path: root/libbible.cc

#include "libbible.h"
#include <sword/versekey.h>
#include <sword/markupfiltmgr.h>
#include <sword/swmodule.h>
#include <sword/swmgr.h>
#include <algorithm>

using namespace sword;
using namespace std;

SWMgr library(new MarkupFilterMgr(FMT_XHTML));

vector<string> getBooks(SWModule *target) {
    vector<string> books;
    VerseKey *key = (VerseKey *) target->getKey();
    for(char t = 1; t <= key->getTestamentMax(); t++) {
        key->setTestament(t);
        for(char b = 1; b <= key->getBookMax(); b++) {
            key->setBook(b);
            // Bug (whose fault??) in JPS; they CLAIM to have two testaments,
            // but they only have one, which causes repeats.
            if(std::find(books.begin(), books.end(), key->getBookName()) != books.end()) {
                continue;
            }
            // Another issue (maybe bug?) Some translations are NT only,
            // but still report OT books/chapters.
            if(string(target->renderText()).empty()) {
                continue;
            }
            books.push_back(key->getBookName());
        }
    }
    return books;
}

map<string, vector<string>> libbible::getModules() {
    library.load();
    map<string, vector<string>> mods;
    ModMap::iterator it;
    for (it = library.getModules().begin(); it != library.getModules().end(); it++) {
        string modName = (*it).second->getName();
        SWModule *target = library.getModule(modName.c_str());
        mods[modName] = getBooks(target);
    }
    return mods;
}

vector<libbible::passage> libbible::getPassages(string modName, string book) {
    vector<libbible::passage> passages;
    SWModule *target = library.getModule(modName.c_str());
    if(target == nullptr) {
        // Module doesn't exist
        return passages;
    }
    target->setKey((book + " " + "1").c_str());
    VerseKey *key = (VerseKey *) target->getKey();
    int maxChapter = key->getChapterMax();
    for(int chapter = 1; chapter <= maxChapter; chapter++) {
        string ref = book + ' ' + to_string(chapter);
        target->setKey(ref.c_str());
        VerseKey *key = (VerseKey *) target->getKey();
        libbible::passage pass;
        pass.modName = modName;
        pass.book = string(key->getBookName());
        pass.bookShort = string(key->getBookAbbrev());
        pass.chapterStart = chapter;
        pass.chapterEnd = chapter;
        pass.verseStart = 1;
        pass.verseEnd = key->getVerseMax();
        passages.push_back(pass);
    }
    return passages;
}

libbible::text getEmptyText(VerseKey *key) {
    libbible::text t;
    t.chapter = key->getChapter();
    t.verse = key->getVerse();
    t.book = key->getBookName();
    t.bookShort = key->getBookAbbrev();
    return t;
}

libbible::passage libbible::getPassage(string modName, string reference) {
    libbible::passage pass;
    pass.modName = modName;
    SWModule *target = library.getModule(pass.modName.c_str());
    if(target == nullptr || reference.empty()) {
        // Bad input
        return pass;
    }
    vector<string> validBooks = getBooks(target);
    //printf("Hey, I'm inferring missing parts!\n");
    // Let's use the target to help us
    target->setKey(reference.c_str());
    VerseKey *key = (VerseKey *) target->getKey();
    pass.book = string(key->getBookName());
    // Hold on a moment, is this book even legal?
    if(find(validBooks.begin(), validBooks.end(), pass.book) == validBooks.end()) {
        key->setBookName(validBooks[0].c_str());
        pass.book = string(key->getBookName());
    }
    pass.bookShort = string(key->getBookAbbrev());
    pass.chapterStart = key->getChapter();
    pass.verseStart = key->getVerse();
    //printf("Results so far: book: %s; chapterStart: %d; verseStart: %d\n", pass.book.c_str(), pass.chapterStart, pass.verseStart);
    // And now we just need chapterEnd and verseEnd. Yippee.
    string ref = string(reference);
    ref.erase(remove(ref.begin(), ref.end(), ' '), ref.end());
    if(ref.find('-') == string::npos) {
        // There's no range!
        if(ref.find(':') == string::npos) {
            // It's a full chapter reference
            pass.chapterEnd = pass.chapterStart;
            pass.verseEnd = key->getVerseMax();
        } else {
            // It's a single verse reference
            pass.chapterEnd = pass.chapterStart;
            pass.verseEnd = pass.verseStart;
            //printf("Hey, it's a single verse reference!\n");
        }
    } else {
        if(ref.find(':') == string::npos) {
            // It's a multi-full-chapter reference
            pass.chapterEnd = stoi(ref.substr(ref.find_last_of('-')+1));
            key->setChapter(pass.chapterEnd);
            pass.verseEnd = key->getVerseMax();
        } else {
            // It falls in categories c:v-v or c:v-c:v (or, technically, c-c:v)
            string rangeEnd = ref.substr(ref.find_last_of('-')+1);
            if(rangeEnd.find(':') == string::npos) {
                // It's c:v-v
                pass.verseEnd = stoi(rangeEnd);
                pass.chapterEnd = pass.chapterStart;
            } else {
                // It's c:v-c:v (or c-c:v, but code is the same)
                pass.chapterEnd = stoi(rangeEnd.substr(0, rangeEnd.find(':')));
                pass.verseEnd = stoi(rangeEnd.substr(rangeEnd.find(':')+1));
            }
        }
    }
    return pass;
}

vector<libbible::text> libbible::getText(libbible::passage pass) {
    vector<libbible::text> texts;
    SWModule *target = library.getModule(pass.modName.c_str());
    if(target == nullptr) {
        // Module doesn't exist
        return texts;
    }
    if(pass.book.empty()) {
        pass.book = pass.bookShort;
    }
    target->setKey((pass.book
                + " " + to_string(pass.chapterStart)
                + ":" + to_string(pass.verseStart)).c_str());
    VerseKey *key = (VerseKey *) target->getKey();

    bool endOfParagraph = false;

    for(; key->getChapter() < pass.chapterEnd ||
            (key->getChapter() == pass.chapterEnd && key->getVerse() <= pass.verseEnd);
            (*key)++) {

        string text = string(target->renderText());
        //printf("Processing text \"%s\"", text.c_str());

        // Handle ¶ symbol if at beginning of line
        /*if(text.find("¶") == 0) {
          text.erase(0, 1);
        // Append \n to text in previous texts (if applicable)
        if(! texts.empty()) {
        texts.back().text += '\n';
        }
        endOfParagraph = true;
        }*/

        texts.push_back(getEmptyText(key));

        /* Things to enforce:
         * 1. No whitespace at start of verse
         */

        // Find and replace everything in "subs" map
        /*const map<string, string> subs = {{"¶", "\n\t"}};
          for(auto const &repl : subs) {
          string::size_type location;
          while((location = text.find(repl.first)) != string::npos) {
          text.replace(location, repl.first.size(), repl.second);
          }
          }*/

        if(key->getVerse() == 1 || endOfParagraph) {
            if(find(texts.back().modifiers.begin(), texts.back().modifiers.end(), "paragraph") == texts.back().modifiers.end()) {
                texts.back().modifiers.push_back("paragraph");
            }
            endOfParagraph = false;
        }

        // Variable to accumulate unterminated spans
        std::vector<string> spans;
        bool spansChanged = false;
        bool hasAddedText = false;
        // Iterate over text
        for(auto i = text.begin(); i != text.end(); i++) {
            if(*i != '<') {
                if(spansChanged) {
                    spansChanged = false;
                    if(!texts.back().text.empty()) {
                        texts.push_back(getEmptyText(key));
                    }
                    for(string modifier : spans) {
                        if(find(texts.back().modifiers.begin(), texts.back().modifiers.end(), modifier) == texts.back().modifiers.end()) {
                            texts.back().modifiers.push_back(modifier);
                        }
                    }
                }
                if(! hasAddedText && (*i == ' ' || *i == '\t')) {
                    continue;
                }
                if(*i == "¶"[0] && *(i+1) == "¶"[1]) {
                    i++;
                    if(hasAddedText) {
                        texts.back().text += '\n';
                    } else {
                        // Append \n to text in previous texts (if applicable)
                        if(texts.size() > 1) {
                            texts[texts.size()-2].text += '\n';
                        }
                        texts.back().modifiers.push_back("paragraph");
                        continue;
                    }
                }
                texts.back().text += *i;
                hasAddedText = true;
            }
            else {
                string span;
                for(; i != text.end(); i++) {
                    span.push_back(*i);
                    if(*i == '>') {
                        // The end of the span will be "</span>".
                        if(span == "</span>") {
                            if(! spans.empty()) {
                                spans.pop_back();
                                spansChanged = true;
                            }
                        } else {
                            // The span will be formatted "<span class=\"NAME\">"
                            // We want just the NAME
                            size_t start = span.find_first_of('"')+1;
                            size_t end = span.find_last_of('"');
                            if(start > 0) {
                                spans.push_back(span.substr(start, end-start));
                                spansChanged = true;
                            }
                        }
                        break;
                    }
                }
            }
        }
        endOfParagraph = (text[text.length()-1] == '\n');
    }
    return texts;
}