mirror of
https://github.com/kiwix/libkiwix.git
synced 2025-06-26 10:11:30 +00:00
Format all the code using clang-format.
Add a script `format_code.sh` to easily format the code.
This commit is contained in:
Matthieu Gautier
@ -24,25 +24,26 @@
|
||||
#include <winsock2.h>
|
||||
#include <ws2tcpip.h>
|
||||
#else
|
||||
#include <net/if.h>
|
||||
#include <netdb.h>
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
#include <unistd.h>
|
||||
#include <sys/ioctl.h>
|
||||
#include <sys/socket.h>
|
||||
#include <sys/types.h>
|
||||
#include <net/if.h>
|
||||
#include <netdb.h>
|
||||
#include <unistd.h>
|
||||
#endif
|
||||
|
||||
#include <iostream>
|
||||
#include <vector>
|
||||
#include <string>
|
||||
#include <map>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
namespace kiwix {
|
||||
std::map<std::string, std::string> getNetworkInterfaces();
|
||||
std::string getBestPublicIp();
|
||||
namespace kiwix
|
||||
{
|
||||
std::map<std::string, std::string> getNetworkInterfaces();
|
||||
std::string getBestPublicIp();
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -26,8 +26,9 @@
|
||||
#include <unistd.h>
|
||||
#endif
|
||||
|
||||
namespace kiwix {
|
||||
void sleep(unsigned int milliseconds);
|
||||
namespace kiwix
|
||||
{
|
||||
void sleep(unsigned int milliseconds);
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -20,18 +20,18 @@
|
||||
#ifndef KIWIX_PATHTOOLS_H
|
||||
#define KIWIX_PATHTOOLS_H
|
||||
|
||||
#include <fcntl.h>
|
||||
#include <limits.h>
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
#include <sys/stat.h>
|
||||
#include <sys/types.h>
|
||||
#include <fstream>
|
||||
#include <ios>
|
||||
#include <iostream>
|
||||
#include <sstream>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
#include <sstream>
|
||||
#include <iostream>
|
||||
#include <fstream>
|
||||
#include <string.h>
|
||||
#include <stdio.h>
|
||||
#include <sys/types.h>
|
||||
#include <fcntl.h>
|
||||
#include <sys/stat.h>
|
||||
#include <ios>
|
||||
#include <limits.h>
|
||||
|
||||
#ifdef _WIN32
|
||||
#include <direct.h>
|
||||
@ -41,20 +41,21 @@
|
||||
|
||||
using namespace std;
|
||||
|
||||
bool isRelativePath(const string &path);
|
||||
bool isRelativePath(const string& path);
|
||||
string computeAbsolutePath(const string path, const string relativePath);
|
||||
string computeRelativePath(const string path, const string absolutePath);
|
||||
string removeLastPathElement(const string path, const bool removePreSeparator = false,
|
||||
const bool removePostSeparator = false);
|
||||
string appendToDirectory(const string &directoryPath, const string &filename);
|
||||
string removeLastPathElement(const string path,
|
||||
const bool removePreSeparator = false,
|
||||
const bool removePostSeparator = false);
|
||||
string appendToDirectory(const string& directoryPath, const string& filename);
|
||||
|
||||
unsigned int getFileSize(const string &path);
|
||||
string getFileSizeAsString(const string &path);
|
||||
bool fileExists(const string &path);
|
||||
bool makeDirectory(const string &path);
|
||||
bool copyFile(const string &sourcePath, const string &destPath);
|
||||
string getLastPathElement(const string &path);
|
||||
unsigned int getFileSize(const string& path);
|
||||
string getFileSizeAsString(const string& path);
|
||||
bool fileExists(const string& path);
|
||||
bool makeDirectory(const string& path);
|
||||
bool copyFile(const string& sourcePath, const string& destPath);
|
||||
string getLastPathElement(const string& path);
|
||||
string getExecutablePath();
|
||||
string getCurrentDirectory();
|
||||
bool writeTextFile(const string &path, const string &content);
|
||||
bool writeTextFile(const string& path, const string& content);
|
||||
#endif
|
||||
|
||||
@ -22,11 +22,15 @@
|
||||
|
||||
#include <unicode/regex.h>
|
||||
#include <unicode/ucnv.h>
|
||||
#include <string>
|
||||
#include <map>
|
||||
#include <string>
|
||||
|
||||
bool matchRegex(const std::string &content, const std::string ®ex);
|
||||
std::string replaceRegex(const std::string &content, const std::string &replacement, const std::string ®ex);
|
||||
std::string appendToFirstOccurence(const std::string &content, const std::string regex, const std::string &replacement);
|
||||
bool matchRegex(const std::string& content, const std::string& regex);
|
||||
std::string replaceRegex(const std::string& content,
|
||||
const std::string& replacement,
|
||||
const std::string& regex);
|
||||
std::string appendToFirstOccurence(const std::string& content,
|
||||
const std::string regex,
|
||||
const std::string& replacement);
|
||||
|
||||
#endif
|
||||
|
||||
@ -22,44 +22,46 @@
|
||||
|
||||
#include <unicode/unistr.h>
|
||||
|
||||
#include <iostream>
|
||||
#include <vector>
|
||||
#include <string>
|
||||
#include <fstream>
|
||||
#include <iostream>
|
||||
#include <sstream>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
#include "pathTools.h"
|
||||
|
||||
namespace kiwix {
|
||||
|
||||
namespace kiwix
|
||||
{
|
||||
#ifndef __ANDROID__
|
||||
|
||||
std::string beautifyInteger(const unsigned int number);
|
||||
std::string beautifyFileSize(const unsigned int number);
|
||||
std::string urlEncode(const std::string &c);
|
||||
void printStringInHexadecimal(const char *s);
|
||||
void printStringInHexadecimal(UnicodeString s);
|
||||
void stringReplacement(std::string& str, const std::string& oldStr, const std::string& newStr);
|
||||
std::string encodeDiples(const std::string& str);
|
||||
std::string beautifyInteger(const unsigned int number);
|
||||
std::string beautifyFileSize(const unsigned int number);
|
||||
std::string urlEncode(const std::string& c);
|
||||
void printStringInHexadecimal(const char* s);
|
||||
void printStringInHexadecimal(UnicodeString s);
|
||||
void stringReplacement(std::string& str,
|
||||
const std::string& oldStr,
|
||||
const std::string& newStr);
|
||||
std::string encodeDiples(const std::string& str);
|
||||
|
||||
#endif
|
||||
|
||||
std::string removeAccents(const std::string &text);
|
||||
void loadICUExternalTables();
|
||||
std::string urlDecode(const std::string &c);
|
||||
std::string removeAccents(const std::string& text);
|
||||
void loadICUExternalTables();
|
||||
std::string urlDecode(const std::string& c);
|
||||
|
||||
std::vector<std::string> split(const std::string&, const std::string&);
|
||||
std::vector<std::string> split(const char*, const char*);
|
||||
std::vector<std::string> split(const std::string&, const char*);
|
||||
std::vector<std::string> split(const char*, const std::string&);
|
||||
std::vector<std::string> split(const std::string&, const std::string&);
|
||||
std::vector<std::string> split(const char*, const char*);
|
||||
std::vector<std::string> split(const std::string&, const char*);
|
||||
std::vector<std::string> split(const char*, const std::string&);
|
||||
|
||||
std::string ucAll(const std::string &word);
|
||||
std::string lcAll(const std::string &word);
|
||||
std::string ucFirst(const std::string &word);
|
||||
std::string lcFirst(const std::string &word);
|
||||
std::string toTitle(const std::string &word);
|
||||
std::string ucAll(const std::string& word);
|
||||
std::string lcAll(const std::string& word);
|
||||
std::string ucFirst(const std::string& word);
|
||||
std::string lcFirst(const std::string& word);
|
||||
std::string toTitle(const std::string& word);
|
||||
|
||||
std::string normalize(const std::string &word);
|
||||
std::string normalize(const std::string& word);
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -22,5 +22,4 @@
|
||||
|
||||
#include "library.h"
|
||||
|
||||
|
||||
#endif
|
||||
@ -22,86 +22,85 @@
|
||||
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string>
|
||||
#include <string.h>
|
||||
#include <vector>
|
||||
#include <stack>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
#include "common/stringTools.h"
|
||||
#include "common/regexTools.h"
|
||||
#include "common/stringTools.h"
|
||||
|
||||
#define KIWIX_LIBRARY_VERSION "20110515"
|
||||
|
||||
using namespace std;
|
||||
|
||||
namespace kiwix {
|
||||
namespace kiwix
|
||||
{
|
||||
enum supportedIndexType { UNKNOWN, XAPIAN };
|
||||
|
||||
enum supportedIndexType { UNKNOWN, XAPIAN };
|
||||
class Book
|
||||
{
|
||||
public:
|
||||
Book();
|
||||
~Book();
|
||||
|
||||
class Book {
|
||||
static bool sortByLastOpen(const Book& a, const Book& b);
|
||||
static bool sortByTitle(const Book& a, const Book& b);
|
||||
static bool sortBySize(const Book& a, const Book& b);
|
||||
static bool sortByDate(const Book& a, const Book& b);
|
||||
static bool sortByCreator(const Book& a, const Book& b);
|
||||
static bool sortByPublisher(const Book& a, const Book& b);
|
||||
static bool sortByLanguage(const Book& a, const Book& b);
|
||||
string getHumanReadableIdFromPath();
|
||||
|
||||
public:
|
||||
Book();
|
||||
~Book();
|
||||
string id;
|
||||
string path;
|
||||
string pathAbsolute;
|
||||
string last;
|
||||
string indexPath;
|
||||
string indexPathAbsolute;
|
||||
supportedIndexType indexType;
|
||||
string title;
|
||||
string description;
|
||||
string language;
|
||||
string creator;
|
||||
string publisher;
|
||||
string date;
|
||||
string url;
|
||||
string name;
|
||||
string tags;
|
||||
string origId;
|
||||
string articleCount;
|
||||
string mediaCount;
|
||||
bool readOnly;
|
||||
string size;
|
||||
string favicon;
|
||||
string faviconMimeType;
|
||||
};
|
||||
|
||||
static bool sortByLastOpen(const Book &a, const Book &b);
|
||||
static bool sortByTitle(const Book &a, const Book &b);
|
||||
static bool sortBySize(const Book &a, const Book &b);
|
||||
static bool sortByDate(const Book &a, const Book &b);
|
||||
static bool sortByCreator(const Book &a, const Book &b);
|
||||
static bool sortByPublisher(const Book &a, const Book &b);
|
||||
static bool sortByLanguage(const Book &a, const Book &b);
|
||||
string getHumanReadableIdFromPath();
|
||||
class Library
|
||||
{
|
||||
public:
|
||||
Library();
|
||||
~Library();
|
||||
|
||||
string id;
|
||||
string path;
|
||||
string pathAbsolute;
|
||||
string last;
|
||||
string indexPath;
|
||||
string indexPathAbsolute;
|
||||
supportedIndexType indexType;
|
||||
string title;
|
||||
string description;
|
||||
string language;
|
||||
string creator;
|
||||
string publisher;
|
||||
string date;
|
||||
string url;
|
||||
string name;
|
||||
string tags;
|
||||
string origId;
|
||||
string articleCount;
|
||||
string mediaCount;
|
||||
bool readOnly;
|
||||
string size;
|
||||
string favicon;
|
||||
string faviconMimeType;
|
||||
};
|
||||
|
||||
class Library {
|
||||
|
||||
public:
|
||||
Library();
|
||||
~Library();
|
||||
|
||||
string version;
|
||||
bool addBook(const Book &book);
|
||||
bool removeBookByIndex(const unsigned int bookIndex);
|
||||
vector <kiwix::Book> books;
|
||||
|
||||
/*
|
||||
* 'current' is the variable storing the current content/book id
|
||||
* in the library. This is used to be able to load per default a
|
||||
* content. As Kiwix may work with many library XML files, you may
|
||||
* have "current" defined many time with different values. The
|
||||
* last XML file read has the priority, Although we do not have an
|
||||
* library object for each file, we want to be able to fallback to
|
||||
* an 'old' current book if the one which should be load
|
||||
* failed. That is the reason why we need a stack here
|
||||
*/
|
||||
stack<string> current;
|
||||
};
|
||||
string version;
|
||||
bool addBook(const Book& book);
|
||||
bool removeBookByIndex(const unsigned int bookIndex);
|
||||
vector<kiwix::Book> books;
|
||||
|
||||
/*
|
||||
* 'current' is the variable storing the current content/book id
|
||||
* in the library. This is used to be able to load per default a
|
||||
* content. As Kiwix may work with many library XML files, you may
|
||||
* have "current" defined many time with different values. The
|
||||
* last XML file read has the priority, Although we do not have an
|
||||
* library object for each file, we want to be able to fallback to
|
||||
* an 'old' current book if the one which should be load
|
||||
* failed. That is the reason why we need a stack here
|
||||
*/
|
||||
stack<string> current;
|
||||
};
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -20,73 +20,89 @@
|
||||
#ifndef KIWIX_MANAGER_H
|
||||
#define KIWIX_MANAGER_H
|
||||
|
||||
#include <string>
|
||||
#include <sstream>
|
||||
#include <time.h>
|
||||
#include <sstream>
|
||||
#include <string>
|
||||
|
||||
#include <pugixml.hpp>
|
||||
|
||||
#include "common/base64.h"
|
||||
#include "common/regexTools.h"
|
||||
#include "common/pathTools.h"
|
||||
#include "common/regexTools.h"
|
||||
#include "library.h"
|
||||
#include "reader.h"
|
||||
|
||||
using namespace std;
|
||||
|
||||
namespace kiwix {
|
||||
namespace kiwix
|
||||
{
|
||||
enum supportedListMode { LASTOPEN, REMOTE, LOCAL };
|
||||
enum supportedListSortBy { TITLE, SIZE, DATE, CREATOR, PUBLISHER };
|
||||
|
||||
enum supportedListMode { LASTOPEN, REMOTE, LOCAL };
|
||||
enum supportedListSortBy { TITLE, SIZE, DATE, CREATOR, PUBLISHER };
|
||||
class Manager
|
||||
{
|
||||
public:
|
||||
Manager();
|
||||
~Manager();
|
||||
|
||||
class Manager {
|
||||
bool readFile(const string path, const bool readOnly = true);
|
||||
bool readFile(const string nativePath,
|
||||
const string UTF8Path,
|
||||
const bool readOnly = true);
|
||||
bool readXml(const string xml,
|
||||
const bool readOnly = true,
|
||||
const string libraryPath = "");
|
||||
bool writeFile(const string path);
|
||||
bool removeBookByIndex(const unsigned int bookIndex);
|
||||
bool removeBookById(const string id);
|
||||
bool setCurrentBookId(const string id);
|
||||
string getCurrentBookId();
|
||||
bool setBookIndex(const string id,
|
||||
const string path,
|
||||
const supportedIndexType type);
|
||||
bool setBookIndex(const string id, const string path);
|
||||
bool setBookPath(const string id, const string path);
|
||||
string addBookFromPathAndGetId(const string pathToOpen,
|
||||
const string pathToSave = "",
|
||||
const string url = "",
|
||||
const bool checkMetaData = false);
|
||||
bool addBookFromPath(const string pathToOpen,
|
||||
const string pathToSave = "",
|
||||
const string url = "",
|
||||
const bool checkMetaData = false);
|
||||
Library cloneLibrary();
|
||||
bool getBookById(const string id, Book& book);
|
||||
bool getCurrentBook(Book& book);
|
||||
unsigned int getBookCount(const bool localBooks, const bool remoteBooks);
|
||||
bool updateBookLastOpenDateById(const string id);
|
||||
void removeBookPaths();
|
||||
bool listBooks(const supportedListMode mode,
|
||||
const supportedListSortBy sortBy,
|
||||
const unsigned int maxSize,
|
||||
const string language,
|
||||
const string creator,
|
||||
const string publisher,
|
||||
const string search);
|
||||
vector<string> getBooksLanguages();
|
||||
vector<string> getBooksCreators();
|
||||
vector<string> getBooksPublishers();
|
||||
vector<string> getBooksIds();
|
||||
|
||||
public:
|
||||
Manager();
|
||||
~Manager();
|
||||
string writableLibraryPath;
|
||||
|
||||
bool readFile(const string path, const bool readOnly = true);
|
||||
bool readFile(const string nativePath, const string UTF8Path, const bool readOnly = true);
|
||||
bool readXml(const string xml, const bool readOnly = true, const string libraryPath = "");
|
||||
bool writeFile(const string path);
|
||||
bool removeBookByIndex(const unsigned int bookIndex);
|
||||
bool removeBookById(const string id);
|
||||
bool setCurrentBookId(const string id);
|
||||
string getCurrentBookId();
|
||||
bool setBookIndex(const string id, const string path, const supportedIndexType type);
|
||||
bool setBookIndex(const string id, const string path);
|
||||
bool setBookPath(const string id, const string path);
|
||||
string addBookFromPathAndGetId(const string pathToOpen, const string pathToSave = "", const string url = "",
|
||||
const bool checkMetaData = false);
|
||||
bool addBookFromPath(const string pathToOpen, const string pathToSave = "", const string url = "",
|
||||
const bool checkMetaData = false);
|
||||
Library cloneLibrary();
|
||||
bool getBookById(const string id, Book &book);
|
||||
bool getCurrentBook(Book &book);
|
||||
unsigned int getBookCount(const bool localBooks, const bool remoteBooks);
|
||||
bool updateBookLastOpenDateById(const string id);
|
||||
void removeBookPaths();
|
||||
bool listBooks(const supportedListMode mode, const supportedListSortBy sortBy, const unsigned int maxSize,
|
||||
const string language, const string creator, const string publisher, const string search);
|
||||
vector<string> getBooksLanguages();
|
||||
vector<string> getBooksCreators();
|
||||
vector<string> getBooksPublishers();
|
||||
vector<string> getBooksIds();
|
||||
vector<std::string> bookIdList;
|
||||
|
||||
string writableLibraryPath;
|
||||
protected:
|
||||
kiwix::Library library;
|
||||
|
||||
vector<std::string> bookIdList;
|
||||
|
||||
protected:
|
||||
kiwix::Library library;
|
||||
|
||||
bool readBookFromPath(const string path, Book *book = NULL);
|
||||
bool parseXmlDom(const pugi::xml_document &doc, const bool readOnly, const string libraryPath);
|
||||
|
||||
private:
|
||||
void checkAndCleanBookPaths(Book &book, const string &libraryPath);
|
||||
};
|
||||
bool readBookFromPath(const string path, Book* book = NULL);
|
||||
bool parseXmlDom(const pugi::xml_document& doc,
|
||||
const bool readOnly,
|
||||
const string libraryPath);
|
||||
|
||||
private:
|
||||
void checkAndCleanBookPaths(Book& book, const string& libraryPath);
|
||||
};
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
154
include/reader.h
154
include/reader.h
@ -20,85 +20,105 @@
|
||||
#ifndef KIWIX_READER_H
|
||||
#define KIWIX_READER_H
|
||||
|
||||
#include <zim/zim.h>
|
||||
#include <zim/file.h>
|
||||
#include <zim/article.h>
|
||||
#include <zim/fileiterator.h>
|
||||
#include <stdio.h>
|
||||
#include <string>
|
||||
#include <zim/article.h>
|
||||
#include <zim/file.h>
|
||||
#include <zim/fileiterator.h>
|
||||
#include <zim/zim.h>
|
||||
#include <exception>
|
||||
#include <sstream>
|
||||
#include <map>
|
||||
#include <sstream>
|
||||
#include <string>
|
||||
#include "common/pathTools.h"
|
||||
#include "common/stringTools.h"
|
||||
|
||||
using namespace std;
|
||||
|
||||
namespace kiwix {
|
||||
namespace kiwix
|
||||
{
|
||||
class Reader
|
||||
{
|
||||
public:
|
||||
Reader(const string zimFilePath);
|
||||
~Reader();
|
||||
|
||||
class Reader {
|
||||
void reset();
|
||||
unsigned int getArticleCount() const;
|
||||
unsigned int getMediaCount() const;
|
||||
unsigned int getGlobalCount() const;
|
||||
string getZimFilePath() const;
|
||||
string getId() const;
|
||||
string getRandomPageUrl() const;
|
||||
string getFirstPageUrl() const;
|
||||
string getMainPageUrl() const;
|
||||
bool getMetatag(const string& url, string& content) const;
|
||||
string getTitle() const;
|
||||
string getDescription() const;
|
||||
string getLanguage() const;
|
||||
string getName() const;
|
||||
string getTags() const;
|
||||
string getDate() const;
|
||||
string getCreator() const;
|
||||
string getPublisher() const;
|
||||
string getOrigId() const;
|
||||
bool getFavicon(string& content, string& mimeType) const;
|
||||
bool getPageUrlFromTitle(const string& title, string& url) const;
|
||||
bool getMimeTypeByUrl(const string& url, string& mimeType) const;
|
||||
bool getContentByUrl(const string& url,
|
||||
string& content,
|
||||
unsigned int& contentLength,
|
||||
string& contentType) const;
|
||||
bool getContentByEncodedUrl(const string& url,
|
||||
string& content,
|
||||
unsigned int& contentLength,
|
||||
string& contentType,
|
||||
string& baseUrl) const;
|
||||
bool getContentByEncodedUrl(const string& url,
|
||||
string& content,
|
||||
unsigned int& contentLength,
|
||||
string& contentType) const;
|
||||
bool getContentByDecodedUrl(const string& url,
|
||||
string& content,
|
||||
unsigned int& contentLength,
|
||||
string& contentType,
|
||||
string& baseUrl) const;
|
||||
bool getContentByDecodedUrl(const string& url,
|
||||
string& content,
|
||||
unsigned int& contentLength,
|
||||
string& contentType) const;
|
||||
bool searchSuggestions(const string& prefix,
|
||||
unsigned int suggestionsCount,
|
||||
const bool reset = true);
|
||||
bool searchSuggestionsSmart(const string& prefix,
|
||||
unsigned int suggestionsCount);
|
||||
bool urlExists(const string& url) const;
|
||||
bool hasFulltextIndex() const;
|
||||
std::vector<std::string> getTitleVariants(const std::string& title) const;
|
||||
bool getNextSuggestion(string& title);
|
||||
bool getNextSuggestion(string& title, string& url);
|
||||
bool canCheckIntegrity() const;
|
||||
bool isCorrupted() const;
|
||||
bool parseUrl(const string& url, char* ns, string& title) const;
|
||||
unsigned int getFileSize() const;
|
||||
zim::File* getZimFileHandler() const;
|
||||
bool getArticleObjectByDecodedUrl(const string& url,
|
||||
zim::Article& article) const;
|
||||
|
||||
public:
|
||||
Reader(const string zimFilePath);
|
||||
~Reader();
|
||||
protected:
|
||||
zim::File* zimFileHandler;
|
||||
zim::size_type firstArticleOffset;
|
||||
zim::size_type lastArticleOffset;
|
||||
zim::size_type currentArticleOffset;
|
||||
zim::size_type nsACount;
|
||||
zim::size_type nsICount;
|
||||
std::string zimFilePath;
|
||||
|
||||
void reset();
|
||||
unsigned int getArticleCount() const;
|
||||
unsigned int getMediaCount() const;
|
||||
unsigned int getGlobalCount() const;
|
||||
string getZimFilePath() const;
|
||||
string getId() const;
|
||||
string getRandomPageUrl() const;
|
||||
string getFirstPageUrl() const;
|
||||
string getMainPageUrl() const;
|
||||
bool getMetatag(const string &url, string &content) const;
|
||||
string getTitle() const;
|
||||
string getDescription() const;
|
||||
string getLanguage() const;
|
||||
string getName() const;
|
||||
string getTags() const;
|
||||
string getDate() const;
|
||||
string getCreator() const;
|
||||
string getPublisher() const;
|
||||
string getOrigId() const;
|
||||
bool getFavicon(string &content, string &mimeType) const;
|
||||
bool getPageUrlFromTitle(const string &title, string &url) const;
|
||||
bool getMimeTypeByUrl(const string &url, string &mimeType) const;
|
||||
bool getContentByUrl(const string &url, string &content, unsigned int &contentLength, string &contentType) const;
|
||||
bool getContentByEncodedUrl(const string &url, string &content, unsigned int &contentLength, string &contentType, string &baseUrl) const;
|
||||
bool getContentByEncodedUrl(const string &url, string &content, unsigned int &contentLength, string &contentType) const;
|
||||
bool getContentByDecodedUrl(const string &url, string &content, unsigned int &contentLength, string &contentType, string &baseUrl) const;
|
||||
bool getContentByDecodedUrl(const string &url, string &content, unsigned int &contentLength, string &contentType) const;
|
||||
bool searchSuggestions(const string &prefix, unsigned int suggestionsCount, const bool reset = true);
|
||||
bool searchSuggestionsSmart(const string &prefix, unsigned int suggestionsCount);
|
||||
bool urlExists(const string &url) const;
|
||||
bool hasFulltextIndex() const;
|
||||
std::vector<std::string> getTitleVariants(const std::string &title) const;
|
||||
bool getNextSuggestion(string &title);
|
||||
bool getNextSuggestion(string &title, string &url);
|
||||
bool canCheckIntegrity() const;
|
||||
bool isCorrupted() const;
|
||||
bool parseUrl(const string &url, char *ns, string &title) const;
|
||||
unsigned int getFileSize() const;
|
||||
zim::File* getZimFileHandler() const;
|
||||
bool getArticleObjectByDecodedUrl(const string &url, zim::Article &article) const;
|
||||
|
||||
protected:
|
||||
zim::File* zimFileHandler;
|
||||
zim::size_type firstArticleOffset;
|
||||
zim::size_type lastArticleOffset;
|
||||
zim::size_type currentArticleOffset;
|
||||
zim::size_type nsACount;
|
||||
zim::size_type nsICount;
|
||||
std::string zimFilePath;
|
||||
|
||||
std::vector< std::vector<std::string> > suggestions;
|
||||
std::vector< std::vector<std::string> >::iterator suggestionsOffset;
|
||||
|
||||
private:
|
||||
std::map<const std::string, unsigned int> parseCounterMetadata() const;
|
||||
};
|
||||
std::vector<std::vector<std::string>> suggestions;
|
||||
std::vector<std::vector<std::string>>::iterator suggestionsOffset;
|
||||
|
||||
private:
|
||||
std::map<const std::string, unsigned int> parseCounterMetadata() const;
|
||||
};
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -22,73 +22,77 @@
|
||||
|
||||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <string>
|
||||
#include <unicode/putil.h>
|
||||
#include <algorithm>
|
||||
#include <vector>
|
||||
#include <locale>
|
||||
#include <cctype>
|
||||
#include <locale>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
#include <vector>
|
||||
#include "common/pathTools.h"
|
||||
#include "common/stringTools.h"
|
||||
#include <unicode/putil.h>
|
||||
#include "kiwix_config.h"
|
||||
|
||||
using namespace std;
|
||||
|
||||
namespace kiwix {
|
||||
class Reader;
|
||||
class Result {
|
||||
public:
|
||||
virtual ~Result() {};
|
||||
virtual std::string get_url() = 0;
|
||||
virtual std::string get_title() = 0;
|
||||
virtual int get_score() = 0;
|
||||
virtual std::string get_snippet() = 0;
|
||||
virtual int get_wordCount() = 0;
|
||||
virtual int get_size() = 0;
|
||||
};
|
||||
namespace kiwix
|
||||
{
|
||||
class Reader;
|
||||
class Result
|
||||
{
|
||||
public:
|
||||
virtual ~Result(){};
|
||||
virtual std::string get_url() = 0;
|
||||
virtual std::string get_title() = 0;
|
||||
virtual int get_score() = 0;
|
||||
virtual std::string get_snippet() = 0;
|
||||
virtual int get_wordCount() = 0;
|
||||
virtual int get_size() = 0;
|
||||
};
|
||||
|
||||
struct SearcherInternal;
|
||||
class Searcher
|
||||
{
|
||||
public:
|
||||
Searcher(const string& xapianDirectoryPath, Reader* reader);
|
||||
~Searcher();
|
||||
|
||||
struct SearcherInternal;
|
||||
class Searcher {
|
||||
|
||||
public:
|
||||
Searcher(const string &xapianDirectoryPath, Reader* reader);
|
||||
~Searcher();
|
||||
|
||||
void search(std::string &search, unsigned int resultStart,
|
||||
unsigned int resultEnd, const bool verbose=false);
|
||||
Result* getNextResult();
|
||||
void restart_search();
|
||||
unsigned int getEstimatedResultCount();
|
||||
bool setProtocolPrefix(const std::string prefix);
|
||||
bool setSearchProtocolPrefix(const std::string prefix);
|
||||
void reset();
|
||||
void setContentHumanReadableId(const string &contentHumanReadableId);
|
||||
void search(std::string& search,
|
||||
unsigned int resultStart,
|
||||
unsigned int resultEnd,
|
||||
const bool verbose = false);
|
||||
Result* getNextResult();
|
||||
void restart_search();
|
||||
unsigned int getEstimatedResultCount();
|
||||
bool setProtocolPrefix(const std::string prefix);
|
||||
bool setSearchProtocolPrefix(const std::string prefix);
|
||||
void reset();
|
||||
void setContentHumanReadableId(const string& contentHumanReadableId);
|
||||
|
||||
#ifdef ENABLE_CTPP2
|
||||
string getHtml();
|
||||
string getHtml();
|
||||
#endif
|
||||
|
||||
protected:
|
||||
std::string beautifyInteger(const unsigned int number);
|
||||
void closeIndex() ;
|
||||
void searchInIndex(string &search, const unsigned int resultStart,
|
||||
const unsigned int resultEnd, const bool verbose=false);
|
||||
|
||||
Reader* reader;
|
||||
SearcherInternal* internal;
|
||||
std::string searchPattern;
|
||||
std::string protocolPrefix;
|
||||
std::string searchProtocolPrefix;
|
||||
std::string template_ct2;
|
||||
unsigned int resultCountPerPage;
|
||||
unsigned int estimatedResultCount;
|
||||
unsigned int resultStart;
|
||||
unsigned int resultEnd;
|
||||
std::string contentHumanReadableId;
|
||||
};
|
||||
protected:
|
||||
std::string beautifyInteger(const unsigned int number);
|
||||
void closeIndex();
|
||||
void searchInIndex(string& search,
|
||||
const unsigned int resultStart,
|
||||
const unsigned int resultEnd,
|
||||
const bool verbose = false);
|
||||
|
||||
Reader* reader;
|
||||
SearcherInternal* internal;
|
||||
std::string searchPattern;
|
||||
std::string protocolPrefix;
|
||||
std::string searchProtocolPrefix;
|
||||
std::string template_ct2;
|
||||
unsigned int resultCountPerPage;
|
||||
unsigned int estimatedResultCount;
|
||||
unsigned int resultStart;
|
||||
unsigned int resultEnd;
|
||||
std::string contentHumanReadableId;
|
||||
};
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
@ -21,70 +21,76 @@
|
||||
#define KIWIX_XAPIAN_SEARCHER_H
|
||||
|
||||
#include <xapian.h>
|
||||
#include "searcher.h"
|
||||
#include "reader.h"
|
||||
#include "searcher.h"
|
||||
|
||||
#include <map>
|
||||
#include <string>
|
||||
|
||||
using namespace std;
|
||||
|
||||
namespace kiwix {
|
||||
namespace kiwix
|
||||
{
|
||||
class XapianSearcher;
|
||||
|
||||
class XapianSearcher;
|
||||
class XapianResult : public Result
|
||||
{
|
||||
public:
|
||||
XapianResult(XapianSearcher* searcher, Xapian::MSetIterator& iterator);
|
||||
virtual ~XapianResult(){};
|
||||
|
||||
class XapianResult : public Result {
|
||||
public:
|
||||
XapianResult(XapianSearcher* searcher, Xapian::MSetIterator& iterator);
|
||||
virtual ~XapianResult() {};
|
||||
virtual std::string get_url();
|
||||
virtual std::string get_title();
|
||||
virtual int get_score();
|
||||
virtual std::string get_snippet();
|
||||
virtual int get_wordCount();
|
||||
virtual int get_size();
|
||||
|
||||
virtual std::string get_url();
|
||||
virtual std::string get_title();
|
||||
virtual int get_score();
|
||||
virtual std::string get_snippet();
|
||||
virtual int get_wordCount();
|
||||
virtual int get_size();
|
||||
private:
|
||||
XapianSearcher* searcher;
|
||||
Xapian::MSetIterator iterator;
|
||||
Xapian::Document document;
|
||||
};
|
||||
|
||||
private:
|
||||
XapianSearcher* searcher;
|
||||
Xapian::MSetIterator iterator;
|
||||
Xapian::Document document;
|
||||
};
|
||||
class NoXapianIndexInZim : public exception
|
||||
{
|
||||
virtual const char* what() const throw()
|
||||
{
|
||||
return "There is no fulltext index in the zim file";
|
||||
}
|
||||
};
|
||||
|
||||
class NoXapianIndexInZim: public exception {
|
||||
virtual const char* what() const throw() {
|
||||
return "There is no fulltext index in the zim file";
|
||||
}
|
||||
};
|
||||
class XapianSearcher
|
||||
{
|
||||
friend class XapianResult;
|
||||
|
||||
class XapianSearcher {
|
||||
friend class XapianResult;
|
||||
public:
|
||||
XapianSearcher(const string &xapianDirectoryPath, Reader* reader);
|
||||
virtual ~XapianSearcher() {};
|
||||
void searchInIndex(string &search, const unsigned int resultStart, const unsigned int resultEnd,
|
||||
const bool verbose=false);
|
||||
virtual Result* getNextResult();
|
||||
void restart_search();
|
||||
public:
|
||||
XapianSearcher(const string& xapianDirectoryPath, Reader* reader);
|
||||
virtual ~XapianSearcher(){};
|
||||
void searchInIndex(string& search,
|
||||
const unsigned int resultStart,
|
||||
const unsigned int resultEnd,
|
||||
const bool verbose = false);
|
||||
virtual Result* getNextResult();
|
||||
void restart_search();
|
||||
|
||||
Xapian::MSet results;
|
||||
Xapian::MSet results;
|
||||
|
||||
protected:
|
||||
void closeIndex();
|
||||
void openIndex(const string &xapianDirectoryPath);
|
||||
void setup_queryParser();
|
||||
|
||||
Reader* reader;
|
||||
Xapian::Database readableDatabase;
|
||||
std::string language;
|
||||
std::string stopwords;
|
||||
Xapian::QueryParser queryParser;
|
||||
Xapian::Stem stemmer;
|
||||
Xapian::SimpleStopper stopper;
|
||||
Xapian::MSetIterator current_result;
|
||||
std::map<std::string, int> valuesmap;
|
||||
};
|
||||
protected:
|
||||
void closeIndex();
|
||||
void openIndex(const string& xapianDirectoryPath);
|
||||
void setup_queryParser();
|
||||
|
||||
Reader* reader;
|
||||
Xapian::Database readableDatabase;
|
||||
std::string language;
|
||||
std::string stopwords;
|
||||
Xapian::QueryParser queryParser;
|
||||
Xapian::Stem stemmer;
|
||||
Xapian::SimpleStopper stopper;
|
||||
Xapian::MSetIterator current_result;
|
||||
std::map<std::string, int> valuesmap;
|
||||
};
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
Reference in New Issue
Block a user