ashikk
/
wikimedia-museumproject
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Make the parser a 2 pass system 

- first run parser with --download
- then run the exif script
- finally run parser without any flags
This commit is contained in:
Ashik K 2023-03-11 15:18:01 +01:00
parent 0727f60927
commit 73533fd0ae
1 changed files with 70 additions and 67 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

137
parser.cc
View File

@ -137,6 +137,7 @@ int main(int argc, char **argv)
if (subjv.IsArray()) { if (subjv.IsArray()) {
for (auto i = 0; i<subjv.GetArray().Size(); i++) { for (auto i = 0; i<subjv.GetArray().Size(); i++) {
subjects += subjv[i].GetString(); subjects += subjv[i].GetString();
subjects += " ";
} }
std::replace( description.begin(), description.end(), ',', ':'); std::replace( description.begin(), description.end(), ',', ':');
std::replace( description.begin(), description.end(), '\n', ' '); std::replace( description.begin(), description.end(), '\n', ' ');
@ -175,80 +176,82 @@ int main(int argc, char **argv)
} }
} }
if (is_cc_license) { if (!download_mode) {
char exif_file[128]; if (is_cc_license) {
snprintf(exif_file, sizeof exif_file, "%s-%s.jpeg.exif.json", article_id.c_str(), mediaid.c_str()); char exif_file[128];
// Open the file snprintf(exif_file, sizeof exif_file, "%s-%s.jpeg.exif.json", article_id.c_str(), mediaid.c_str());
FILE* exiffp = fopen(exif_file, "rb"); // Open the file
if (!exiffp) { FILE* exiffp = fopen(exif_file, "rb");
std::cerr << "Error: unable to open file" << std::string(exif_file) if (!exiffp) {
<< std::endl; std::cerr << "Error: unable to open file" << std::string(exif_file)
return -1; << std::endl;
} return -1;
}
char exifreadBuffer[4096]; char exifreadBuffer[4096];
rapidjson::FileReadStream exifis(exiffp, exifreadBuffer, sizeof(exifreadBuffer)); rapidjson::FileReadStream exifis(exiffp, exifreadBuffer, sizeof(exifreadBuffer));
rapidjson::Document exifdoc; rapidjson::Document exifdoc;
exifdoc.ParseStream(exifis); exifdoc.ParseStream(exifis);
// Check if the document is valid // Check if the document is valid
if (exifdoc.HasParseError()) { if (exifdoc.HasParseError()) {
std::cerr << "Error: failed to parse JSON document exif data" std::cerr << "Error: failed to parse JSON document exif data"
<< std::endl; << std::endl;
} }
fclose(exiffp); fclose(exiffp);
std::string exif_model, exif_iso, exif_focallength, exif_exposuretime, exif_aperture, exif_fnumber, exif_datetimeoriginal; std::string exif_model, exif_iso, exif_focallength, exif_exposuretime, exif_aperture, exif_fnumber, exif_datetimeoriginal;
if (exifdoc.IsArray()) { if (exifdoc.IsArray()) {
if (exifdoc[0].IsObject()) { if (exifdoc[0].IsObject()) {
if (exifdoc[0].GetObject().HasMember("Model")) { if (exifdoc[0].GetObject().HasMember("Model")) {
exif_model = exifdoc[0].GetObject()["Model"].GetString(); exif_model = exifdoc[0].GetObject()["Model"].GetString();
} }
if (exifdoc[0].GetObject().HasMember("ISO")) { if (exifdoc[0].GetObject().HasMember("ISO")) {
exif_iso = std::to_string(exifdoc[0].GetObject()["ISO"].GetInt()); exif_iso = std::to_string(exifdoc[0].GetObject()["ISO"].GetInt());
} }
if (exifdoc[0].GetObject().HasMember("FocalLength")) { if (exifdoc[0].GetObject().HasMember("FocalLength")) {
exif_focallength = exifdoc[0].GetObject()["FocalLength"].GetString(); exif_focallength = exifdoc[0].GetObject()["FocalLength"].GetString();
} }
if (exifdoc[0].GetObject().HasMember("ExposureTime")) { if (exifdoc[0].GetObject().HasMember("ExposureTime")) {
if (exifdoc[0].GetObject()["ExposureTime"].IsString()) if (exifdoc[0].GetObject()["ExposureTime"].IsString())
exif_exposuretime = exifdoc[0].GetObject()["ExposureTime"].GetString(); exif_exposuretime = exifdoc[0].GetObject()["ExposureTime"].GetString();
} }
if (exifdoc[0].GetObject().HasMember("ApertureValue")) { if (exifdoc[0].GetObject().HasMember("ApertureValue")) {
exif_aperture = std::to_string(exifdoc[0].GetObject()["ApertureValue"].GetDouble()); exif_aperture = std::to_string(exifdoc[0].GetObject()["ApertureValue"].GetDouble());
} }
if (exifdoc[0].GetObject().HasMember("DateTimeOriginal")) { if (exifdoc[0].GetObject().HasMember("DateTimeOriginal")) {
exif_datetimeoriginal = exifdoc[0].GetObject()["DateTimeOriginal"].GetString(); exif_datetimeoriginal = exifdoc[0].GetObject()["DateTimeOriginal"].GetString();
}
} }
} }
out_file1<<
/* article["artifact.defaultMediaIdentifier"].GetString()<<
", "<< article["artifact.defaultPictureIndex"].GetInt()<<
", "<< article["artifact.defaultPictureDimension"].GetString()<<
", "<<*/
article_id <<
", "<< title <<
", "<< yearstr(yearb) <<
", "<< yearstr(yeare) <<
", "<< description <<
", "<< imglink <<
", "<< article_id+"-"+ mediaid +".jpeg" <<
", "<< subjects <<
", "<< publishdate <<
", "<<"Länge leve Kosta! exhibition" <<
", "<<"Kulturparken Småland / Smålands museum" <<
", "<< exif_model <<
", "<< exif_iso <<
", "<< exif_focallength <<
", "<< exif_exposuretime <<
", "<< exif_aperture <<
", "<< exif_datetimeoriginal<<
", "<< license <<
std::endl;
} }
out_file1<<
/* article["artifact.defaultMediaIdentifier"].GetString()<<
", "<< article["artifact.defaultPictureIndex"].GetInt()<<
", "<< article["artifact.defaultPictureDimension"].GetString()<<
", "<<*/
article_id <<
", "<< title <<
", "<< yearstr(yearb) <<
", "<< yearstr(yeare) <<
", "<< description <<
", "<< imglink <<
", "<< article_id+"-"+ mediaid +".jpeg" <<
", "<< subjects <<
", "<< publishdate <<
", "<<"Länge leve Kosta! exhibition" <<
", "<<"Kulturparken Småland / Smålands museum" <<
", "<< exif_model <<
", "<< exif_iso <<
", "<< exif_focallength <<
", "<< exif_exposuretime <<
", "<< exif_aperture <<
", "<< exif_datetimeoriginal<<
", "<< license <<
std::endl;
} }
} }
} }