Skip to content

Instantly share code, notes, and snippets.

@yangchenyun
Last active February 19, 2023 10:10
  • Star 47 You must be signed in to star a gist
  • Fork 11 You must be signed in to fork a gist
Star You must be signed in to star a gist
Save yangchenyun/a1c123935d82f5e25d57 to your computer and use it in GitHub Desktop.
Get back my books from Kindle
/*
* @fileoverview Program to free the content in kindle books as plain HTML.
*
* This is largely based on reverse engineering kindle cloud app
* (https://read.amazon.com) to read book data from webSQL.
*
* Access to kindle library is required to download this book.
*/
// The Kindle Compression Module copied from http://read.amazon.com application
// The script reuses the same logic to decompress the fragments
var KindleCompression = function() {
function h(a, c, g) {
var f, g = g > 0 ? g : b;
for (f in c)
c[f] >= g && (g = c[f] + 1);
f = g;
for (var d in a)
for (var g = a[d], e = 2; e <= g.length; e++) {
var h = g.substr(0, e);
c.hasOwnProperty(h) || (c[h] = f++)
}
return c
}
function j(c, e) {
var l, h = b;
l = "";
for (var r = 0; r < e.length; ) {
var o = e.charAt(r);
r++;
o.charCodeAt(0) <= d ? c.hasOwnProperty(l + o) ? l += o : (l.length > 0 && h < f && (c[l + o] = h,
h++,
h === a && (h = g)),
l = o) : l = ""
}
return c
}
// This is called when lzExpandWithStaticDictionary doesn't have a dictionary passed in
// in our case, we always construct the dictionary before decompress
// function e() {
// if (defaultDictionary === void 0 || defaultDictionary === {})
// defaultDictionary = j({}, defaultDictionaryString);
// return defaultDictionary
// }
var d = 9983
, c = d + 1
, b = c + 100 + 1
, f = 65533
, a = 55295
, g = 57344;
return {
lzCompress: function(k) {
var e = {}, l = [], h, r = b, o, q, s;
o = h = "";
for (var t = 0; t < k.length; ) {
var u = k.charAt(t);
t++;
if (u.charCodeAt(0) <= d) {
for (; o.length > 0; ) {
q = Math.min(100, o.length);
s = o.substr(0, q);
o = o.substr(q);
l.push(c + q);
for (q = 0; q < s.length; q++)
l.push(s.charCodeAt(q))
}
e.hasOwnProperty(h + u) ? h += u : (h.length > 0 && (l.push(h.length === 1 ? h.charCodeAt(0) : e[h]),
r < f && (e[h + u] = r,
r++,
r === a && (r = g))),
h = u)
} else
h.length > 0 && (l.push(h.length === 1 ?
h.charCodeAt(0) : e[h]),
h = ""),
o += u
}
for (h.length > 0 && l.push(h.length === 1 ? h.charCodeAt(0) : e[h]); o.length > 0; ) {
q = Math.min(100, o.length);
s = o.substr(0, q);
o = o.substr(q);
l.push(c + q);
for (q = 0; q < s.length; q++)
l.push(s.charCodeAt(q))
}
for (i = 0; i < l.length; i++)
l[i] = String.fromCharCode(l[i]);
return l.join("")
},
lzExpand: function(k) {
for (var e = {}, l = [], h, r = b, o = "", q, s = 0; s < k.length; ) {
h = k.charCodeAt(s);
s++;
if (h <= d)
h = String.fromCharCode(h);
else if (h >= b)
(h = e[h]) || (h = o + q);
else {
o = h - c;
l.push(k.substr(s, o));
s += o;
o = "";
continue
}
l.push(h);
q = h.charAt(0);
r < f && o.length > 0 && (e[r] = o + q,
r++,
r === a && (r = g));
o = h
}
return l.join("")
},
lzBuildDictionary: j,
lzGetDecompressionDictionary: function(a) {
var b = [], c;
for (c in a)
b[a[c]] = c;
return b
},
lzAddStringsToDictionary: h,
lzAddNumbersToDictionary: function(a, b) {
for (var c = [], g = 100; g < 1E3; g++)
c.push("" + g);
return h(c, a, b)
},
lzCompressWithStaticDictionary: function(a, b) {
if (b === void 0 || b === {})
b = e();
var g = [], f, h, o, q;
h = f = "";
for (var s = 0; s < a.length; ) {
var t = a.charAt(s);
s++;
if (t.charCodeAt(0) <= d) {
for (; h.length > 0; ) {
o = Math.min(100,
h.length);
q = h.substr(0, o);
h = h.substr(o);
g.push(c + o);
for (o = 0; o < q.length; o++)
g.push(q.charCodeAt(o))
}
b.hasOwnProperty(f + t) ? f += t : (f.length > 0 && g.push(f.length === 1 ? f.charCodeAt(0) : b[f]),
f = t)
} else
f.length > 0 && (g.push(f.length === 1 ? f.charCodeAt(0) : b[f]),
f = ""),
h += t
}
for (f.length > 0 && g.push(f.length === 1 ? f.charCodeAt(0) : b[f]); h.length > 0; ) {
o = Math.min(100, h.length);
q = h.substr(0, o);
h = h.substr(o);
g.push(c + o);
for (o = 0; o < q.length; o++)
g.push(q.charCodeAt(o))
}
for (i = 0; i < g.length; i++)
g[i] = String.fromCharCode(g[i]);
return g.join("")
},
lzExpandWithStaticDictionary: function(a, g, f) {
// NOTE: g is always defined in our case
// if (g === void 0 || g === []) {
// if (defaultDeDictionary === void 0 || defaultDeDictionary === []) {
// e();
// defaultDeDictionary = [];
// for (var h in defaultDictionary)
// defaultDeDictionary[defaultDictionary[h]] = h
// }
// g = defaultDeDictionary
// }
h = d;
var r = b;
f !== void 0 && (h = f - 1,
r = f);
for (var f = [], o = 0; o < a.length; ) {
var q = a.charCodeAt(o);
o++;
q <= h ? f.push(String.fromCharCode(q)) : q >= r ? f.push(g[q]) : (q -= c,
f.push(a.substr(o, q)),
o += q)
}
return f.join("")
}
}
}()
function s(metadata) { // a is bookinfo.metadata
var b = {};
if (metadata.cpr !== void 0) {
KindleCompression.lzAddStringsToDictionary(metadata.cpr, b),
KindleCompression.lzAddNumbersToDictionary(b);
return KindleCompression.lzGetDecompressionDictionary(b);
}
if (metadata.cprJson !== void 0) {
KindleCompression.lzAddStringsToDictionary(metadata.cprJson, b, 256),
KindleCompression.lzAddNumbersToDictionary(b, 256);
return KindleCompression.lzGetDecompressionDictionary(b);
}
}
var fs = require('fs');
var path = require('path');
var sqlite3 = require('sqlite3').verbose();
//
// http://read.amazon.com stores the ebook with webSQL, which is a sqlite accessible in Chrome
// To locate the sqlite file: http://ahoj.io/how-to-delete-web-sql-database-in-google-chrome
var KINDLE_DB = os.homedir() + '/Library/Application\ Support/Google/Chrome/Default/databases/https_read.amazon.com_0/17';
var db = new sqlite3.Database(KINDLE_DB);
// The following hack is from reverse engineering how kindle cloud app reads data
db.all("select metadata from 'bookinfo'", function(err, rows) {
rows.forEach(function (row) {
var metadata = JSON.parse(row.metadata);
var title = metadata.title;
var authors = metadata.authorList.join(',');
// used for dictionary request at https://read.amazon.com/dict/getDefinition?asin=<asin>&word=<word>
var asin = metadata.asin;
var ca = s(metadata);
console.log('staring process book: ' + title);
var HtmlHeader = '<html><head>' +
'<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">' +
'<meta name="author" content="' + authors + '">' +
'</head><body id="' + asin + '">';
var HtmlFile = path.join(os.tmpdir(), title.replace(/\s+/g, '-') + '.html');
fs.writeFile(HtmlFile, HtmlHeader);
console.log("created the file with HTML headers.");
db.all("select id, piece, other from 'fragments' order by id", function(err, rows) {
rows.forEach(function (row) {
var id = row.id;
var compressedFragmentData = row.piece;
var uncompressedFragmentData;
var imageDataMap = JSON.parse(row.other).imageData || {};
uncompressedFragmentData = KindleCompression.lzExpandWithStaticDictionary(
row.piece, ca);
// replace image path with base64 encoded string
for (var image in imageDataMap) {
uncompressedFragmentData = uncompressedFragmentData.replace(
'dataUrl="' + image + '"',
'src="' + imageDataMap[image] + '"');
}
fs.appendFile(HtmlFile, uncompressedFragmentData);
});
});
fs.appendFile(HtmlFile, '</body></html>');
console.log("created the file at: " + HtmlFile);
});
});
@indychris33
Copy link

hi! thanks for creating this great javascript! I have a question--how does one use it? :)

@marcolazzeri
Copy link

Fantastic job! I run this script in a debian (jessie) environment. To answer to a previous question node.js is needed to run this script and I had to install the nodejs-osenv package too (maybe a debian related issue).
A first change I did was to include a "require" directive at line 206 to cope with a "os object not found":
var os = require('osenv');
A second change was to modify the name of the function to get the home directory (from homedir to home: maybe a linux issue?) and obviously to change the location of the book database (all the change are at line 214):
var KINDLE_DB = os.home() + "/.config/google-chrome/Default/databases/https_read.amazon.com_0/10"
These changes let the script to run successfully and my books were dumped in html format. However they seemed to be mixed one another. Since the book database is an sqlite one I inspected the schema and I guessed that the query at line 238 should have been modified adding a "where clause":
db.all("select id, piece, other from 'fragments' where asin = '" + asin + "' order by id", function(err, rows) {
This change worked for me and the books are now properly dumped in html format.
I hope this comment could help.
Thanks a lot for this invaluable script!

@bitumin
Copy link

bitumin commented Aug 15, 2016

Thanks to @yangchenyun for the script and to @marcolazzeri for his corrections. It works ;)
A cached version of the webpage suggested by the script author about "how to locate the sqlite database" when using chrome/chromium can be found here: http://archive.is/7wF4E
I have noticed the inner links (chapters, notes) are broken, since they try to call "KindleContentInterface.gotoPosition(6,12399)" and KindleContentInterface is not defined anymore after the conversion.
Any ideas of how to replace all this goToPosition() methods with simple html internal links/anchors during the execution of the script?

@philiprhoades
Copy link

philiprhoades commented Sep 18, 2016

Thanks to @yangchenyun for the script and to @marcolazzeri for his corrections from me too! Also thanks to @bitumin for the db link.
I had to change this line:

var KINDLE_DB = os.home() + "/.config/google-chrome/Default/databases/https_read.amazon.com_0/10"
to:
var KINDLE_DB = os.home() + "/.config/google-chrome/Default/databases/https_read.amazon.com_0/75"

I have also made some trivial changes to suit me better and when I get the chance I will work my way through the code to better understand how it works.

@apehex
Copy link

apehex commented Oct 2, 2016

Great job extracting the actual encryption functions, that helped me a ton ! goddamn the proprietary apps and crap, i want cold hard pdf :)

i've made a fork to handle internal links and translate the javascript function into actual anchors
here you go @bitumin : https://gist.github.com/EelMood/84140e557065ac3d73f669f120429ae1

with a few tweaks similar to @marcolazzeri suggestions i was able to convert my kindle books from the french cloud reader, under chrome / debian.

thank you all, cheers !

@tokland
Copy link

tokland commented Nov 13, 2016

I just had to edit KINDLE_DB, it works like a charm. Using calibre, you can then convert the generated html file to pdf, epub or whatever. Thanks very much!

@barthelemypousset
Copy link

hi,
This script looks awesome, but i can't manage to use it...
I've installed nodejs and i'm using the command "node fetch_kindle.js" while connected to lire.amazon.fr with the ebook i want to be exported open(on Chrome), But nothing happen.
Can you explain me how to use this script ?

@d10r
Copy link

d10r commented Jan 26, 2017

Thanks for this script!
I've made a repo out of it, with small changes:

  • added package.json (dependencies)
  • configurable file path
  • some instructions

https://github.com/d10r/kindle-fetch

@darkerego
Copy link

You are awesome! This is absolutely brilliant, thanks so much. I hate DRM and am no fan of the Kindle app, or the Kindle, for that matter. The only thing I had to do to get this to work was to run npm install process and change the cont css = multi line variable to this format (my node version must be older):

// improve readability
const css = '\
<style> \
body { \
    margin: 0 auto; \
    max-width: 50em; \
    background: #FFFAFD; \
    font-size:100%; \
    line-height:1.5; \
} \
img { \
    max-width: 100%; \
} \
</style>'

@JohnKulp
Copy link

Hello! I've been messing around with this script, and I'm wondering if the db that I have from read.amazon.co.jp are in a slightly different format. I wouldn't think so, but it seems like even though the HTML metadata is being created correctly, and the pages with resource links set up, the images themselves aren't being extracted. If someone could PM me, I'll send over the trouble-causing files as well as the broken HTML result.

@jasonc310771
Copy link

Hi,
I am completely lost with this. How do I actually use this ? I have looked at the other versions and yours is the only one that shows 'usage' but where do I put the JS and where do I enter the commands above ?

I was hoping that it would work from Chrome Browser.

I changed the https_read.amazon.com_0/17 to https_read.amazon.com_0/2 as this is the last file in the list.

@HeathenGirl
Copy link

This is wonderful script. What about CSS styles? Are they stored somewhere inside sqlite database or they are "inlined" into fragments?

@AGuyCoding
Copy link

should this script still work?

@ICWiz
Copy link

ICWiz commented Apr 5, 2019

These need to be changed so that the script works on newer versions of Node.js:

line 235
fs.writeFile(HtmlFile, HtmlHeader);
fs.writeFile needs to be changed fs.writeFileSync

line 252
fs.appendFile(HtmlFile, uncompressedFragmentData);
line 256
fs.appendFile(HtmlFile, '</body></html>');
fs.appendFile needs to be changed to fs.appendFileSync

Also, this script can work with Mac Safari as well, just go to line 214:

var KINDLE_DB = os.homedir() + '/Library/Application\ Support/Google/Chrome/Default/databases/https_read.amazon.com_0/17';

change it to:

var KINDLE_DB = '/Users/<user>/Library/Safari/Databases/https_read.amazon.com_0/<db_file>';

The db file name will be a long UID ending in .db.

@glachancecmaisonneuve
Copy link

Make forks instead of describing changes please

@varun-ramraj
Copy link

I have made the changes suggested by ICWiz above along with a couple of package version updates and have issued a pull request to the https://github.com/d10r/kindle-fetch repository.

@iongradea
Copy link

Hi guys,

For all those wondering if the script still works, it does :) ! Just been using it to convert a book tonight.
Have a look at the different version for node, chrome and macOS I have on this issue : d10r/kindle-fetch#9.
Thx to all the people that did that amazing work ! Been struggling for several week with a reliable and free way to crack DRM (hate DRM) !!!

Cheers !

@dineshkoravi
Copy link

I tried to convert a file. The output file is 0KB. please help me.

image

@iongradea
Copy link

iongradea commented Feb 16, 2021

It doesn't seem to work no more !
I used this script https://github.com/d10r/kindle-fetch from @d10r and end up with the following error :

/home/iongradea/DEV/project_kindle_2_epub/kindle-fetch/node_modules/sqlite3/lib/trace.js:27
                    throw err;
                    ^
no metadata found, probably not a valid file

After downloading sqlite3 to look at the data base there is no bookinfo (line 274) and the content is pretty empty. It seems the browser doesn't download automatically and I can't manage to right click and ask to download. Seems it has been removed on Amazon Cloud Reader,

Here the print witht sqlit3, I checked all the tables :

sqlite> .tables
__WebKitDatabaseInfoTable__  keyValue                   
bookdata                     metrics                    
journal   

I would be more than happy to be wrong and it still works !
If someone has any clue !

Cheers,
Ion

Fiy more details :

  • node version : v8.10.0
  • Linux version : ubuntu 18.04 LTS
  • chrome://version
Google Chrome | 88.0.4324.150 (Official Build) (64-bit)
-- | --
Revision | e6c2a846b660e61f059daee2c46217af42913240-refs/branch-heads/4324@{#2103}
OS | Linux
JavaScript | V8 8.8.278.15
User Agent | Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.150 Safari/537.36
Command Line | /usr/bin/google-chrome-stable --flag-switches-begin --flag-switches-end --origin-trial-disabled-features=SecurePaymentConfirmation
Executable Path | /opt/google/chrome/google-chrome
Profile Path | /home/iongradea/.config/google-chrome/Default

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment