From 1979956f03c8bea80b68e2e2b65a75b19045a7be Mon Sep 17 00:00:00 2001
From: tastytea <tastytea@tastytea.de>
Date: Mon, 24 May 2021 15:35:49 +0200
Subject: [PATCH] Add basic search functionality and context output.

---
 CMakeLists.txt     |  2 +-
 src/CMakeLists.txt |  3 +-
 src/main.cpp       |  6 ++--
 src/search.cpp     | 81 +++++++++++++++++++++++++++++++++++++++++++++-
 src/search.hpp     | 10 ++++--
 5 files changed, 94 insertions(+), 8 deletions(-)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 34f4a62..f764439 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -22,7 +22,7 @@ set(CMAKE_CXX_EXTENSIONS OFF)
 include(cmake/debug_flags.cmake)
 
 # All dependencies except test dependencies.
-find_package(Boost 1.65.0 REQUIRED COMPONENTS program_options locale)
+find_package(Boost 1.65.0 REQUIRED COMPONENTS locale program_options regex)
 find_package(Gettext REQUIRED)
 find_package(Filesystem REQUIRED COMPONENTS Final Experimental)
 find_package(LibArchive 3.2 REQUIRED)
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index bf0db77..19ece0c 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -17,8 +17,9 @@ unset(headers_src)
 
 target_link_libraries(${PROJECT_NAME}_lib
   PUBLIC
-  Boost::program_options
   Boost::locale
+  Boost::program_options
+  Boost::regex
   std::filesystem
   LibArchive::LibArchive
   fmt::fmt)
diff --git a/src/main.cpp b/src/main.cpp
index dbc773b..d47bd4a 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -85,15 +85,15 @@ int main(int argc, char *argv[])
                 search::options opts;
                 if (vm.count("basic-regexp") > 0)
                 {
-                    opts.regex = search::regex_type::basic;
+                    opts.regex = search::regex_kind::basic;
                 }
                 if (vm.count("extended-regexp") > 0)
                 {
-                    opts.regex = search::regex_type::extended;
+                    opts.regex = search::regex_kind::extended;
                 }
                 if (vm.count("perl-regexp") > 0)
                 {
-                    opts.regex = search::regex_type::perl;
+                    opts.regex = search::regex_kind::perl;
                 }
                 if (vm.count("grep") > 0)
                 {
diff --git a/src/search.cpp b/src/search.cpp
index a8aa441..af08d27 100644
--- a/src/search.cpp
+++ b/src/search.cpp
@@ -17,7 +17,12 @@
 #include "search.hpp"
 
 #include "fs-compat.hpp"
+#include "zip.hpp"
 
+#include <boost/regex.hpp>
+
+#include <algorithm>
+#include <string>
 #include <string_view>
 #include <vector>
 
@@ -27,7 +32,81 @@ namespace epubgrep::search
 std::vector<match> search(const fs::path &filepath, std::string_view regex,
                           const options &opts)
 {
-    return {};
+    boost::regex::flag_type flags{};
+
+    switch (opts.regex)
+    {
+    case regex_kind::basic:
+    {
+        flags = opts.grep_like ? boost::regex::grep : boost::regex::basic;
+        break;
+    }
+    case regex_kind::extended:
+    {
+        flags = opts.grep_like ? boost::regex::egrep : boost::regex::extended;
+        break;
+    }
+    case regex_kind::perl:
+    {
+        flags = boost::regex::perl;
+        break;
+    }
+    }
+
+    if (opts.ignore_case)
+    {
+        flags |= boost::regex::icase;
+    }
+
+    boost::regex re(regex.data(), flags);
+    std::vector<match> matches;
+    for (const auto &entry : zip::list(filepath))
+    {
+        auto document{zip::read_file(filepath, entry)};
+        std::string::const_iterator begin{document.begin()};
+        std::string::const_iterator end{document.end()};
+        boost::match_results<std::string::const_iterator> match_result;
+        while (boost::regex_search(begin, end, match_result, re,
+                                   boost::match_default))
+        {
+            match match; // FIXME: Rename variable or struct.
+            match.filepath = entry;
+            match.text = match_result[0];
+            match.context = context(match_result, opts.context);
+
+            matches.emplace_back(match);
+            begin = match_result[0].second;
+        }
+    }
+
+    return matches;
+}
+
+match_context
+context(const boost::match_results<std::string::const_iterator> &match,
+        std::uint64_t words)
+{
+    const auto &prefix{match.prefix().str()};
+    const auto &suffix{match.suffix().str()};
+    size_t pos_before{prefix.length()};
+    size_t pos_after{};
+    ++words;
+
+    while (words != 0)
+    {
+        if (pos_before != std::string::npos)
+        {
+            pos_before = prefix.rfind(' ', pos_before) - 1;
+        }
+
+        if (pos_after != std::string::npos)
+        {
+            pos_after = suffix.find(' ', pos_after) + 1;
+        }
+        words -= 1;
+    }
+
+    return {prefix.substr(pos_before + 2), suffix.substr(0, pos_after - 1)};
 }
 
 } // namespace epubgrep::search
diff --git a/src/search.hpp b/src/search.hpp
index 8a08600..08e7620 100644
--- a/src/search.hpp
+++ b/src/search.hpp
@@ -19,6 +19,8 @@
 
 #include "fs-compat.hpp"
 
+#include <boost/regex.hpp>
+
 #include <cstdint>
 #include <string>
 #include <string_view>
@@ -39,7 +41,7 @@ struct match
     std::string page;      //!< The page number, if available.
 };
 
-enum class regex_type
+enum class regex_kind
 {
     basic,
     extended,
@@ -48,7 +50,7 @@ enum class regex_type
 
 struct options
 {
-    regex_type regex{regex_type::basic};
+    regex_kind regex{regex_kind::basic};
     bool grep_like{false};
     bool ignore_case{false};
     bool nostrip{false};
@@ -59,6 +61,10 @@ struct options
                                         std::string_view regex,
                                         const options &opts);
 
+[[nodiscard]] match_context
+context(const boost::match_results<std::string::const_iterator> &match,
+        std::uint64_t words);
+
 } // namespace epubgrep::search
 
 #endif // EPUBGREP_SEARCH_HPP