt.Init__doc("Property:P2");
fxt.Init__pid("de", "de0", 2);
fxt.Init__pid("en", "en0", 2);
fxt.Test__ResolvePropertyId("P2" , "P2"); // match exact
fxt.Test__ResolvePropertyId("p2" , "P2"); // match by lower-case
fxt.Test__ResolvePropertyId("P3" , null); // do not match if unknown doc
fxt.Test__ResolvePropertyId("de0", "P2"); // match by label
fxt.Test__ResolvePropertyId("en0", null); // do not match if label is not same lang as wiki
fxt.Test__ResolvePropertyId("fr0", null); // do not match if unknown label
}
}
class ResolvePropertyId__fxt {
private final Scrib_invoke_func_fxt fxt = new Scrib_invoke_func_fxt();
private final Wdata_wiki_mgr_fxt wdata_fxt = new Wdata_wiki_mgr_fxt();
private Scrib_lib lib;
public void Init() {
fxt.Clear_for_lib("de.wikipedia.org", "de");
lib = fxt.Core().Lib_wikibase().Init();
wdata_fxt.Init(fxt.Parser_fxt(), false);
}
public void Init__doc(String title) {
Wdata_doc_bldr bldr = wdata_fxt.Wdoc_bldr(title);
wdata_fxt.Init__docs__add(bldr.Xto_wdoc());
}
public void Init__pid(String lang_key, String pid_name, int pid) {
wdata_fxt.Init_pids_add(lang_key, pid_name, pid);
}
public void Test__ResolvePropertyId(String arg, String expd) {
fxt.Test_scrib_proc_str(lib, Scrib_lib_wikibase.Invk_resolvePropertyId, Object_.Ary(arg), expd);
}
}

5
400_xowa/src/gplx/xowa/xtns/wbases/Wdata_wiki_mgr.java
View File

@ -194,7 +194,8 @@ public class Wdata_wiki_mgr implements Gfo_evt_itm, Gfo_invk {
}
return false;
}
public static void Log_missing_qid(Xop_ctx ctx, byte[] qid) {
ctx.Wiki().Appe().Usr_dlg().Log_many("", "", "qid not found in wikidata; qid=~{0} page=~{1}", String_.new_u8(qid), String_.new_u8(ctx.Page().Url_bry_safe()));
public static void Log_missing_qid(Xop_ctx ctx, String type, byte[] id) {
if (id == null) id = Bry_.Empty;
ctx.Wiki().Appe().Usr_dlg().Log_many("", "", "Unknown id in wikidata; type=~{0} id=~{1} page=~{2}", type, id, ctx.Page().Url_bry_safe());
}
}