Compare commits
4
Commits
codex-test
..
1.0
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
6ed48022ae | ||
|
|
7026e246bb | ||
|
|
804876656d | ||
|
|
e5cb11c659 |
No files matched your search
+31
-1
@@ -1,11 +1,34 @@
|
||||
cmake_minimum_required(VERSION 3.16)
|
||||
|
||||
project(kindle2notes LANGUAGES CXX)
|
||||
project(kindle2notes VERSION 1.0 LANGUAGES CXX)
|
||||
|
||||
set(CMAKE_CXX_STANDARD 17)
|
||||
set(CMAKE_CXX_STANDARD_REQUIRED ON)
|
||||
set(CMAKE_CXX_EXTENSIONS OFF)
|
||||
|
||||
set(KINDLE2NOTES_VERSION "${PROJECT_VERSION}")
|
||||
find_package(Git QUIET)
|
||||
if(Git_FOUND)
|
||||
execute_process(
|
||||
COMMAND ${GIT_EXECUTABLE} describe --tags --long --dirty --always
|
||||
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
|
||||
RESULT_VARIABLE KINDLE2NOTES_GIT_RESULT
|
||||
OUTPUT_VARIABLE KINDLE2NOTES_GIT_DESCRIBE
|
||||
OUTPUT_STRIP_TRAILING_WHITESPACE
|
||||
ERROR_QUIET
|
||||
)
|
||||
|
||||
if(KINDLE2NOTES_GIT_RESULT EQUAL 0 AND NOT KINDLE2NOTES_GIT_DESCRIBE STREQUAL "")
|
||||
if(KINDLE2NOTES_GIT_DESCRIBE MATCHES "^(.+)-0-g[0-9a-f]+$")
|
||||
set(KINDLE2NOTES_VERSION "${CMAKE_MATCH_1}")
|
||||
else()
|
||||
set(KINDLE2NOTES_VERSION "${KINDLE2NOTES_GIT_DESCRIBE}")
|
||||
endif()
|
||||
endif()
|
||||
endif()
|
||||
|
||||
message(STATUS "kindle2notes version: ${KINDLE2NOTES_VERSION}")
|
||||
|
||||
function(kindle2notes_enable_warnings target)
|
||||
if(MSVC)
|
||||
target_compile_options(${target} PRIVATE /W4)
|
||||
@@ -15,13 +38,20 @@ function(kindle2notes_enable_warnings target)
|
||||
endfunction()
|
||||
|
||||
add_executable(kindle2notes
|
||||
typefactory/convertasciidoc.cpp
|
||||
typefactory/convertasciidoc.h
|
||||
typefactory/convertlatex.cpp
|
||||
typefactory/convertlatex.h
|
||||
typefactory/convertmarkdown.cpp
|
||||
typefactory/convertmarkdown.h
|
||||
typefactory/convertorg.cpp
|
||||
typefactory/convertorg.h
|
||||
typefactory/converter.cpp
|
||||
typefactory/converter.h
|
||||
typefactory/main.cpp
|
||||
typefactory/typefactory.h
|
||||
)
|
||||
kindle2notes_enable_warnings(kindle2notes)
|
||||
target_compile_definitions(kindle2notes PRIVATE
|
||||
KINDLE2NOTES_VERSION="${KINDLE2NOTES_VERSION}"
|
||||
)
|
||||
@@ -7,6 +7,8 @@ Supported output formats:
|
||||
|
||||
- LaTeX (`--type=latex`)
|
||||
- Markdown (`--type=md`)
|
||||
- Org mode (`--type=org`)
|
||||
- AsciiDoc (`--type=adoc`)
|
||||
|
||||
Highlights are grouped by book title while preserving the order in which
|
||||
books first appear in the input file.
|
||||
@@ -63,6 +65,43 @@ Markdown block quote, followed by an empty notes heading:
|
||||
Empty lines inside highlights are preserved as `>` lines, and Markdown syntax
|
||||
characters are escaped automatically.
|
||||
|
||||
### Org mode
|
||||
|
||||
Books and clipping metadata become Org headings. Highlight text is placed in
|
||||
a quote block followed by an empty notes heading:
|
||||
|
||||
```org
|
||||
* Book Title
|
||||
|
||||
** - Your Highlight on Location 123-125 | Added on ...
|
||||
|
||||
#+begin_quote
|
||||
Highlighted text line 1
|
||||
Highlighted text line 2
|
||||
#+end_quote
|
||||
|
||||
*** Notes
|
||||
```
|
||||
|
||||
### AsciiDoc
|
||||
|
||||
Books and clipping metadata become section headings. Highlight text is placed
|
||||
in a quote block followed by an empty notes section:
|
||||
|
||||
```adoc
|
||||
== Book Title
|
||||
|
||||
=== - Your Highlight on Location 123-125 | Added on ...
|
||||
|
||||
[quote]
|
||||
____
|
||||
Highlighted text line 1
|
||||
Highlighted text line 2
|
||||
____
|
||||
|
||||
==== Notes
|
||||
```
|
||||
|
||||
## Architecture
|
||||
|
||||
Input parsing is implemented as a stage-driven pipeline with one handler for
|
||||
@@ -92,6 +131,12 @@ The resulting executable is:
|
||||
build/kindle2notes
|
||||
```
|
||||
|
||||
When the source is built from Git, CMake obtains the version from
|
||||
`git describe`. A build made exactly from a release tag reports only that tag
|
||||
(for example, `1.0`). Later commits include their distance from the tag and
|
||||
the abbreviated commit hash (for example, `1.0-3-gabc1234`). A modified
|
||||
working tree also receives the `-dirty` suffix.
|
||||
|
||||
## Usage
|
||||
|
||||
Generate LaTeX:
|
||||
@@ -112,17 +157,42 @@ Generate Markdown:
|
||||
--type=md
|
||||
```
|
||||
|
||||
Generate Org mode:
|
||||
|
||||
```bash
|
||||
./build/kindle2notes \
|
||||
--input "My Clippings.txt" \
|
||||
--output notes.org \
|
||||
--type=org
|
||||
```
|
||||
|
||||
Generate AsciiDoc:
|
||||
|
||||
```bash
|
||||
./build/kindle2notes \
|
||||
--input "My Clippings.txt" \
|
||||
--output notes.adoc \
|
||||
--type=adoc
|
||||
```
|
||||
|
||||
Short options are also available:
|
||||
|
||||
```bash
|
||||
./build/kindle2notes -i input.txt -o notes.md -t md
|
||||
```
|
||||
|
||||
Show the version compiled into the executable:
|
||||
|
||||
```bash
|
||||
./build/kindle2notes --version
|
||||
```
|
||||
|
||||
| Argument | Short form | Description |
|
||||
| --- | --- | --- |
|
||||
| `--input` | `-i` | Path to the Kindle clippings file |
|
||||
| `--output` | `-o` | Path to the generated file |
|
||||
| `--type` | `-t` | Output format: `latex` or `md` |
|
||||
| `--type` | `-t` | Output format: `latex`, `md`, `org`, or `adoc` |
|
||||
| `--version` | `-v` | Show the compiled version and exit |
|
||||
| `--help` | `-h` | Show usage information |
|
||||
|
||||
If an unsupported output type is requested, the program prints the list of
|
||||
|
||||
@@ -1,295 +0,0 @@
|
||||
#pragma once
|
||||
|
||||
#include <type_traits>
|
||||
#include <utility>
|
||||
#include <variant>
|
||||
|
||||
/** @brief Missing transition policy. */
|
||||
enum class missing_transition_policy {
|
||||
/** @brief Missing transition is considered required by validation. */
|
||||
strict,
|
||||
|
||||
/** @brief Missing transition is allowed (event ignored). */
|
||||
ignore
|
||||
};
|
||||
|
||||
/** @brief Marker for undefined transition. */
|
||||
struct no_transition {};
|
||||
|
||||
/** @brief Default guard: always allow. */
|
||||
struct always_allow {
|
||||
template <typename TContext, typename TFrom, typename TEvent>
|
||||
bool operator() (TContext &, const TFrom &, const TEvent &) const noexcept {
|
||||
return true;
|
||||
}
|
||||
};
|
||||
|
||||
/** @brief Default action: do nothing. */
|
||||
struct do_nothing {
|
||||
template <typename TContext, typename TFrom, typename TTo, typename TEvent>
|
||||
void operator() (TContext &, const TFrom &, const TTo &, const TEvent &) const noexcept {
|
||||
}
|
||||
};
|
||||
|
||||
/**
|
||||
* @brief Transition descriptor.
|
||||
*
|
||||
* @tparam TTo Destination state type.
|
||||
* @tparam TGuard Guard functor type.
|
||||
* @tparam TAction Action functor type.
|
||||
*/
|
||||
template <typename TTo, typename TGuard = always_allow, typename TAction = do_nothing>
|
||||
struct transition_to {
|
||||
using to_state = TTo;
|
||||
using guard = TGuard;
|
||||
using action = TAction;
|
||||
};
|
||||
|
||||
/**
|
||||
* @brief Transition table mapping: specialize to define transitions.
|
||||
* Default: no_transition.
|
||||
*/
|
||||
template <typename TFrom, typename TEvent>
|
||||
struct transition {
|
||||
using type = no_transition;
|
||||
};
|
||||
|
||||
/** @brief Optional mapping to names for logs. */
|
||||
template <typename T>
|
||||
struct name_of {
|
||||
static constexpr const char *value = "?";
|
||||
};
|
||||
|
||||
/**
|
||||
* @brief Logger interface (default no-op).
|
||||
*
|
||||
* You can implement:
|
||||
* - ignored_transition(from, event)
|
||||
* - unhandled_transition(from, event) // policy=strict but missing transition at runtime
|
||||
* - guard_blocked(from, event)
|
||||
*/
|
||||
struct null_logger {
|
||||
void ignored_transition (const char *, const char *) noexcept {}
|
||||
void unhandled_transition (const char *, const char *) noexcept {}
|
||||
void guard_blocked (const char *, const char *) noexcept {}
|
||||
};
|
||||
|
||||
/** @brief Hooks: state-only. */
|
||||
template <typename TState>
|
||||
struct on_exit_state {
|
||||
template <typename TContext>
|
||||
static void call (TContext &, const TState &) noexcept {}
|
||||
};
|
||||
|
||||
template <typename TState>
|
||||
struct on_entry_state {
|
||||
template <typename TContext>
|
||||
static void call (TContext &, const TState &) noexcept {}
|
||||
};
|
||||
|
||||
/** @brief Hooks: event-specific. */
|
||||
template <typename TState, typename TEvent>
|
||||
struct on_exit_event {
|
||||
template <typename TContext>
|
||||
static void call (TContext &, const TState &, const TEvent &) noexcept {}
|
||||
};
|
||||
|
||||
template <typename TState, typename TEvent>
|
||||
struct on_entry_event {
|
||||
template <typename TContext>
|
||||
static void call (TContext &, const TState &, const TEvent &) noexcept {}
|
||||
};
|
||||
|
||||
/**
|
||||
* @brief Factory: construct destination state.
|
||||
*
|
||||
* IMPORTANT: Now includes TContext in template parameters to avoid templated make<TContext>().
|
||||
*
|
||||
* Default: To{}.
|
||||
* Specialize state_factory<TContext, From, Event, To> as needed.
|
||||
*/
|
||||
template <typename TContext, typename TFrom, typename TEvent, typename TTo>
|
||||
struct state_factory {
|
||||
static TTo make (TContext &, const TFrom &, const TEvent &) {
|
||||
return TTo{};
|
||||
}
|
||||
};
|
||||
|
||||
/**
|
||||
* @brief Per (State,Event) missing transition policy.
|
||||
*
|
||||
* Default is ignore. Mark only required pairs strict.
|
||||
*/
|
||||
template <typename TFrom, typename TEvent>
|
||||
struct missing_transition {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::ignore;
|
||||
};
|
||||
|
||||
/** @brief Helper: is T one of Ts... */
|
||||
template <typename T, typename... Ts>
|
||||
struct is_one_of : std::disjunction<std::is_same<T, Ts>...> {};
|
||||
|
||||
/** @brief Helper: transition exists? */
|
||||
template <typename TFrom, typename TEvent>
|
||||
struct has_transition
|
||||
: std::bool_constant < !std::is_same_v<typename transition<TFrom, TEvent>::type, no_transition >> {};
|
||||
|
||||
/** @brief Factory validation (call-form). */
|
||||
template <typename TContext, typename TFrom, typename TEvent, typename TTo, typename = void>
|
||||
struct is_factory_invocable : std::false_type {};
|
||||
|
||||
template <typename TContext, typename TFrom, typename TEvent, typename TTo>
|
||||
struct is_factory_invocable <
|
||||
TContext, TFrom, TEvent, TTo,
|
||||
std::void_t<decltype (state_factory<TContext, TFrom, TEvent, TTo>::make (
|
||||
std::declval<TContext &>(),
|
||||
std::declval<const TFrom &>(),
|
||||
std::declval<const TEvent &>())) >>
|
||||
: std::bool_constant<std::is_same_v<
|
||||
decltype (state_factory<TContext, TFrom, TEvent, TTo>::make (
|
||||
std::declval<TContext &>(),
|
||||
std::declval<const TFrom &>(),
|
||||
std::declval<const TEvent &>())),
|
||||
TTo>> {};
|
||||
|
||||
/**
|
||||
* @brief FSM with runtime state storage and explicit compile-time validation.
|
||||
*
|
||||
* dispatch():
|
||||
* - never static_asserts on missing transitions (avoids State×Event explosion with std::visit)
|
||||
* - handles missing transitions via missing_transition<State,Event>::policy (ignore/strict)
|
||||
*
|
||||
* validate_events<Ev...>():
|
||||
* - compile-time checks ONLY for pairs marked strict.
|
||||
*/
|
||||
template <typename TContext, typename TLogger, typename... TStates>
|
||||
class fsm {
|
||||
public:
|
||||
using state_variant = std::variant<TStates...>;
|
||||
|
||||
template <typename TInitial,
|
||||
typename = std::enable_if_t<is_one_of<TInitial, TStates...>::value>>
|
||||
explicit fsm (TContext &ctx, TInitial initial, TLogger logger = TLogger{})
|
||||
: m_ctx (ctx)
|
||||
, m_logger (std::move (logger))
|
||||
, m_state (std::move (initial)) {
|
||||
std::visit (init_entry_visitor{*this}, m_state);
|
||||
}
|
||||
|
||||
template <typename TEvent>
|
||||
bool dispatch (const TEvent &ev) {
|
||||
dispatch_visitor<TEvent> v{*this, ev};
|
||||
return std::visit (v, m_state);
|
||||
}
|
||||
|
||||
template <typename... TEvents>
|
||||
static constexpr void validate_events() {
|
||||
(validate_one_event<TEvents>(), ...);
|
||||
}
|
||||
|
||||
const state_variant &state() const noexcept {
|
||||
return m_state;
|
||||
}
|
||||
state_variant &state() noexcept {
|
||||
return m_state;
|
||||
}
|
||||
|
||||
private:
|
||||
template <typename TEvent>
|
||||
struct dispatch_visitor {
|
||||
fsm &self;
|
||||
const TEvent &ev;
|
||||
|
||||
template <typename TFrom>
|
||||
bool operator() (TFrom &from) const {
|
||||
return self.template dispatch_from<TFrom> (from, ev);
|
||||
}
|
||||
};
|
||||
|
||||
struct init_entry_visitor {
|
||||
fsm &self;
|
||||
template <typename TState>
|
||||
void operator() (TState &st) const {
|
||||
on_entry_state<TState>::call (self.m_ctx, st);
|
||||
}
|
||||
};
|
||||
|
||||
template <typename TEvent>
|
||||
struct entry_after_commit_visitor {
|
||||
fsm &self;
|
||||
const TEvent &ev;
|
||||
|
||||
template <typename TState>
|
||||
void operator() (TState &st) const {
|
||||
on_entry_state<TState>::call (self.m_ctx, st);
|
||||
on_entry_event<TState, TEvent>::call (self.m_ctx, st, ev);
|
||||
}
|
||||
};
|
||||
|
||||
template <typename TFrom, typename TEvent>
|
||||
bool dispatch_from (TFrom &from, const TEvent &ev) {
|
||||
using tr = typename transition<TFrom, TEvent>::type;
|
||||
|
||||
if constexpr (std::is_same_v<tr, no_transition>) {
|
||||
if constexpr (missing_transition<TFrom, TEvent>::policy == missing_transition_policy::strict)
|
||||
m_logger.unhandled_transition (name_of<TFrom>::value, name_of<TEvent>::value);
|
||||
else
|
||||
m_logger.ignored_transition (name_of<TFrom>::value, name_of<TEvent>::value);
|
||||
return false;
|
||||
} else {
|
||||
using to_state = typename tr::to_state;
|
||||
using guard_t = typename tr::guard;
|
||||
using action_t = typename tr::action;
|
||||
|
||||
static_assert (is_one_of<to_state, TStates...>::value,
|
||||
"FSM: transition target state not in FSM state list.");
|
||||
|
||||
static_assert (std::is_invocable_r_v<bool, guard_t, TContext &, const TFrom &, const TEvent &>,
|
||||
"FSM: guard must be callable as bool(Context&, const From&, const Event&).");
|
||||
|
||||
static_assert (std::is_invocable_r_v<void, action_t, TContext &, const TFrom &, const to_state &, const TEvent &>,
|
||||
"FSM: action must be callable as void(Context&, const From&, const To&, const Event&).");
|
||||
|
||||
static_assert (is_factory_invocable<TContext, TFrom, TEvent, to_state>::value,
|
||||
"FSM: state_factory<Context,From,Event,To>::make must be callable as "
|
||||
"To make(Context&, const From&, const Event&).");
|
||||
|
||||
guard_t guard{};
|
||||
if (!guard (m_ctx, from, ev)) {
|
||||
m_logger.guard_blocked (name_of<TFrom>::value, name_of<TEvent>::value);
|
||||
return false;
|
||||
}
|
||||
|
||||
on_exit_event<TFrom, TEvent>::call (m_ctx, from, ev);
|
||||
on_exit_state<TFrom>::call (m_ctx, from);
|
||||
|
||||
to_state to = state_factory<TContext, TFrom, TEvent, to_state>::make (m_ctx, from, ev);
|
||||
|
||||
action_t action{};
|
||||
action (m_ctx, from, to, ev);
|
||||
|
||||
m_state = std::move (to);
|
||||
|
||||
std::visit (entry_after_commit_visitor<TEvent> {*this, ev}, m_state);
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
template <typename TEvent>
|
||||
static constexpr void validate_one_event() {
|
||||
(validate_pair<TStates, TEvent>(), ...);
|
||||
}
|
||||
|
||||
template <typename TState, typename TEvent>
|
||||
static constexpr void validate_pair() {
|
||||
if constexpr (missing_transition<TState, TEvent>::policy == missing_transition_policy::strict) {
|
||||
static_assert (has_transition<TState, TEvent>::value,
|
||||
"FSM validation: required (strict) transition is missing for (State, Event).");
|
||||
}
|
||||
}
|
||||
|
||||
private:
|
||||
TContext &m_ctx;
|
||||
TLogger m_logger;
|
||||
state_variant m_state;
|
||||
};
|
||||
@@ -1,10 +0,0 @@
|
||||
TEMPLATE = app
|
||||
CONFIG += console c++17
|
||||
CONFIG -= app_bundle
|
||||
CONFIG -= qt
|
||||
|
||||
SOURCES += \
|
||||
main.cpp
|
||||
|
||||
HEADERS += \
|
||||
fsm.h
|
||||
-372
@@ -1,372 +0,0 @@
|
||||
#include "fsm.h"
|
||||
|
||||
#include <cerrno>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
#include <fstream>
|
||||
#include <getopt.h>
|
||||
#include <iostream>
|
||||
#include <string>
|
||||
#include <unordered_map>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
// ------------------------------------------------------------
|
||||
// Data model
|
||||
// ------------------------------------------------------------
|
||||
|
||||
struct Quote {
|
||||
std::string meta;
|
||||
std::vector<std::string> text_lines;
|
||||
};
|
||||
|
||||
static std::string rstrip_cr (std::string s) {
|
||||
if (!s.empty() && s.back() == '\r')
|
||||
s.pop_back();
|
||||
return s;
|
||||
}
|
||||
|
||||
static std::string latex_escape (const std::string &s) {
|
||||
// Conservative escaping for LaTeX special chars.
|
||||
std::string out;
|
||||
out.reserve (s.size() + s.size() / 8);
|
||||
for (const unsigned char ch : s) {
|
||||
switch (ch) {
|
||||
case '\\':
|
||||
out += "\\textbackslash{}";
|
||||
break;
|
||||
case '{':
|
||||
out += "\\{";
|
||||
break;
|
||||
case '}':
|
||||
out += "\\}";
|
||||
break;
|
||||
case '%':
|
||||
out += "\\%";
|
||||
break;
|
||||
case '$':
|
||||
out += "\\$";
|
||||
break;
|
||||
case '#':
|
||||
out += "\\#";
|
||||
break;
|
||||
case '&':
|
||||
out += "\\&";
|
||||
break;
|
||||
case '_':
|
||||
out += "\\_";
|
||||
break;
|
||||
case '^':
|
||||
out += "\\textasciicircum{}";
|
||||
break;
|
||||
case '~':
|
||||
out += "\\textasciitilde{}";
|
||||
break;
|
||||
default:
|
||||
out.push_back (static_cast<char> (ch));
|
||||
break;
|
||||
}
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
// ------------------------------------------------------------
|
||||
// FSM: parsing Kindle "My Clippings" style format
|
||||
// ------------------------------------------------------------
|
||||
|
||||
struct StExpectTitle {};
|
||||
struct StExpectMeta {};
|
||||
struct StExpectBlank {};
|
||||
struct StCollectText {};
|
||||
|
||||
struct EvTitle {
|
||||
std::string s;
|
||||
};
|
||||
struct EvMeta {
|
||||
std::string s;
|
||||
};
|
||||
struct EvBlank {};
|
||||
struct EvText {
|
||||
std::string s;
|
||||
};
|
||||
struct EvSep {};
|
||||
struct EvEof {};
|
||||
|
||||
struct ParserContext {
|
||||
// group storage
|
||||
std::vector<std::string> order;
|
||||
std::unordered_map<std::string, std::vector<Quote>> by_title;
|
||||
std::unordered_map<std::string, std::size_t> seen;
|
||||
|
||||
// currently parsed quote
|
||||
std::string cur_title;
|
||||
std::string cur_meta;
|
||||
std::vector<std::string> cur_text;
|
||||
|
||||
void start_title (std::string t) {
|
||||
cur_title = std::move (t);
|
||||
cur_meta.clear();
|
||||
cur_text.clear();
|
||||
}
|
||||
void set_meta (std::string m) {
|
||||
cur_meta = std::move (m);
|
||||
}
|
||||
void add_text (std::string line) {
|
||||
cur_text.push_back (std::move (line));
|
||||
}
|
||||
|
||||
void finalize_quote_if_ready() {
|
||||
if (cur_title.empty() || cur_meta.empty()) {
|
||||
cur_meta.clear();
|
||||
cur_text.clear();
|
||||
return;
|
||||
}
|
||||
|
||||
if (seen.find (cur_title) == seen.end()) {
|
||||
seen.emplace (cur_title, order.size());
|
||||
order.push_back (cur_title);
|
||||
}
|
||||
|
||||
by_title[cur_title].push_back (Quote{cur_meta, cur_text});
|
||||
|
||||
cur_meta.clear();
|
||||
cur_text.clear();
|
||||
}
|
||||
};
|
||||
|
||||
struct SilentLogger : null_logger {};
|
||||
|
||||
// Actions
|
||||
struct ActSetTitle {
|
||||
void operator() (ParserContext &ctx, const StExpectTitle &, const StExpectMeta &, const EvTitle &ev) const {
|
||||
ctx.start_title (ev.s);
|
||||
}
|
||||
};
|
||||
|
||||
struct ActSetMeta {
|
||||
void operator() (ParserContext &ctx, const StExpectMeta &, const StExpectBlank &, const EvMeta &ev) const {
|
||||
ctx.set_meta (ev.s);
|
||||
}
|
||||
};
|
||||
|
||||
struct ActAddText {
|
||||
template <typename TFrom>
|
||||
void operator() (ParserContext &ctx, const TFrom &, const StCollectText &, const EvText &ev) const {
|
||||
ctx.add_text (ev.s);
|
||||
}
|
||||
};
|
||||
|
||||
struct ActFinalizeOnSep {
|
||||
template <typename TFrom>
|
||||
void operator() (ParserContext &ctx, const TFrom &, const StExpectTitle &, const EvSep &) const {
|
||||
ctx.finalize_quote_if_ready();
|
||||
}
|
||||
};
|
||||
|
||||
struct ActFinalizeOnEof {
|
||||
template <typename TFrom>
|
||||
void operator() (ParserContext &ctx, const TFrom &, const StExpectTitle &, const EvEof &) const {
|
||||
ctx.finalize_quote_if_ready();
|
||||
}
|
||||
};
|
||||
|
||||
// Transitions
|
||||
template <> struct transition<StExpectTitle, EvTitle> {
|
||||
using type = transition_to<StExpectMeta, always_allow, ActSetTitle>;
|
||||
};
|
||||
template <> struct transition<StExpectTitle, EvBlank> {
|
||||
using type = transition_to<StExpectTitle>;
|
||||
};
|
||||
template <> struct transition<StExpectTitle, EvEof> {
|
||||
using type = transition_to<StExpectTitle, always_allow, ActFinalizeOnEof>;
|
||||
};
|
||||
|
||||
template <> struct transition<StExpectMeta, EvMeta> {
|
||||
using type = transition_to<StExpectBlank, always_allow, ActSetMeta>;
|
||||
};
|
||||
|
||||
template <> struct transition<StExpectBlank, EvBlank> {
|
||||
using type = transition_to<StCollectText>;
|
||||
};
|
||||
template <> struct transition<StExpectBlank, EvText> {
|
||||
using type = transition_to<StCollectText, always_allow, ActAddText>;
|
||||
};
|
||||
|
||||
template <> struct transition<StCollectText, EvText> {
|
||||
using type = transition_to<StCollectText, always_allow, ActAddText>;
|
||||
};
|
||||
template <> struct transition<StCollectText, EvSep> {
|
||||
using type = transition_to<StExpectTitle, always_allow, ActFinalizeOnSep>;
|
||||
};
|
||||
template <> struct transition<StCollectText, EvEof> {
|
||||
using type = transition_to<StExpectTitle, always_allow, ActFinalizeOnEof>;
|
||||
};
|
||||
|
||||
// Strict contract for the pairs we dispatch.
|
||||
template <> struct missing_transition<StExpectTitle, EvTitle> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
template <> struct missing_transition<StExpectTitle, EvBlank> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
template <> struct missing_transition<StExpectTitle, EvEof> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
|
||||
template <> struct missing_transition<StExpectMeta, EvMeta> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
|
||||
template <> struct missing_transition<StExpectBlank, EvBlank> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
template <> struct missing_transition<StExpectBlank, EvText> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
|
||||
template <> struct missing_transition<StCollectText, EvText> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
template <> struct missing_transition<StCollectText, EvSep> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
template <> struct missing_transition<StCollectText, EvEof> {
|
||||
static constexpr missing_transition_policy policy = missing_transition_policy::strict;
|
||||
};
|
||||
|
||||
using ParserFsm = fsm<ParserContext, SilentLogger, StExpectTitle, StExpectMeta, StExpectBlank, StCollectText>;
|
||||
|
||||
static constexpr void validate_fsm_contract() {
|
||||
ParserFsm::validate_events<EvTitle, EvMeta, EvBlank, EvText, EvSep, EvEof>();
|
||||
}
|
||||
|
||||
// ------------------------------------------------------------
|
||||
// CLI
|
||||
// ------------------------------------------------------------
|
||||
|
||||
struct CliArgs {
|
||||
std::string input;
|
||||
std::string output;
|
||||
};
|
||||
|
||||
static void print_usage (const char *argv0) {
|
||||
std::cerr << "Usage: " << argv0 << " --input <file> --output <file>\n";
|
||||
}
|
||||
|
||||
static bool parse_args (int argc, char **argv, CliArgs &out) {
|
||||
static option long_opts[] = {
|
||||
{"input", required_argument, nullptr, 'i'},
|
||||
{"output", required_argument, nullptr, 'o'},
|
||||
{"help", no_argument, nullptr, 'h'},
|
||||
{nullptr, 0, nullptr, 0 }
|
||||
};
|
||||
|
||||
int c = 0;
|
||||
while ((c = ::getopt_long (argc, argv, "i:o:h", long_opts, nullptr)) != -1) {
|
||||
switch (c) {
|
||||
case 'i':
|
||||
out.input = optarg;
|
||||
break;
|
||||
case 'o':
|
||||
out.output = optarg;
|
||||
break;
|
||||
case 'h':
|
||||
print_usage (argv[0]);
|
||||
return false;
|
||||
default:
|
||||
print_usage (argv[0]);
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
if (out.input.empty() || out.output.empty()) {
|
||||
print_usage (argv[0]);
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
// ------------------------------------------------------------
|
||||
// Conversion
|
||||
// ------------------------------------------------------------
|
||||
|
||||
static int convert (const std::string &in_path, const std::string &out_path) {
|
||||
std::ifstream in (in_path);
|
||||
if (!in.is_open()) {
|
||||
std::cerr << "Failed to open input file: " << in_path << " (" << std::strerror (errno) << ")\n";
|
||||
return 2;
|
||||
}
|
||||
|
||||
ParserContext ctx{};
|
||||
ParserFsm fsm (ctx, StExpectTitle{}, SilentLogger{});
|
||||
|
||||
std::string line;
|
||||
while (std::getline (in, line)) {
|
||||
line = rstrip_cr (std::move (line));
|
||||
|
||||
if (line == "==========") {
|
||||
fsm.dispatch (EvSep{});
|
||||
continue;
|
||||
}
|
||||
|
||||
const bool is_blank = line.empty();
|
||||
const auto &st = fsm.state();
|
||||
|
||||
if (std::holds_alternative<StExpectTitle> (st)) {
|
||||
if (is_blank)
|
||||
fsm.dispatch (EvBlank{});
|
||||
else
|
||||
fsm.dispatch (EvTitle{line});
|
||||
} else if (std::holds_alternative<StExpectMeta> (st)) {
|
||||
// Kindle format expects meta here. If missing, keep going.
|
||||
fsm.dispatch (EvMeta{line});
|
||||
} else if (std::holds_alternative<StExpectBlank> (st)) {
|
||||
if (is_blank)
|
||||
fsm.dispatch (EvBlank{});
|
||||
else
|
||||
fsm.dispatch (EvText{line}); // some clippings have no blank line
|
||||
} else {
|
||||
// StCollectText
|
||||
fsm.dispatch (EvText{line});
|
||||
}
|
||||
}
|
||||
|
||||
fsm.dispatch (EvEof{});
|
||||
|
||||
std::ofstream out (out_path, std::ios::trunc);
|
||||
if (!out.is_open()) {
|
||||
std::cerr << "Failed to open output file: " << out_path << " (" << std::strerror (errno) << ")\n";
|
||||
return 3;
|
||||
}
|
||||
|
||||
for (std::size_t i = 0; i < ctx.order.size(); ++i) {
|
||||
const auto &title = ctx.order[i];
|
||||
out << "\\section {" << latex_escape (title) << "}\n";
|
||||
|
||||
const auto it = ctx.by_title.find (title);
|
||||
if (it == ctx.by_title.end())
|
||||
continue;
|
||||
|
||||
for (const auto &q : it->second) {
|
||||
out << " \\subsection {" << latex_escape (q.meta) << "}\n";
|
||||
for (const auto &tl : q.text_lines)
|
||||
out << " " << latex_escape (tl) << "\n";
|
||||
out << " \\subsubsection{notes}\n\n";
|
||||
}
|
||||
|
||||
if (i + 1 < ctx.order.size())
|
||||
out << "\n";
|
||||
}
|
||||
|
||||
return 0;
|
||||
}
|
||||
|
||||
int main (int argc, char **argv) {
|
||||
validate_fsm_contract();
|
||||
|
||||
CliArgs args;
|
||||
if (!parse_args (argc, argv, args))
|
||||
return 1;
|
||||
|
||||
return convert (args.input, args.output);
|
||||
}
|
||||
@@ -0,0 +1,46 @@
|
||||
#include "convertasciidoc.h"
|
||||
|
||||
#include <cerrno>
|
||||
#include <cstring>
|
||||
#include <fstream>
|
||||
#include <iostream>
|
||||
|
||||
bool ConverterAsciiDoc::save (const std::string &output_file_name) {
|
||||
if (!parse())
|
||||
return false;
|
||||
|
||||
std::ofstream output (output_file_name, std::ios::trunc);
|
||||
if (!output.is_open()) {
|
||||
std::cerr << "Failed to open output file: "
|
||||
<< output_file_name
|
||||
<< " (" << std::strerror (errno) << ")\n";
|
||||
set_error_code (3);
|
||||
return false;
|
||||
}
|
||||
|
||||
const auto &book_titles = titles();
|
||||
for (const auto &title : book_titles) {
|
||||
output << "== " << title << "\n\n";
|
||||
|
||||
for (std::size_t quote_index = 0;
|
||||
quote_index < quote_count (title);
|
||||
++quote_index) {
|
||||
output << "=== " << quote_meta (title, quote_index) << "\n\n";
|
||||
output << "[quote]\n____\n";
|
||||
|
||||
const auto &text_lines = quote_text_lines (title, quote_index);
|
||||
std::size_t line_index = 0;
|
||||
if (!text_lines.empty() && text_lines.front().empty())
|
||||
line_index = 1;
|
||||
|
||||
for (; line_index < text_lines.size(); ++line_index)
|
||||
output << text_lines[line_index] << '\n';
|
||||
|
||||
output << "____\n\n";
|
||||
output << "==== Notes\n\n";
|
||||
}
|
||||
}
|
||||
|
||||
set_error_code (0);
|
||||
return true;
|
||||
}
|
||||
@@ -0,0 +1,13 @@
|
||||
#ifndef CONVERTASCIIDOC_H
|
||||
#define CONVERTASCIIDOC_H
|
||||
|
||||
#include "converter.h"
|
||||
|
||||
class ConverterAsciiDoc final : public Converter {
|
||||
public:
|
||||
using Converter::Converter;
|
||||
|
||||
bool save (const std::string &output_file_name) override;
|
||||
};
|
||||
|
||||
#endif // CONVERTASCIIDOC_H
|
||||
@@ -0,0 +1,46 @@
|
||||
#include "convertorg.h"
|
||||
|
||||
#include <cerrno>
|
||||
#include <cstring>
|
||||
#include <fstream>
|
||||
#include <iostream>
|
||||
|
||||
bool ConverterOrg::save (const std::string &output_file_name) {
|
||||
if (!parse())
|
||||
return false;
|
||||
|
||||
std::ofstream output (output_file_name, std::ios::trunc);
|
||||
if (!output.is_open()) {
|
||||
std::cerr << "Failed to open output file: "
|
||||
<< output_file_name
|
||||
<< " (" << std::strerror (errno) << ")\n";
|
||||
set_error_code (3);
|
||||
return false;
|
||||
}
|
||||
|
||||
const auto &book_titles = titles();
|
||||
for (const auto &title : book_titles) {
|
||||
output << "* " << title << "\n\n";
|
||||
|
||||
for (std::size_t quote_index = 0;
|
||||
quote_index < quote_count (title);
|
||||
++quote_index) {
|
||||
output << "** " << quote_meta (title, quote_index) << "\n\n";
|
||||
output << "#+begin_quote\n";
|
||||
|
||||
const auto &text_lines = quote_text_lines (title, quote_index);
|
||||
std::size_t line_index = 0;
|
||||
if (!text_lines.empty() && text_lines.front().empty())
|
||||
line_index = 1;
|
||||
|
||||
for (; line_index < text_lines.size(); ++line_index)
|
||||
output << text_lines[line_index] << '\n';
|
||||
|
||||
output << "#+end_quote\n\n";
|
||||
output << "*** Notes\n\n";
|
||||
}
|
||||
}
|
||||
|
||||
set_error_code (0);
|
||||
return true;
|
||||
}
|
||||
@@ -0,0 +1,13 @@
|
||||
#ifndef CONVERTORG_H
|
||||
#define CONVERTORG_H
|
||||
|
||||
#include "converter.h"
|
||||
|
||||
class ConverterOrg final : public Converter {
|
||||
public:
|
||||
using Converter::Converter;
|
||||
|
||||
bool save (const std::string &output_file_name) override;
|
||||
};
|
||||
|
||||
#endif // CONVERTORG_H
|
||||
@@ -5,13 +5,17 @@ CONFIG -= app_bundle
|
||||
CONFIG -= qt
|
||||
|
||||
SOURCES += \
|
||||
convertasciidoc.cpp \
|
||||
convertlatex.cpp \
|
||||
convertmarkdown.cpp \
|
||||
convertorg.cpp \
|
||||
converter.cpp \
|
||||
main.cpp
|
||||
|
||||
HEADERS += \
|
||||
convertasciidoc.h \
|
||||
convertlatex.h \
|
||||
convertmarkdown.h \
|
||||
convertorg.h \
|
||||
converter.h \
|
||||
typefactory.h
|
||||
+25
-5
@@ -3,25 +3,33 @@
|
||||
#include <iostream>
|
||||
#include <string>
|
||||
|
||||
#include "convertasciidoc.h"
|
||||
#include "convertlatex.h"
|
||||
#include "convertmarkdown.h"
|
||||
#include "convertorg.h"
|
||||
#include "typefactory.h"
|
||||
|
||||
#ifndef KINDLE2NOTES_VERSION
|
||||
#define KINDLE2NOTES_VERSION "unknown"
|
||||
#endif
|
||||
|
||||
using ConverterFactory = TypeFactory<std::string, Converter, std::string>;
|
||||
|
||||
static constexpr std::array<const char *, 2> kSupportedTypes = {
|
||||
"latex", "md"
|
||||
static constexpr std::array<const char *, 4> kSupportedTypes = {
|
||||
"latex", "md", "org", "adoc"
|
||||
};
|
||||
|
||||
struct CliArgs {
|
||||
std::string input;
|
||||
std::string output;
|
||||
std::string type;
|
||||
bool version = false;
|
||||
};
|
||||
|
||||
static void print_usage (const char *argv0) {
|
||||
std::cerr << "Usage: " << argv0
|
||||
<< " --input <file> --output <file> --type <type>\n";
|
||||
<< " --input <file> --output <file> --type <type>\n"
|
||||
<< " " << argv0 << " --version\n";
|
||||
}
|
||||
|
||||
static void print_supported_types() {
|
||||
@@ -33,8 +41,10 @@ static void print_supported_types() {
|
||||
|
||||
static ConverterFactory build_converter_factory() {
|
||||
ConverterFactory factory;
|
||||
factory.registerType<ConverterAsciiDoc> ("adoc");
|
||||
factory.registerType<ConverterLatex> ("latex");
|
||||
factory.registerType<ConverterMarkdown> ("md");
|
||||
factory.registerType<ConverterOrg> ("org");
|
||||
return factory;
|
||||
}
|
||||
|
||||
@@ -43,12 +53,13 @@ static bool parse_args (int argc, char **argv, CliArgs &out) {
|
||||
{"input", required_argument, nullptr, 'i'},
|
||||
{"output", required_argument, nullptr, 'o'},
|
||||
{"type", required_argument, nullptr, 't'},
|
||||
{"version", no_argument, nullptr, 'v'},
|
||||
{"help", no_argument, nullptr, 'h'},
|
||||
{nullptr, 0, nullptr, 0 }
|
||||
};
|
||||
|
||||
int c = 0;
|
||||
while ((c = ::getopt_long (argc, argv, "i:o:t:h", long_opts, nullptr)) != -1) {
|
||||
while ((c = ::getopt_long (argc, argv, "i:o:t:vh", long_opts, nullptr)) != -1) {
|
||||
switch (c) {
|
||||
case 'i':
|
||||
out.input = optarg;
|
||||
@@ -59,6 +70,9 @@ static bool parse_args (int argc, char **argv, CliArgs &out) {
|
||||
case 't':
|
||||
out.type = optarg;
|
||||
break;
|
||||
case 'v':
|
||||
out.version = true;
|
||||
break;
|
||||
case 'h':
|
||||
print_usage (argv[0]);
|
||||
return false;
|
||||
@@ -68,7 +82,8 @@ static bool parse_args (int argc, char **argv, CliArgs &out) {
|
||||
}
|
||||
}
|
||||
|
||||
if (out.input.empty() || out.output.empty() || out.type.empty()) {
|
||||
if (!out.version &&
|
||||
(out.input.empty() || out.output.empty() || out.type.empty())) {
|
||||
print_usage (argv[0]);
|
||||
return false;
|
||||
}
|
||||
@@ -81,6 +96,11 @@ int main (int argc, char **argv) {
|
||||
if (!parse_args (argc, argv, args))
|
||||
return 1;
|
||||
|
||||
if (args.version) {
|
||||
std::cout << "kindle2notes " << KINDLE2NOTES_VERSION << '\n';
|
||||
return 0;
|
||||
}
|
||||
|
||||
const ConverterFactory factory = build_converter_factory();
|
||||
if (!factory.contains (args.type)) {
|
||||
std::cerr << "Unsupported output type: " << args.type << '\n';
|
||||
|
||||
@@ -5,13 +5,17 @@ CONFIG -= app_bundle
|
||||
CONFIG -= qt
|
||||
|
||||
SOURCES += \
|
||||
convertasciidoc.cpp \
|
||||
convertlatex.cpp \
|
||||
convertmarkdown.cpp \
|
||||
convertorg.cpp \
|
||||
converter.cpp \
|
||||
main.cpp
|
||||
|
||||
HEADERS += \
|
||||
convertasciidoc.h \
|
||||
convertlatex.h \
|
||||
convertmarkdown.h \
|
||||
convertorg.h \
|
||||
converter.h \
|
||||
typefactory.h
|
||||
Reference in new issue
Block a user