** Easy way to make your own Anki deck that includes all the sentences you want from Tatoeba with audio **

Someone already asked about this feature and I decided to make a program for it. It's not perfect but I tried it and found quite useful for me. I tested it under Windows 7. It should work under any other common operating system since it's written in Python 2.7
If you don't know what Anki is, read this article https://en.wikipedia.org/wiki/Anki_(software)
You can import as many sentences as you want from any list or favorite page (I tested it with https://tatoeba.org/eng/sentences_lists/show/4022 and https://tatoeba.org/eng/favorites/of_user/odexed )

** Short instruction for Windows users **

1. You are supposed to have Anki installed in your computer. If not, you can get it here http://ankisrs.net/ It's easy to install.

2. If you don't have Python or it's not 2x you can just download this zip archive
https://drive.google.com/file/d...ew?usp=sharing

3. Unzip it into any folder and open Tatoeba_anki.py with any text editor (I recommend you to do it with Notepad++)
All you need to do is to edit at least 4 variables on the top of the file. For example, you want to make Anki cards from the quotes of Barack Obama with audio and translations into your language. You set the first variable UrlListOfSentences = 'https://tatoeba.org/eng/sentenc...ow/4022/none/'
If your native language is French and you only interested in English sentences from this list, set
srclang = ["en"] (it should be in square brackets because you can set several languages like srclang = ["en","es"] )
set audio3letterslangcodes = ["eng"] (this will be the same as srclang but in 3 letter language codes as it is in Tatoeba)
and targetlang = "fr" (only one target language per deck)

Important! If you set srclang you should set audio3letterslangcodes accordingly.

4. Save the file and close it. Run Prepare_deck.bat and wait for a while.
If program works correctly you should see a black console window that will close in a while. After this you will see a new folder 'foranki' in the same directory. This folder contains the csv file and media files with audio.

Important! I recommend you to edit and save the file with Notepad++ because default Notepad includes BOM information when you save it. You can also use any Python IDE to edit it.

5. Copy all mp3 files into Anki media folder (it's called collection.media and you can usually find it in your Documents folder) manually if you didn't set copymediafilestoankifolder = True and the ankimediafolder variable. If you set these variables just skip this step.

6. Open Anki and import new deck of cards. Choose 'exampledeck.csv' as a source.

*** If you have another Operating system ***
You can get the source code here https://github.com/alexanderk40...deckgeneration
and edit it in the way you want. It should work with most Python varieties 2x.

Once you imported the deck, you can export it as apkg file and use in your cellphone. I also tested it.

@bill I personally liked the favorites more than lists before they were changed just because they were structured in a very comfortable way. Now it's too much work to pass them into lists. Besides, I already have several lists that I use for other purposes. And last but not least, it's much easier to press a heart icon than to click the list icon, then click the combobox menu, look for a list I need, click it again and press OK.

I've also noticed that now I can't filter my favorite sentences by language. If I want to see all my favorite Spanish sentences there is no way to do it.

Good idea. It could solve the problem. By the way, I would like to have some search feature for my old comments too.

I agree on this and I understand the reasons. But as far as I know very few people have so many favorite sentences and even if they have, they don't open and check them very often so at least for me it was kind of necessary evil.
(I don't insist nor impose my opinion, just share some of my thoughts)

I can't say it bothers me too much to wait until this 1 page loads. But thank you, anyway.

With these default settings people would have lost a big deal of good non-native sentences. Don't forget that corpus maintainers work on improving the quality. Many non-native sentences have 'OK' tags, marks and comments. I don't think we should just cut this from the search results.

I just don't understand why it should be something different. We already can filter and proofread any sentence. Corpus mantainers do a great job to increase the quality (but we still need more people to do it). Perhaps we should focus on our new contributors to make the things more clear for them, but generally speaking I think we've made a big progress yet.

P.S. Thanks, but tastes differ, you know.

> We could use smart defaults and only display sentences owned by native speakers on the search results

Sorry, but I don't like this idea. There are some people who contribute in their non-native languages but I trust their sentences as if they were owned by a native speaker. For example, I trust Spanish sentences of alexmarcelo, English sentences that posted by Amastan. And I do want to see them among the search results. We already have an option 'Owned by a self-identified native' and I think it's enough for people who don't trust non-native sentences.

This second place sounds like a "Perfect Tatoeba" but there is nothing perfect in the world. Most people would certainly argue about some sentences they see there. Besides this would cause a lot of trouble and make things more difficult since it's not easy to decide what is natural and what isn't.

Let's just try to add natural and correct sentences and use tags and ratings. I think it's the best thing we can do for now. And as to bad sentences, we're all humans prone to error as you said. I agree.

We actually have a rule that says "Don't change what is already correct". I don't suggest making all the sentences unchangeable but perhaps it would make some sense for a small part of them that doubtless correct and natural.

We don't really need to split all Tatoeba into two sections. I like the idea of making the sentence with audio unable to change. Perhaps we could freeze the sentences that have an OK tag and several 'OK' marks either. This way people would know what sentences they can trust.

Just to make things clear, now I have to go through 141 pages in order to look my favorite sentences over. What if I need to find some of them quickly? I used to press Ctrl+F and put a clue word and now I have no idea what to do with it. The search engine can find hundreds of sentences with that word but it's not what I need. I need just the sentences that I harvested because they help me to memorize new words the best way.

I've noticed that my favorite sentences now appear on several pages, but before they were just on one page. Why was it changed?

I see. Thank you.

I wonder where is the copy to clipboard button.

http://i.imgur.com/CCzyBeB.png?1
http://i.imgur.com/OzDsWw0.png?1

I couldn't delete my message. The problem is now solved.

Indeed, it's good to know that. Thank you.

Thank you, I didn't know that.

Why can't I find this sentence? #4636494

https://tatoeba.org/eng/sentenc...rom=und&to=und