Merge pull request #1497 from chrox/android_dict

Fix sdcv on Android cannot handle morphological changes

base

@@ -1 +1 @@
-Subproject commit ae6ea7e96c6695ed23aee84da1065b29e9fed0ea
+Subproject commit d2c53757a098f3f17fec816bea638adc71e2976f

frontend/apps/filemanager/filemanager.lua

@@ -26,6 +26,10 @@ local FileManager = InputContainer:extend{
     -- our own size
     dimen = Geom:new{ w = 400, h = 600 },
     onExit = function() end,
+
+    mv_bin = Device:isAndroid() and "/system/bin/mv" or "/bin/mv",
+    cp_bin = Device:isAndroid() and "/system/bin/cp" or "/bin/cp",
+    rm_bin = Device:isAndroid() and "/system/bin/rm" or "/bin/rm",
 }
 
 function FileManager:init()
@@ -265,9 +269,9 @@ function FileManager:pasteHere(file)
         local dest = lfs.attributes(file, "mode") == "directory" and
             file or file:match("(.*/)")
         if self.cutfile then
-            util.execute("/bin/mv", orig, dest)
+            util.execute(self.mv_bin, orig, dest)
         else
-            util.execute("/bin/cp", "-r", orig, dest)
+            util.execute(self.cp_bin, "-r", orig, dest)
         end
     end
 end
@@ -275,7 +279,7 @@ end
 function FileManager:deleteFile(file)
     local InfoMessage = require("ui/widget/infomessage")
     DEBUG("File to remove", util.realpath(file))
-    local rm = util.execute("/bin/rm", "-rf", util.realpath(file))
+    local rm = util.execute(self.rm_bin, "-rf", util.realpath(file))
     DEBUG("rm status", rm)
     if rm == 0 then
         UIManager:show(InfoMessage:new{

frontend/apps/reader/modules/readerdictionary.lua

@@ -70,12 +70,18 @@ function ReaderDictionary:stardictLookup(word, box)
         word = require("util").stripePunctuations(word)
         DEBUG("stripped word:", word)
         -- escape quotes and other funny characters in word
-        local std_out = io.popen("./sdcv --utf8-input --utf8-output -nj "
-            .. ("%q"):format(word) .. " --data-dir " .. self.data_dir, "r")
         local results_str = nil
-        if std_out then
-            results_str = std_out:read("*all")
-            std_out:close()
+        if Device:isAndroid() then
+            local A = require("android")
+            results_str = A.stdout("./sdcv", "--utf8-input", "--utf8-output",
+                    "-nj", word, "--data-dir", self.data_dir)
+        else
+            local std_out = io.popen("./sdcv --utf8-input --utf8-output -nj "
+                .. ("%q"):format(word) .. " --data-dir " .. self.data_dir, "r")
+            if std_out then
+                results_str = std_out:read("*all")
+                std_out:close()
+            end
         end
         --DEBUG("result str:", word, results_str)
         local ok, results = pcall(JSON.decode, results_str)

frontend/ui/widget/textboxwidget.lua

@@ -3,8 +3,7 @@ local Widget = require("ui/widget/widget")
 local RenderText = require("ui/rendertext")
 local Screen = require("device").screen
 local Geom = require("ui/geometry")
-
--- TODO: rename string:gsplit definition
+local util = require("util")
 
 --[[
 A TextWidget that handles long text wrapping
@@ -64,45 +63,6 @@ function TextBoxWidget:_wrapGreedyAlg(h_list)
     return v_list
 end
 
---[[
-Lua doesn't have a string.split() function and most of the time
-you don't really need it because string.gmatch() is enough.
-However string.gmatch() has one significant disadvantage for me:
-You can't split a string while matching both the delimited
-strings and the delimiters themselves without tracking positions
-and substrings. The string.gsplit() function below takes care of
-this problem.
-Author: Peter Odding
-License: MIT/X11
-Source: http://snippets.luacode.org/snippets/String_splitting_130
---]]
-function string:gsplit(pattern, capture)
-    pattern = pattern and tostring(pattern) or '%s+'
-    if (''):find(pattern) then
-        error('pattern matches empty string!', 2)
-    end
-    return coroutine.wrap(function()
-        local index = 1
-        repeat
-            local first, last = self:find(pattern, index)
-            if first and last then
-                if index < first then
-                    coroutine.yield(self:sub(index, first - 1))
-                end
-                if capture then
-                    coroutine.yield(self:sub(first, last))
-                end
-                index = last + 1
-            else
-                if index <= #self then
-                    coroutine.yield(self:sub(index))
-                end
-                break
-            end
-        until index > #self
-    end)
-end
-
 function TextBoxWidget:_getVerticalList(alg)
     if self.vertical_list then
         return self.vertical_list
@@ -110,10 +70,10 @@ function TextBoxWidget:_getVerticalList(alg)
     -- build horizontal list
     local h_list = {}
     local line_count = 0
-    for line in self.text:gsplit("\n", true) do
+    for line in util.gsplit(self.text, "\n", true) do
         for words in line:gmatch("[\32-\127\192-\255]+[\128-\191]*") do
-            for word in words:gsplit("%s+", true) do
-                for w in word:gsplit("%p+", true) do
+            for word in util.gsplit(words, "%s+", true) do
+                for w in util.gsplit(word, "%p+", true) do
                     local word_box = {}
                     word_box.word = w
                     word_box.width = RenderText:sizeUtf8Text(0, Screen:getWidth(), self.face, w, true, self.bold).x

frontend/util.lua

@@ -8,4 +8,43 @@ function util.stripePunctuations(word)
     return word:gsub("\226[\128-\131][\128-\191]",''):gsub("^%p+",''):gsub("%p+$",'')
 end
 
+--[[
+Lua doesn't have a string.split() function and most of the time
+you don't really need it because string.gmatch() is enough.
+However string.gmatch() has one significant disadvantage for me:
+You can't split a string while matching both the delimited
+strings and the delimiters themselves without tracking positions
+and substrings. The gsplit function below takes care of
+this problem.
+Author: Peter Odding
+License: MIT/X11
+Source: http://snippets.luacode.org/snippets/String_splitting_130
+--]]
+function util.gsplit(str, pattern, capture)
+    pattern = pattern and tostring(pattern) or '%s+'
+    if (''):find(pattern) then
+        error('pattern matches empty string!', 2)
+    end
+    return coroutine.wrap(function()
+        local index = 1
+        repeat
+            local first, last = str:find(pattern, index)
+            if first and last then
+                if index < first then
+                    coroutine.yield(str:sub(index, first - 1))
+                end
+                if capture then
+                    coroutine.yield(str:sub(first, last))
+                end
+                index = last + 1
+            else
+                if index <= #str then
+                    coroutine.yield(str:sub(index))
+                end
+                break
+            end
+        until index > #str
+    end)
+end
+
 return util

platform/android/llapp_main.lua

@@ -27,7 +27,7 @@ A.LOGI("intent file path " .. (file or ""))
 pcall(function() dofile("/sdcard/koreader/patch.lua") end)
 
 -- set proper permission for sdcv
-io.popen("chmod 755 ./sdcv")
+A.execute("chmod", "755", "./sdcv")
 
 -- create fake command-line arguments
 arg = {"-d", file or "/sdcard"}

platform/android/luajit-launcher

@@ -1 +1 @@
-Subproject commit a00e3d1c7f73756a3c3cf45b1828c2b55a903503
+Subproject commit ddcf09b0f8b60e8e3137b818c973aee3ff76d3f4

spec/unit/util_spec.lua

@@ -0,0 +1,36 @@
+require("commonrequire")
+
+local util = require("util")
+
+describe("util module", function()
+    it("should strip punctuations around word", function()
+        assert.is_equal(util.stripePunctuations("\"hello world\""), "hello world")
+        assert.is_equal(util.stripePunctuations("\"hello world?\""), "hello world")
+        assert.is_equal(util.stripePunctuations("\"hello, world?\""), "hello, world")
+        assert.is_equal(util.stripePunctuations("“你好“"), "你好")
+        assert.is_equal(util.stripePunctuations("“你好?“"), "你好")
+    end)
+    it("should split string with patterns", function()
+        local sentence = "Hello world, welcome to KoReader!"
+        local words = {}
+        for word in util.gsplit(sentence, "%s+", false) do
+            table.insert(words, word)
+        end
+        assert.are_same(words, {"Hello", "world,", "welcome", "to", "KoReader!"})
+    end)
+    it("should split command line arguments with quotation", function()
+        local command = "./sdcv -nj \"words\" \"a lot\" 'more or less' --data-dir=dict"
+        local argv = {}
+        for arg1 in util.gsplit(command, "[\"'].-[\"']", true) do
+            for arg2 in util.gsplit(arg1, "^[^\"'].-%s+", true) do
+                for arg3 in util.gsplit(arg2, "[\"']", false) do
+                    local trimed = arg3:gsub("^%s*(.-)%s*$", "%1")
+                    if trimed ~= "" then
+                        table.insert(argv, trimed)
+                    end
+                end
+            end
+        end
+        assert.are_same(argv, {"./sdcv", "-nj", "words", "a lot", "more or less", "--data-dir=dict"})
+    end)
+end)