MuPDF-赋予Similer功能以根据文本C语言制作情感数组

如何解决MuPDF-赋予Similer功能以根据文本C语言制作情感数组

我正在研究用C语言开发的ndk lib。我想为其添加功能。为此，我需要一个简单的功能，该功能类似于波纹管功能（它已经存在于lib中）。波纹管功能适用于给定文本中的搜索词，如果找到匹配项，它将以数组形式返回结果。同样，我希望从给定的页面文本中获取一系列句子。

JNIEXPORT jobjectArray JNICALL
JNI_FN(Mupdfpage_search)(jnienv * env,jobject thiz,jlong dochandle,jlong pagehandle,jstring text)
{

    renderdocument_t *doc = (renderdocument_t*) (long) dochandle;
    renderpage_t *page = (renderpage_t*) (long) pagehandle;
    // DEBUG("Mupdfpage(%p).search(%p,%p)",thiz,doc,page);

    if (!doc || !page)
    {
        return NULL;
    }

    const char *str = (*env)->GetStringUTFChars(env,text,NULL);
    if (str == NULL)
    {
        return NULL;
    }

    ArrayListHelper alh;
    PageTextBoxHelper ptbh;
    CharacterHelper ch;

    if (!ArrayListHelper_init(&alh,env) || !PageTextBoxHelper_init(&ptbh,env)|| !CharacterHelper_init(&ch,env))
    {
        DEBUG("search(): JNI helper initialization Failed"); //,pagehandle);
        return NULL;
    }
    jobject arrayList = ArrayListHelper_create(&alh);
    // DEBUG("Mupdfpage(%p).search(%p,%p): array: %p",page,arrayList);
    if (!arrayList)
    {
        return NULL;
    }

    fz_rect *hit_bBox = NULL;

    fz_stext_sheet *sheet = NULL;
    fz_stext_page *pagetext = NULL;
    fz_device *dev = NULL;
    int pos;
    int len;
    int i,n;
    int hit_count = 0;

    fz_try(doc->ctx)
    {
        fz_rect rect;

        // DEBUG("Mupdfpage(%p).search(%p,%p): load page text",page);

        fz_bound_page(doc->ctx,page->page,&rect);
        sheet = fz_new_stext_sheet(doc->ctx);
        pagetext = fz_new_stext_page(doc->ctx,&rect);
        dev = fz_new_stext_device(doc->ctx,sheet,pagetext,NULL);
        fz_run_page(doc->ctx,dev,&fz_identity,NULL);

        // DEBUG("Mupdfpage(%p).search(%p,%p): free text device",page);

        fz_close_device(doc->ctx,dev);
        fz_drop_device(doc->ctx,dev);
        dev = NULL;

        len = textlen(pagetext);

        // DEBUG("Mupdfpage(%p).search(%p,%p): text length: %d",len);

        for (pos = 0; pos < len; pos++)
        {
            fz_rect rr = fz_empty_rect;
            // DEBUG("Mupdfpage(%p).search(%p,%p): match %d",pos);

            n = match(doc->ctx,&ch,str,pos);
            if (n > 0)
            {
//                DEBUG("Mupdfpage(%p).search(%p,%p): match found: %d,%d",pos,n);
                for (i = 0; i < n; i++)
                {
                    fz_rect tmp_rr = bBoxcharat(doc->ctx,pos + i);
                    rr = *fz_union_rect(&rr,&tmp_rr);
                }

                if (!fz_is_empty_rect(&rr))
                {
                    int coords[4];
                    coords[0] = (rr.x0);
                    coords[1] = (rr.y0);
                    coords[2] = (rr.x1);
                    coords[3] = (rr.y1);
//                    DEBUG("Mupdfpage(%p).search(%p,%p): found rectangle (%d,%d - %d,%d)",coords[0],coords[1],coords[2],coords[3]);
                    jobject ptb = PageTextBoxHelper_create(&ptbh);
                    if (ptb)
                    {
                        // DEBUG("Mupdfpage(%p).search(%p,%p): rect %p",ptb);
                        PageTextBoxHelper_setRect(&ptbh,ptb,coords);
                        // PageTextBoxHelper_setText(&ptbh,txt);
                        // DEBUG("Mupdfpage(%p).search(%p,%p): add rect %p to array %p",arrayList);
                        ArrayListHelper_add(&alh,arrayList,ptb);
                    }
                }
            }
        }
    } fz_always(doc->ctx)
    {
        // DEBUG("Mupdfpage(%p).search(%p,%p): free resources",page);
        if (pagetext)
        {
            fz_drop_stext_page(doc->ctx,pagetext);
        }
        if (sheet)
        {
            fz_drop_stext_sheet(doc->ctx,sheet);
        }
        if (dev)
        {
            fz_drop_device(doc->ctx,dev);
        }
    }fz_catch(doc->ctx)
    {
        jclass cls;
        (*env)->ReleaseStringUTFChars(env,str);
        cls = (*env)->FindClass(env,"java/lang/OutOfMemoryError");
        if (cls != NULL)
        {
            (*env)->ThrowNew(env,cls,"Out of memory in MuPDFCore_searchPage");
        }
        (*env)->DeleteLocalRef(env,cls);
        return NULL;
    }

    (*env)->ReleaseStringUTFChars(env,str);

    return arrayList;
}