phantop
/
quatalog-scraper
mirror of https://github.com/quatalog/quatalog.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Uhh idk

This commit is contained in:
3eef8a28f26fb2bcc514e6f1938929a1f931762 2023-02-04 15:06:20 -05:00
parent 7ad5f2ff55
commit d6dc6b080c
1 changed files with 59 additions and 60 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

119
CourseOfferingsScraper.cpp
View File

@ -5,19 +5,13 @@
#include<filesystem>
#include<dist/jsoncpp.cpp>
struct course_sections_data_t {
int credMin = Json::Value::maxInt;
int credMax = 0;
int seatsTaken = 0;
int capacity = 0;
int remaining = 0;
};
namespace fs = std::filesystem;
void handle_term(const fs::directory_entry&,Json::Value&);
void handle_prefix(const Json::Value&,const std::string&,Json::Value&);
void handle_course(const Json::Value&,const std::string&,Json::Value&);
void handle_attribute(Json::Value&,const std::string&);
course_sections_data_t handle_sections(const Json::Value&);
void handle_attribute(const std::string&,Json::Value&);
void handle_attributes(const Json::Value&,Json::Value&);
void handle_sections(const Json::Value&,Json::Value&);
int main(const int argc,const char** argv) {
if(argc < 2) {
@ -69,88 +63,93 @@ void handle_term(const fs::directory_entry& term_entry,
#ifdef DEBUG
std::cerr << '\t' << courses_filename << std::endl;
std::cerr << '\t' << prereqs_filename << std::endl;
std::cerr << "Processing term " << term << "..." << std::endl;
#endif
std::cerr << "Processing term " << term << "..." << std::endl;
terms_offered["all_terms"].append(term);
Json::Value courses;
courses_file >> courses;
for(auto prefix : courses) {
handle_prefix(prefix,term,terms_offered);
}
}
void handle_prefix(const Json::Value& prefix,const std::string& term,Json::Value& terms_offered) {
#ifdef DEBUG
std::cerr << "\tProcessing prefix "
<< prefix["code"]
<< " - "
<< prefix["name"]
<< "..."
<< std::endl;
std::cerr << "\tProcessing prefix "
<< prefix["code"]
<< " - "
<< prefix["name"]
<< "..."
<< std::endl;
#endif
for(auto course : prefix["courses"]) {
handle_course(course,term,terms_offered);
}
for(auto course : prefix["courses"]) {
handle_course(course,term,terms_offered);
}
}
void handle_course(const Json::Value& course,const std::string& term,Json::Value& terms_offered) {
const auto course_code = course["id"].asString();
auto& course_term = terms_offered[course_code][term];
course_term["title"] = course["title"];
Json::Value sections = course["sections"];
course_sections_data_t course_sections_data = handle_sections(sections);
course_term["credits"]["min"] = course_sections_data.credMin;
course_term["credits"]["max"] = course_sections_data.credMax;
course_term["seats"]["taken"] = course_sections_data.seatsTaken;
course_term["seats"]["capacity"] = course_sections_data.capacity;
course_term["seats"]["remaining"] = course_sections_data.remaining;
auto& course_term = terms_offered[course_code][term];
course_term["title"] = course["title"];
handle_sections(sections,course_term);
handle_attributes(course["sections"][0],course_term);
}
void handle_sections(const Json::Value& sections,Json::Value& course_term) {
int credMin = Json::Value::maxInt, credMax = 0;
int seatsTaken = 0, capacity = 0, remaining = 0;
for(auto section : sections) {
// Get min/max credits *of all sections*
// (RCOS looking at you)
credMin = std::min(credMin,section["credMin"].asInt());
credMax = std::max(credMax,section["credMax"].asInt());
// Add seating data of all sections together.
// remaining might get clobbered by some sections
// having negative seats, but this probably won't
// be too much of an issue
seatsTaken += section["act"].asInt();
capacity += section["cap"].asInt();
remaining += section["rem"].asInt();
}
course_term["credits"]["min"] = credMin;
course_term["credits"]["max"] = credMax;
course_term["seats"]["taken"] = seatsTaken;
course_term["seats"]["capacity"] = capacity;
course_term["seats"]["remaining"] = remaining;
}
void handle_attributes(const Json::Value& section,Json::Value& course_term) {
// Need to do this because for some reason attributes are by
// section instead of by course
const auto section = course["sections"][0];
// This mess is basically C++'s string split but not using
// as much memory as an actual string split
const auto delim = std::regex(" and |, ");
const auto attributes_str = section["attribute"].asString();
const auto end_itr = std::sregex_token_iterator();
auto attributes_itr = std::sregex_token_iterator(
attributes_str.begin(),
attributes_str.end(),
delim,
-1
);
// Makes the JSON list of attributes
while(attributes_itr != end_itr
&& !attributes_itr->str().empty()) {
handle_attribute(course_term,attributes_itr->str());
handle_attribute(attributes_itr->str(),course_term);
attributes_itr++;
}
}
course_sections_data_t handle_sections(const Json::Value& sections) {
course_sections_data_t csd;
//csd.credMin = Json::Value::maxInt;
//csd.credMax = 0;
//csd.seatsTaken = 0, capacity = 0, remaining = 0;
for(auto section : sections) {
// Get min/max credits *of all sections*
// (RCOS looking at you)
csd.credMin = std::min(csd.credMin,section["credMin"].asInt());
csd.credMax = std::max(csd.credMax,section["credMax"].asInt());
// Add seating data of all sections together.
// remaining might get clobbered by some sections
// having negative seats, but this probably won't
// be too much of an issue
csd.seatsTaken += section["act"].asInt();
csd.capacity += section["cap"].asInt();
csd.remaining += section["rem"].asInt();
}
return csd;
}
void handle_attribute(Json::Value& course_term,const std::string& attribute) {
void handle_attribute(const std::string& attribute,Json::Value& course_term) {
std::string attr;
if(attribute == "Communication Intensive") {
attr = "[CI]";