jeanlemotan
585 lines
14 KiB
C++
585 lines
14 KiB
C++
#pragma once
|
|
|
|
#include "tl/detail/prologue.h"
|
|
#include "tl/functional.h"
|
|
#include <EASTL/vector.h>
|
|
#include <tl/fixed_vector.h>
|
|
|
|
#include "tl/string.h"
|
|
#include "tl/optional.h"
|
|
|
|
namespace tl
|
|
{
|
|
namespace algorithm
|
|
{
|
|
enum class empty_token_policy
|
|
{
|
|
discard,
|
|
keep
|
|
};
|
|
|
|
template <class DstContainer, class SrcContainer>
|
|
DstContainer to_lower_ascii_copy(const SrcContainer& str) noexcept
|
|
{
|
|
if constexpr (tl::is_same_v<DstContainer, string>)
|
|
{
|
|
if (str.empty())
|
|
return string();
|
|
|
|
const string::size_type s = str.size();
|
|
fixed_vector<char, 512> b(s);
|
|
char const* string = str.data();
|
|
for (string::size_type i = 0; i < s; i++)
|
|
b[i] = (char)ascii::tolower(string[i]);
|
|
|
|
return { b.data(), s };
|
|
}
|
|
else
|
|
{
|
|
DstContainer dst;
|
|
dst.resize(str.size());
|
|
eastl::transform(tl::begin(str), tl::end(str), tl::begin(dst), ascii::tolower);
|
|
return dst;
|
|
}}
|
|
|
|
template <class Container>
|
|
Container to_lower_ascii_copy(const Container& str) noexcept
|
|
{
|
|
return to_lower_ascii_copy<Container, Container>(str);
|
|
}
|
|
|
|
|
|
template <class DstContainer, class SrcContainer>
|
|
DstContainer to_upper_ascii_copy(const SrcContainer& str) noexcept
|
|
{
|
|
if constexpr (tl::is_same_v<DstContainer, string>)
|
|
{
|
|
if (str.empty())
|
|
return string();
|
|
|
|
const string::size_type s = str.size();
|
|
fixed_vector<char, 512> b(s);
|
|
char const* string = str.data();
|
|
for (string::size_type i = 0; i < s; i++)
|
|
b[i] = (char)ascii::toupper(string[i]);
|
|
|
|
return { b.data(), s };
|
|
}
|
|
else
|
|
{
|
|
DstContainer dst;
|
|
dst.resize(str.size());
|
|
eastl::transform(tl::begin(str), tl::end(str), tl::begin(dst), ascii::toupper);
|
|
return dst;
|
|
}
|
|
}
|
|
|
|
template <class Container>
|
|
Container to_upper_ascii_copy(const Container& str) noexcept
|
|
{
|
|
return to_upper_ascii_copy<Container, Container>(str);
|
|
}
|
|
|
|
template <class Container>
|
|
void to_lower_ascii(Container& str) noexcept
|
|
{
|
|
eastl::transform(tl::begin(str), tl::end(str), tl::begin(str), ascii::tolower);
|
|
}
|
|
|
|
template <>
|
|
inline void to_lower_ascii(string& str) noexcept
|
|
{
|
|
str = to_lower_ascii_copy(str);
|
|
}
|
|
|
|
template <class Container>
|
|
void to_upper_ascii(Container& str) noexcept
|
|
{
|
|
eastl::transform(tl::begin(str), tl::end(str), tl::begin(str), ascii::toupper);
|
|
}
|
|
|
|
template <>
|
|
inline void to_upper_ascii(string& str) noexcept
|
|
{
|
|
str = to_upper_ascii_copy(str);
|
|
}
|
|
|
|
template <class String>
|
|
cppcoro::generator<const String&> split_on_any(const String& str, char sep, empty_token_policy token_policy = empty_token_policy::discard) noexcept
|
|
{
|
|
String token;
|
|
const char* cstr = str.data();
|
|
size_t offset = 0;
|
|
const size_t endOffset = str.size();
|
|
while (offset <= endOffset)
|
|
{
|
|
const char* nextCstr = static_cast<const char*>(::memchr(cstr + offset, sep, endOffset - offset));
|
|
const size_t nextOffset = nextCstr ? nextCstr - cstr : endOffset;
|
|
|
|
if (offset < nextOffset || token_policy == empty_token_policy::keep)
|
|
{
|
|
token.clear();
|
|
token.append(cstr + offset, cstr + nextOffset);
|
|
co_yield token;
|
|
}
|
|
offset = nextOffset + 1;
|
|
}
|
|
}
|
|
|
|
template <class String>
|
|
cppcoro::generator<const String&> split_on_any(const String& str, tl::string_view sep, empty_token_policy token_policy = empty_token_policy::discard) noexcept
|
|
{
|
|
const size_t sep_size = sep.size();
|
|
if (sep_size == 1)
|
|
{
|
|
for (auto const& s: split_on_any(str, sep[0], token_policy))
|
|
co_yield s;
|
|
co_return;
|
|
}
|
|
|
|
if (sep_size == 0)
|
|
{
|
|
TL_PLAIN_FAIL("'sep' cannot be empty");
|
|
co_return;
|
|
}
|
|
|
|
String token;
|
|
const char* cstr = str.data();
|
|
const char* sepStr = sep.data();
|
|
size_t offset = 0;
|
|
const size_t endOffset = str.size();
|
|
while (offset <= endOffset)
|
|
{
|
|
size_t nextOffset = offset;
|
|
for (; nextOffset < endOffset; ++nextOffset)
|
|
{
|
|
if (::memchr(sepStr, cstr[nextOffset], sep_size) != nullptr)
|
|
break;
|
|
}
|
|
|
|
if (offset < nextOffset || token_policy == empty_token_policy::keep)
|
|
{
|
|
token.clear();
|
|
token.append(cstr + offset, cstr + nextOffset);
|
|
co_yield token;
|
|
}
|
|
offset = nextOffset + 1;
|
|
}
|
|
}
|
|
|
|
template <class String>
|
|
cppcoro::generator<const String&> split_on_all(const String& str, tl::string_view sep, empty_token_policy token_policy = empty_token_policy::discard) noexcept
|
|
{
|
|
if (sep.empty())
|
|
{
|
|
TL_PLAIN_FAIL("'sep' cannot be empty");
|
|
co_return;
|
|
}
|
|
|
|
size_t p0 = 0;
|
|
size_t p1 = 0;
|
|
const String empty_token;
|
|
String token;
|
|
while ((p1 = str.find(sep, p0)) != String::npos)
|
|
{
|
|
if (p1 > p0)
|
|
{
|
|
token.clear();
|
|
token.append(str, p0, p1 - p0);
|
|
co_yield token;
|
|
}
|
|
else if (token_policy == empty_token_policy::keep)
|
|
co_yield empty_token;
|
|
|
|
p0 = p1 + sep.size();
|
|
}
|
|
|
|
if (str.size() > p0)
|
|
{
|
|
token.clear();
|
|
token.append(str, p0, String::npos);
|
|
co_yield token;
|
|
}
|
|
else if (token_policy == empty_token_policy::keep)
|
|
co_yield empty_token;
|
|
}
|
|
|
|
template <class String>
|
|
String left(const String& text, size_t length) noexcept
|
|
{
|
|
size_t textSize = text.size();
|
|
return text.substr(0, length > textSize ? textSize : length);
|
|
}
|
|
|
|
template <class String>
|
|
String right(const String& text, size_t length) noexcept
|
|
{
|
|
const size_t textSize = text.size();
|
|
size_t startPos = textSize >= length ? textSize - length : 0;
|
|
return text.substr(startPos);
|
|
}
|
|
|
|
template <class String>
|
|
bool starts_with(const String& text, const String& prefix) noexcept
|
|
{
|
|
const size_t textSize = text.size();
|
|
const size_t prefixSize = prefix.size();
|
|
if (textSize < prefixSize)
|
|
return false;
|
|
|
|
return memcmp(text.data(), prefix.data(), prefixSize) == 0;
|
|
}
|
|
|
|
template <class String>
|
|
bool starts_with_ci(const String& text, const String& prefix) noexcept
|
|
{
|
|
const size_t textSize = text.size();
|
|
const size_t prefixSize = prefix.size();
|
|
if (textSize < prefixSize)
|
|
return false;
|
|
|
|
return ascii::memicmp(text.data(), prefix.data(), prefixSize) == 0;
|
|
}
|
|
|
|
template <class String>
|
|
bool ends_with(const String& text, const String& suffix) noexcept
|
|
{
|
|
const size_t textSize = text.size();
|
|
const size_t suffixSize = suffix.size();
|
|
if (textSize < suffixSize)
|
|
return false;
|
|
|
|
return memcmp(text.data() + (textSize - suffixSize), suffix.data(), suffixSize) == 0;
|
|
}
|
|
|
|
template <class String>
|
|
bool ends_with_ci(const String& text, const String& suffix) noexcept
|
|
{
|
|
const size_t textSize = text.size();
|
|
const size_t suffixSize = suffix.size();
|
|
if (textSize < suffixSize)
|
|
return false;
|
|
|
|
return ascii::memicmp(text.data() + (textSize - suffixSize), suffix.data(), suffixSize) == 0;
|
|
}
|
|
|
|
template <class String>
|
|
String remove_prefix(const String& text, const String& prefix) noexcept
|
|
{
|
|
if (starts_with(text, prefix))
|
|
return text.substr(prefix.size(), text.size() - prefix.size());
|
|
|
|
return text;
|
|
}
|
|
|
|
template <class String>
|
|
String remove_prefix_ci(const String& text, const String& prefix) noexcept
|
|
{
|
|
if (starts_with_ci(text, prefix))
|
|
return text.substr(prefix.size(), text.size() - prefix.size());
|
|
|
|
return text;
|
|
}
|
|
|
|
template <class String>
|
|
String remove_suffix(const String& text, const String& suffix) noexcept
|
|
{
|
|
if (ends_with(text, suffix))
|
|
return text.substr(0, text.size() - suffix.size());
|
|
|
|
return text;
|
|
}
|
|
|
|
template <class String>
|
|
String remove_suffix_ci(const String& text, const String& suffix) noexcept
|
|
{
|
|
if (ends_with_ci(text, suffix))
|
|
return text.substr(0, text.size() - suffix.size());
|
|
|
|
return text;
|
|
}
|
|
|
|
template <class String>
|
|
eastl::optional<size_t> contains(const String& text, const String& subText) noexcept
|
|
{
|
|
const size_t offset = text.find(subText);
|
|
if (offset != String::npos)
|
|
return offset;
|
|
|
|
return nullopt;
|
|
}
|
|
|
|
template <class String>
|
|
eastl::optional<size_t> contains_ci(const String& text, const String& subText) noexcept
|
|
{
|
|
const size_t offset = text.find_ci(subText);
|
|
if (offset != String::npos)
|
|
return offset;
|
|
|
|
return nullopt;
|
|
}
|
|
|
|
template <class String>
|
|
String quote(const String& text, const String& quote) noexcept
|
|
{
|
|
String result = quote;
|
|
result += text;
|
|
result += quote;
|
|
return result;
|
|
}
|
|
|
|
template <class String>
|
|
String unquote(const String& text, const String& quote) noexcept
|
|
{
|
|
if (starts_with(text, quote) && ends_with(text, quote) && text.size() >= quote.size() * 2)
|
|
return text.substr(quote.size(), text.size() - quote.size() * 2);
|
|
|
|
return text;
|
|
}
|
|
|
|
static const char* s_defaultWhiteSpaceCStr = " \t\f\v\n\r";
|
|
|
|
template <class String>
|
|
String trim_left_any_of(const String& text, const String& characters) noexcept
|
|
{
|
|
const typename String::size_type startPos = text.find_first_not_of(characters);
|
|
if (startPos == String::npos)
|
|
return String();
|
|
|
|
return text.substr(startPos);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_left_any_of_ci(const String& text, const String& characters) noexcept
|
|
{
|
|
const typename String::size_type startPos = text.find_first_not_of_ci(characters);
|
|
if (startPos == String::npos)
|
|
return String();
|
|
|
|
return text.substr(startPos);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_left(const String& text) noexcept
|
|
{
|
|
static String s_defaultWhiteSpace(s_defaultWhiteSpaceCStr);
|
|
return trim_left_any_of(text, s_defaultWhiteSpace);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_left_ci(const String& text) noexcept
|
|
{
|
|
static String s_defaultWhiteSpace(s_defaultWhiteSpaceCStr);
|
|
return trim_left_any_of_ci(text, s_defaultWhiteSpace);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_right_any_of(const String& text, const String& characters) noexcept
|
|
{
|
|
const typename String::size_type startPos = text.find_last_not_of(characters);
|
|
if (startPos == String::npos)
|
|
return String();
|
|
|
|
return text.substr(0, startPos + 1);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_right_any_of_ci(const String& text, const String& characters) noexcept
|
|
{
|
|
const typename String::size_type startPos = text.find_last_not_of_ci(characters);
|
|
if (startPos == String::npos)
|
|
return String();
|
|
|
|
return text.substr(0, startPos + 1);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_right(const String& text) noexcept
|
|
{
|
|
static String s_defaultWhiteSpace(s_defaultWhiteSpaceCStr);
|
|
return trim_right_any_of(text, String(s_defaultWhiteSpace));
|
|
}
|
|
|
|
template <class String>
|
|
String trim_right_ci(const String& text) noexcept
|
|
{
|
|
static String s_defaultWhiteSpace(s_defaultWhiteSpaceCStr);
|
|
return trim_right_any_of_ci(text, String(s_defaultWhiteSpace));
|
|
}
|
|
|
|
template <class String>
|
|
String trim(const String& text) noexcept
|
|
{
|
|
return trim_right(trim_left(text));
|
|
}
|
|
|
|
template <class String>
|
|
String trim_ci(const String& text) noexcept
|
|
{
|
|
return trim_right_ci(trim_left_ci(text));
|
|
}
|
|
|
|
template <class String>
|
|
String trim_any_of(const String& text, const String& characters) noexcept
|
|
{
|
|
return trim_right_any_of(trim_left_any_of(text, characters), characters);
|
|
}
|
|
|
|
template <class String>
|
|
String trim_any_of_ci(const String& text, const String& characters) noexcept
|
|
{
|
|
return trim_right_any_of_ci(trim_left_any_of_ci(text, characters), characters);
|
|
}
|
|
|
|
template <class String, class RepString>
|
|
String replace(const String& text, const String& target, const function<eastl::optional<RepString>(size_t, size_t)>& functor) noexcept
|
|
{
|
|
if (target.empty())
|
|
return text;
|
|
|
|
eastl::fixed_vector<size_t, 512> positions;
|
|
|
|
size_t position = 0;
|
|
while ((position = text.find(target, position)) != String::npos)
|
|
{
|
|
positions.push_back(position);
|
|
position += target.length();
|
|
}
|
|
|
|
String result;
|
|
result.reserve(text.size());
|
|
|
|
auto start = text.begin();
|
|
for (auto it = positions.begin(); it != positions.end(); ++it)
|
|
{
|
|
const size_t position = *it;
|
|
size_t index = eastl::distance(positions.begin(), it);
|
|
|
|
eastl::optional<RepString> replacementString = functor(index, positions.size());
|
|
if (!replacementString)
|
|
continue;
|
|
|
|
auto end = text.begin() + position;
|
|
result.append(start, end); //copy up to the result
|
|
result.append(*replacementString);
|
|
start = end + target.size();
|
|
}
|
|
|
|
if (start != text.end())
|
|
result.append(start, text.end()); //copy the end run
|
|
|
|
return result;
|
|
}
|
|
|
|
template <class String, typename Fun>
|
|
String replace(const String& text, const String& target, Fun f) noexcept
|
|
{
|
|
return replace(text, target, function<eastl::optional<String>(size_t, size_t)>(f));
|
|
}
|
|
|
|
template <class Func>
|
|
size_t replace_first(const string& text, const string& target, const string& replacement, size_t startingOffset, const Func& functor) noexcept
|
|
{
|
|
if (target.empty())
|
|
{
|
|
functor(text);
|
|
return string::npos;
|
|
}
|
|
|
|
const size_t startPos = text.find(target, startingOffset);
|
|
if (startPos == string::npos)
|
|
{
|
|
functor(text);
|
|
return string::npos;
|
|
}
|
|
|
|
eastl::string result(text.eastl_str());
|
|
string res2(result.replace(startPos, target.length(), eastl::string(replacement.eastl_str())));
|
|
|
|
functor(res2);
|
|
return startPos + replacement.length();
|
|
}
|
|
|
|
namespace detail
|
|
{
|
|
template <class Container, class String>
|
|
String join_worker(const Container& container, const String& sep) noexcept
|
|
{
|
|
String dst;
|
|
//reserve some space based on some heuristics:
|
|
// We'll have size - 1 seps
|
|
// And an average string size of 2 chars. The small choice is made to help with the small string case only where the allocator overhead is significant.
|
|
dst.reserve((sep.size() + 2) * container.size());
|
|
|
|
auto end = std::end(container);
|
|
auto it = std::begin(container);
|
|
while (true)
|
|
{
|
|
dst += *it;
|
|
++it;
|
|
if (it == end)
|
|
break;
|
|
|
|
dst += sep;
|
|
}
|
|
return dst;
|
|
}
|
|
|
|
template <class Container>
|
|
string join_worker(const Container& container, const string& sep) noexcept
|
|
{
|
|
eastl::string buffer;
|
|
//reserve some space based on some heuristics:
|
|
// We'll have size - 1 seps
|
|
// And an average string size of 2 chars. The small choice is made to help with the small string case only where the allocator overhead is significant.
|
|
buffer.reserve((sep.size() + 2) * container.size());
|
|
|
|
auto end = std::end(container);
|
|
auto it = std::begin(container);
|
|
while (true)
|
|
{
|
|
const string& s = *it;
|
|
buffer.append(s.data(), s.size());
|
|
++it;
|
|
if (it == end)
|
|
break;
|
|
|
|
buffer.append(sep.data(), sep.size());
|
|
}
|
|
return string(buffer);
|
|
}
|
|
|
|
//joiner used when the Sep and String are of the same type
|
|
template <class Container, class Sep, class String>
|
|
std::enable_if_t<
|
|
std::is_same_v<
|
|
std::remove_const_t<std::remove_reference_t<String>>,
|
|
std::remove_const_t<std::remove_reference_t<Sep>>>,
|
|
String> join(const Container& container, const Sep& sep) noexcept
|
|
{
|
|
return join_worker(container, sep);
|
|
}
|
|
|
|
//joiner used when the Sep and String are different types
|
|
template <class Container, class Sep, class String>
|
|
std::enable_if_t<
|
|
!std::is_same_v<
|
|
std::remove_const_t<std::remove_reference_t<String>>,
|
|
std::remove_const_t<std::remove_reference_t<Sep>>>,
|
|
String> join(const Container& container, const Sep& sep) noexcept
|
|
{
|
|
return join_worker(container, String(sep));
|
|
}
|
|
} //namespace detail
|
|
|
|
template <class Container, class Sep, class String>
|
|
String join(const Container& container, const Sep& sep) noexcept
|
|
{
|
|
if (container.empty())
|
|
return String();
|
|
|
|
return detail::join<Container, Sep, String>(container, sep);
|
|
}
|
|
}
|
|
}
|