You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

283 lines
8.0 KiB

#define MICHLIB_NOSOURCE
#include "COPERNICUS.h"
#include "mirrorfuncs.h"
#include <libxml/parser.h>
#include <libxml/tree.h>
using michlib::GPL;
const MString COPERNICUSData::caturl = "https://stac.marine.copernicus.eu/metadata/catalog.stac.json";
std::pair<Json::Value, MString> COPERNICUSData::GetJSON(const MString& url)
{
Json::Reader reader;
Json::Value obj;
MString content;
auto [val, suc] = cache->Get(url);
if(suc)
content = std::move(val);
else
{
michlib::message(url + " not found in cache, downloading");
auto [out, res] = GetUrl(chandle, url);
if(res != CURLE_OK) return {obj, MString("Can't download JSON: ") + curlerr};
cache->Put(url, out, 3600);
content = std::move(out);
}
reader.parse(content.Buf(), content.Buf() + content.Len(), obj, false);
return {obj, ""};
}
MString COPERNICUSData::ReadURL(const Json::Value& cat, const MString& prod)
{
const auto& links = cat["links"];
if(links.type() != Json::arrayValue) return "";
for(Json::ArrayIndex i = 0; i < links.size(); i++)
{
const auto& titl = links[i]["title"];
const auto& href = links[i]["href"];
if(titl.type() == Json::stringValue && href.type() == Json::stringValue)
{
MString str(titl.asString().c_str());
if(str == prod) return MString(href.asString().c_str());
}
}
return "";
}
std::pair<std::vector<struct FileInfo>, MString> COPERNICUSData::ReadRemoteFileList(const MString& url)
{
LIBXML_TEST_VERSION
std::vector<struct FileInfo> out;
MString bucket, prefix;
// Split url on prefix and bucket
{
size_t pos = url.Len();
size_t count = 0;
for(size_t i = 0; i < url.Len(); i++)
{
if(url[i] == '/') count++;
if(count == 4)
{
pos = i;
break;
}
}
if(pos == url.Len()) return {out, "Can't parse url: " + url};
bucket = url.SubStr(1, pos);
prefix = url.SubStr(pos + 2, url.Len() - pos - 1);
}
MString cont;
bool next = true;
while(next)
{
MString url = bucket + "?list-type=2&prefix=" + prefix;
if(cont.Exist()) url += "&continuation-token=" + cont;
cont = "";
auto [data, res] = GetUrl(chandle, url);
if(res != CURLE_OK) return {out, MString("Can't download ") + url + ": " + curlerr};
xmlDocPtr doc = xmlReadMemory(data.Buf(), data.Len(), "data.xml", nullptr, 0);
if(doc == nullptr) return {out, MString("Can't download ") + url + ": XML parse error"};
auto cur = xmlDocGetRootElement(doc);
if(cur == nullptr)
{
xmlFreeDoc(doc);
return {out, MString("Can't download ") + url + ": empty XML"};
}
if(xmlStrEqual(cur->name, (const xmlChar*)"ListBucketResult") == 0)
{
xmlFreeDoc(doc);
return {out, MString("Can't download ") + url + ": unknown XML"};
}
for(const auto* n = cur->children; n; n = n->next)
{
if(xmlStrEqual(n->name, (const xmlChar*)"NextContinuationToken") == 1)
{
auto* content = xmlNodeGetContent(n);
cont = (char*)content;
xmlFree(content);
}
if(xmlStrEqual(n->name, (const xmlChar*)"Contents") == 1)
{
MString fname;
MDateTime mtime;
size_t size = 0;
for(const auto* c = n->children; c; c = c->next)
{
if(xmlStrEqual(c->name, (const xmlChar*)"Key") == 1)
{
auto* content = xmlNodeGetContent(c);
fname = (char*)content;
xmlFree(content);
}
if(xmlStrEqual(c->name, (const xmlChar*)"LastModified") == 1)
{
auto* content = xmlNodeGetContent(c);
mtime.FromString((char*)content);
xmlFree(content);
}
if(xmlStrEqual(c->name, (const xmlChar*)"Size") == 1)
{
auto* content = xmlNodeGetContent(c);
size = MString((char*)content).ToInteger<size_t>();
xmlFree(content);
}
}
out.emplace_back(bucket + "/" + fname, fname.SubStr(prefix.Len() + 2, fname.Len() - prefix.Len() - 1), mtime, size);
}
}
xmlFreeDoc(doc);
next = cont.Exist();
}
std::sort(out.begin(), out.end(), [](const struct FileInfo& a, const struct FileInfo& b) { return a.name < b.name; });
return {out, ""};
}
MString COPERNICUSData::Mirror(const CLArgs& args)
{
GPL.UsePrefix("COPERNICUS");
// Local directory
MString mirrorroot = GPL.ParameterSValue("MirrorTo", "");
if(!mirrorroot.Exist()) return "Local mirror directory not specified";
// Cache
cache.reset(CreateCache(GPL.ParameterSValue("Cache", "")));
if(!cache)
{
michlib::errmessage("Can't init cache");
cache.reset(new FakeCache);
}
curl_easy_setopt(chandle, CURLOPT_ERRORBUFFER, curlerr);
if(!args.contains("product")) return "Copernicus product not specified";
MString prod = args.at("product");
Json::Value product;
MString produrl;
// Get catalog
{
auto [cat, err] = GetJSON(caturl);
if(err.Exist()) return "Can't download catalog: " + err;
if(cat["title"].type() != Json::stringValue || cat["title"].asString() != "Copernicus Marine Data Store") return "Can't parse catalog";
catalog = std::move(cat);
}
// Get product
{
auto url = ReadURL(catalog, prod);
if(!url.Exist()) return "Url for product " + prod + " not found in catalog";
produrl = DirName(caturl) + "/" + url;
auto [pr, err] = GetJSON(produrl);
if(err.Exist()) return "Can't download product information from " + produrl + ": " + err;
product = std::move(pr);
}
std::vector<MString> dsets;
if(args.contains("dataset"))
dsets.push_back(args.at("dataset"));
else
{
const auto& links = product["links"];
if(links.type() != Json::arrayValue) return "Can't find information about datasets";
for(Json::ArrayIndex i = 0; i < links.size(); i++)
{
const auto& rel = links[i]["rel"];
const auto& titl = links[i]["title"];
if(rel.type() == Json::stringValue && titl.type() == Json::stringValue && rel.asString() == "item") dsets.push_back(titl.asString().c_str());
}
}
for(const auto& dset: dsets)
{
michlib::message("Mirroring " + dset);
auto url = ReadURL(product, dset);
if(!url.Exist()) return "Url for dataset " + dset + " not found in product description";
MString dseturl = DirName(produrl) + "/" + url;
auto [ds, err] = GetJSON(dseturl);
if(err.Exist()) return "Can't download dataset information from " + dseturl + ": " + err;
const auto& href = ds["assets"]["native"]["href"];
if(href.type() != Json::stringValue) return "Can't find data for dataset " + dset + " from product " + prod;
url = href.asString().c_str();
MString locroot = mirrorroot + "/" + prod + "/" + dset;
auto [lfiles, lerr] = ReadLocalFileList(locroot);
if(lerr.Exist()) return lerr;
auto [rfiles, rerr] = ReadRemoteFileList(url);
if(rerr.Exist()) return rerr;
std::vector<size_t> down, rem;
std::vector<std::pair<size_t, size_t>> upd;
{
size_t rpos = 0, lpos = 0;
while(rpos != rfiles.size() || lpos != lfiles.size())
{
if(rpos == rfiles.size())
while(lpos != lfiles.size()) rem.push_back(lpos++);
if(lpos == lfiles.size())
while(rpos != rfiles.size()) down.push_back(rpos++);
if(rpos == rfiles.size() || lpos == lfiles.size()) continue;
if(rfiles[rpos].name < lfiles[lpos].name)
down.push_back(rpos++);
else if(lfiles[lpos].name < rfiles[rpos].name)
rem.push_back(lpos++);
else
{
auto delta = rfiles[rpos].mtime.Epoch() - lfiles[lpos].mtime.Epoch();
if(delta < 0) delta = -delta;
if(delta > 0 || rfiles[rpos].size != lfiles[lpos].size) upd.emplace_back(rpos, lpos);
lpos++;
rpos++;
}
}
}
michlib::message(MString("New files: ") + down.size());
michlib::message(MString("Obsolete files: ") + rem.size());
michlib::message(MString("Modified files: ") + upd.size());
for(size_t i = 0; i < down.size(); i++)
{
size_t ri = down[i];
auto err = DownloadFile(chandle, rfiles[ri], locroot);
if(err.Exist()) return err;
}
for(size_t i = 0; i < rem.size(); i++)
{
size_t li = rem[i];
auto err = RemoveFile(lfiles[li]);
if(err.Exist()) return err;
}
for(size_t i = 0; i < upd.size(); i++)
{
size_t ri = upd[i].first;
size_t li = upd[i].second;
auto err = UpdateFile(chandle, rfiles[ri], lfiles[li], locroot);
if(err.Exist()) return err;
}
}
return "";
}