This seemed logical as Anki is “content-agnostic”. However, Anki does provide excellent support for Latex, and that seemed the way to go.

After a bit of research, I found the following combination that should work:

Pygments: A fine syntax highlighting engine written in Python that supports Latex among its many output formats. It can also be run directly via the pygmentize command-line tool. Incredibly useful!

minted: A Latex package that wraps Pygments, so you only need to wrap your snippet in \begin{minted}<snippet>\end{minted} to get syntax highlighting.

\begin{minted}{cpp}
#include
using namespace std;
int main()
{
 return 0;
}
\end{minted}

These are the steps I followed to get these tools to work with Anki.

Install Pygments: I followed instructions given on the Pygments page and used easy_install. I had to install the python-setuptools package to get easy_install on my machine.  (i.e. sudo aptitude install python-setuptools)

Install Latex: On Ubuntu, this is the texlive (or texlive-full) package. Check here for detailed instructions on setting up Latex. I already had it installed (for another project; I don’t think it’s installed by default on a Ubuntu box).

Install minted: This is a Latex package that uses Pygments for syntax highlighting a code snippet. I downloaded the zip file from the package downloads and used the below instructions to incorporate it into my Latex installation.

$ cp ~/Downloads/minted-v1.7.zip .
$ unp minted-v1.7.zip
 Archive: minted-v1.7.zip
 inflating: minted.dtx
 inflating: minted.ins
 inflating: minted.pdf
 inflating: README
 inflating: Makefile
$ latex minted.ins
$ sudo mkdir /usr/share/texmf/tex/latex/minted
$ sudo cp minted.sty /usr/share/texmf/tex/latex/minted
$ sudo texhash
 Updating /usr/local/share/texmf/ls-R...
 texhash: Updating /var/lib/texmf/ls-R-TEXMFMAIN...
 texhash: Updating /var/lib/texmf/ls-R-TEXLIVE...
 texhash: Updating /var/lib/texmf/ls-R...
 texhash: Done.

At this point, I needed to verify that I had the package correctly installed.
I created the below test.tex Latex file:

\documentclass[12pt]{article}
\special{papersize=3in,5in}
\usepackage[utf8]{inputenc}
\usepackage{amssymb,amsmath}
\usepackage{minted}
\pagestyle{empty}
\setlength{\parindent}{0in}
\begin{document}
\begin{minted}{cpp}

#include
using namespace std;
int main()
{
 return 0;
}

\end{minted}

\end{document}

I ran on the command line: “latex test.tex“ to be greeted with this error:

Package ifplatform Warning:
 shell escape is disabled, so I can only detect \ifwindows.

))

! Package minted Error: You must invoke LaTeX with the -shell-escape flag.

See the minted package documentation for explanation.
Type H  for immediate help.
 ...

l.7 \pagestyle
 {empty}
?

So, I ran it again with the “-shell-escape” argument and this time it ran successfully.

To be doubly sure, I ran the pdflatex command with the same arguments (“pdflatex -shell-escape test.tex“) and checked the pdf output file.

Now that I had minted working on my machine, it was time to get Anki to support this. I could either add this to the Latex environment for all the decks, or choose to add it for individual decks.  The former is a lot more involved than the latter so I will describe that first.

Anki provides a plugin to configure the Latex environment it uses. Go to File > Download > Shared Plugin, and download the “Configure latex” plugin. Restart Anki. Now, go to Settings > Plugins > “Open Plugin Folder…”. You should the “Configure latex.py” file in this folder. Open it.

In the “latex.latexPreamble” section, add this line: “\\usepackage{minted}\n” below the existing \usepackage line. Close the file and restart Anki. This makes the minted package available in the Latex environment used for the decks. Note that this seems to affect only newly created decks.

Verify this by creating a new deck and checking the Latex tab in Settings > Deck Properties. You should see the additional line in the Header text area.

To make the minted package available in already existing decks, or make it selectively available to decks that need it, open the deck and manually add the line  (“\usepackage{minted}“) in the Latex tab in Settings > Deck Properties instead of changing it via the “Configure Latex” plugin.

Now, I was almost there. I knew that when I tried to use minted in a flashcard I would likely get the same error as when I tried the command-line version:

! Package minted Error: You must invoke LaTeX with the -shell-escape flag.

See the minted package documentation for explanation.
Type H  for immediate help.
 ...

l.7 \pagestyle
 {empty}
?

Go on, try it (In case Anki has changed since I wrote this).

So, how do I make Anki invoke the latex command with the -shell-escape flag? Well, I looked around and couldn’t find an option to pass parameters to the latex command. I finally tracked down Anki’s latex plugin file to /usr/share/anki/anki/latex.py. The actual call to the latex command, within this file, was:

if call(["latex", "-interaction=nonstopmode",
 "tmp.tex"], stdout=log, stderr=log, startupinfo=si):
 return (False, errmsg("latex"))

I backed up this file, and edited it (via sudo vim /usr/share/anki/anki/latex.py) to pass the -shell-escape argument. This is the diff:

G61-Ubuntu> cd /usr/share/anki/anki/
G61-Ubuntu> diff latex.py latex.py_original
99c99
< if call(["latex", "-shell-escape", "-interaction=nonstopmode",
---
> if call(["latex", "-interaction=nonstopmode",
G61-Ubuntu>

That’s it. I now could enter the first snippet in this post in an Anki flash card and it would correctly display the syntax-highlighted C++ source code.

Note: On Windows, you typically install MiKTeX as the Latex distribution. I found that if you set the option to “Ask me”  for the “Download packages as needed” option (paraphrasing here; I don’t remember the exact commands) it didn’t download and install the minted package as I expected, but returned an error. After looking at how Anki invokes the latex command on Ubuntu, I assume it’s the “-interaction=nonstopmode”  option that’s preventing this. I googled and found that changing the option to either “Yes” or “No” prevents this error.

Note2: I am not very familiar with the components I describe above, and this may not be the best or easiest way to get syntax highlighting in Anki. If you have any suggestions or updates to what I have above, please drop me a comment.

Note3: I also added a feature request for Anki to allow arguments to the latex command via the Configure Latex plugin, so please check if that is supported in your version before hacking at the file directly.

Along the way I discovered the most amazing music player I’ve ever used, which excels on so many counts, not least the presence of one of the best tagging tools. Then I threw it all away when I moved to Ubuntu. Yes, it’s among the few pieces of software I miss on Linux. ( Atleast until such time as I can get it working under Wine )

Anyway, on Ubuntu I settled on Banshee and recently started adding ratings to the songs. This is a major investment of my time, and it is consequently not unreasonable to expect to:

• Be able to back up my investment in maintaining this information.
• Be able to export the ratings from Banshee into some standard format.

Backup Banshee data

The first is easily achieved: Banshee maintains all its information in the user home directory, under ~/.config/banshee-1/:

$ ls ~/.config/banshee-1/
addin-db-001  banshee.db  downloads  equalizers.xml  log

Excellent! Backing up my Banshee experience is as simple as backing up this folder. Even better, the library information is stored in a SQLite 3 database file  called banshee.db.

Export Ratings information

What’s not so good is that Banshee does not support any formal way to export the ratings information. Neither did googling turn up any pointers. Hence, this post.

The banshee.db file contains a few main tables (CoreArtists, CoreAlbums, CoreTracks), and all it takes is a peek into the db and a little bit of SQL to extract the ratings information from these tables and into a text file. I haven’t bothered to turn this into a more structured format like xml, since I don’t need it.

( If you don’t find the sqlite3 command, you probably need to install the sqlite3 package. )

sqlite3 -header ~/.config/banshee-1/banshee.db "select ct.Uri as FileLocation, ct.Rating, ct.Genre, ct.Title, car.Name Artist, cal.Title as Album from CoreTracks ct inner join CoreArtists car on ct.ArtistID=car.ArtistID inner join CoreAlbums cal on cal.AlbumID=ct.AlbumID" \; 

Sample output:

$ sqlite3 -header ~/.config/banshee-1/banshee.db "select ct.Uri as FileLocation, ct.Rating, ct.Genre, ct.Title, car.Name Artist, cal.Title as Album from CoreTracks ct inner join CoreArtists car on ct.ArtistID=car.ArtistID inner join CoreAlbums cal on cal.AlbumID=ct.AlbumID" \; | head -10
FileLocation|Rating|Genre|Title|Artist|Album
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/10. Room At The Top Of The Stairs.mp3|0|Bluegrass|Room At The Top Of The Stairs|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/01. My Walking Shoes.mp3|0|Bluegrass|My Walking Shoes|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/02. Hound Dawg.mp3|0|Bluegrass|Hound Dawg|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/03. Maggie's Farm.mp3|0|Bluegrass|Maggie's Farm|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/04. Memphis.mp3|0|Bluegrass|Memphis|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/05. Flint Hill Special.mp3|0|Bluegrass|Flint Hill Special|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/06. The Old Home Place.mp3|0|Bluegrass|The Old Home Place|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/07. Uncle Pen.mp3|5|Bluegrass|Uncle Pen|David Grisman, John Hartford & Mike Seeger|Retrograss
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/08. Air Mail Special.mp3|0|Bluegrass|Air Mail Special|David Grisman, John Hartford & Mike Seeger|Retrograss

The columns are self-explanatory. The output contains the song’s Title, Artist, Album, Genre, Rating and FileLocation which should be enough to uniquely identify the file/song and assign ratings to a library in any other format. This should hopefully work even if the physical file location is changed.

There are many other fields and tables in the SQLite db file that I haven’t used. If the above fields aren’t enough, and you need to export some other data ( like Year, Compilation, Disc, Composer etc ), have a peek at the db file yourself:

$ sqlite3 -header ~/.config/banshee-1/banshee.db ".schema" \; | more

....

$ sqlite3 -header ~/.config/banshee-1/banshee.db "select Uri,Rating,Genre,Title from CoreTracks where Rating != 0" \; | head -5
Uri|Rating|Genre|Title
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/07. Uncle Pen.mp3|5|Bluegrass|Uncle Pen
Bluegrass/Jerry Garcia & David Grisman/Not For Kids Only/02. Freight Train.mp3|4|Bluegrass|Freight Train
Bluegrass/Jerry Garcia & David Grisman/Not For Kids Only/07. Hopalong Peter.mp3|5|Bluegrass|Hopalong Peter
Bluegrass/Jerry Garcia & David Grisman/Not For Kids Only/08. Teddy Bears' Picnic.mp3|5|Bluegrass|Teddy Bears' Picnic
$

$ sqlite3 -header ~/.config/banshee-1/banshee.db "select Uri,Rating,Genre,Title from CoreTracks" \; | head -5
Uri|Rating|Genre|Title
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/10. Room At The Top Of The Stairs.mp3|0|Bluegrass|Room At The Top Of The Stairs
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/01. My Walking Shoes.mp3|0|Bluegrass|My Walking Shoes
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/02. Hound Dawg.mp3|0|Bluegrass|Hound Dawg
Bluegrass/David Grisman, John Hartford & Mike Seeger/Retrograss/03. Maggie's Farm.mp3|0|Bluegrass|Maggie's Farm

Also, if you do not want the column header in the output, lose the -header switch.

Caution

We are using an internal implementation artefact that could change at any time. The Banshee version this command works against is given below. If you’re running newer versions, it might be prudent to check the output to see if the sql commands still function as expected.

—

Versions (tested against): Banshee 1.4.3

To give some perspective, I used to have 2 fat32 partitions earlier, both of which could be accessed through nslu2. I reformatted two of them to ext3 to improve performance (Music played via banshee accessing my 30 GB music library shared via nslu2 would stutter and break regularly; numerous people have suggested that this was due to it being a fat32 partition )

Going into more detail ..

My 1TB Western Digital hard disk is partitioned thusly:

# fdisk  -l /dev/sdb

Disk /dev/sdb: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

 Device Boot      Start         End      Blocks  Id System
/dev/sdb1               1       52268   419842678+ 83 Linux
/dev/sdb2           52269      104536   419842710   b Win95 FAT32
/dev/sdb3          104537      121601   137074612+ 83 Linux
#

The goal is to connect this up to the Disk 1 port, and have the 2 ext3 and 1 fat32 partitions shared through nslu2.

However, what happens when I connect it up is that the fat32 partition shows up in /share/hdd/data/HDD_1_1_1 ( but not in /etc/fstab, or the output of mount command ), and other partitions do not show up ( though folders /share/hdd/data/HDD_1_1_2 and /share/hdd/data/HDD_1_1_3 are present ).

# mount
/dev/sda1 on /share/flash/data type ext3 (defaults,usrquota)
/dev/sda2 on /share/flash/conf type ext3 (defaults)
/dev/sda3 on swap type swap (defaults)
# umount /share/hdd/data/HDD_1_1_1
# mount /share/hdd/data/HDD_1_1_1 
mount: can't find /share/hdd/data/HDD_1_1_1 in /etc/fstab

Also, I am unable to mount the ext3 partitions as the command returns with an “Invalid argument” error.

# mount /dev/sdb1 /share/hdd/data/HDD_1_1_2/
mount: mounting /dev/sdb1 on /share/hdd/data/HDD_1_1_2/ failed: Invalid argument

dmesg output shows that nslu2 isn’t able to properly read the hard drive.

hub.c: new USB device 00:01.2-1, assigned address 3
Device descriptor:8 bytes received.
Device descriptor:18 bytes received.
Protocol: Transparent SCSI
*** detect:usb-storage
* scsi_register: usb-storage-1, 
* scsi_reg, hostno:1 prot is 49
* scsi_reg, retval->host_no:1
scsi1 : SCSI emulation for USB Mass Storage devices
! scan_scsis
do read capacity!!!
 read capacity ok!!!
The capacity is 1953525168
*** 49
 Vendor: WD        Model: 10EAVS External   Rev: 1.75
 Type:   Direct-Access                      ANSI SCSI revision: 02
Attached scsi disk sdb at scsi1, channel 0, id 0, lun 0
SCSI device sdb: 1953525168 512-byte hdwr sectors (1000205 MB)
31 port connect!!!!!
 /dev/scsi/host1/bus0/target0/lun0: p1 p2 p3
WARNING: USB Mass Storage data integrity not assured
USB Mass Storage device found at 3
MSDOS FS: IO charset cp437
EXT3-fs: unsupported inode size: 256
EXT2-fs warning (device sd(8,17)): ext2_read_super: mounting ext3 filesystem as ext2

EXT2-fs: unsupported inode size: 256
FAT: bogus logical sector size 0
#

Stuck for ideas at the moment.

I know it wasn’t an issue with the network; wife’s MacBook connected to the wi-fi just fine during this period. Googling the dmesg messages turned up nothing tangible.

[   35.090022] [drm] Initialized drm 1.1.0 20060810
[   35.158146] pci 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[   35.158157] pci 0000:00:02.0: setting latency timer to 64
[   35.158473] [drm] Initialized i915 1.6.0 20080730 on minor 0
[   37.823593] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   37.825808] iwl3945 0000:0c:00.0: firmware: requesting iwlwifi-3945-1.ucode
[   37.967080] Registered led device: iwl-phy0:radio
[   37.967110] Registered led device: iwl-phy0:assoc
[   37.967171] Registered led device: iwl-phy0:RX
[   37.967189] Registered led device: iwl-phy0:TX
[   38.001315] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[   42.887035] [drm:i915_get_vblank_counter] *ERROR* trying to get vblank count for disabled pipe 0
[   65.622137] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   65.820342] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   66.020181] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   66.220181] wlan0: authentication with AP 00:1d:68:6b:f1:39 timed out
[   83.426670] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   83.427419] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   83.624187] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   83.824193] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   84.024185] wlan0: authentication with AP 00:1d:68:6b:f1:39 timed out
[   95.626076] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   95.824064] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   96.024166] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[   96.224170] wlan0: authentication with AP 00:1d:68:6b:f1:39 timed out
[  113.424981] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  113.429815] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  113.632043] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  113.831718] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  114.028034] wlan0: authentication with AP 00:1d:68:6b:f1:39 timed out
[  131.231906] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  131.232738] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  131.432039] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  131.632125] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  131.836074] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  145.627856] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  145.629653] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  145.828070] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  146.028045] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  146.244030] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  163.866027] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  163.866901] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  164.068036] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  164.264039] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  164.468033] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  181.672824] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  181.673582] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  181.876640] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  182.076053] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  182.276051] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  199.470824] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  199.471520] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  199.668087] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  199.868059] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  200.072036] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  215.635841] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  215.643307] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 1
[  215.841221] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 2
[  216.040061] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  216.240038] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  219.016083] wlan0 direct probe responded
<.... snip (more of the same) ...>

[  984.241056] wlan0: direct probe to AP 00:1d:68:6b:f1:39 try 3
[  984.441054] wlan0: direct probe to AP 00:1d:68:6b:f1:39 timed out
[  997.073581] wlan0 direct probe responded
[  997.073588] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  997.276034] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  997.472968] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[  997.681073] wlan0: authentication with AP 00:1d:68:6b:f1:39 timed out
[ 3112.138020] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[ 3112.138664] wlan0: authenticate with AP 00:1d:68:6b:f1:39
[ 3112.166703] wlan0: authenticated
[ 3112.166714] wlan0: associate with AP 00:1d:68:6b:f1:39
[ 3112.367040] wlan0: associate with AP 00:1d:68:6b:f1:39
[ 3112.569548] wlan0: associate with AP 00:1d:68:6b:f1:39
[ 3112.696327] wlan0: RX AssocResp from 00:1d:68:6b:f1:39 (capab=0x411 status=0 aid=2)
[ 3112.696334] wlan0: associated
[ 3112.701082] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready

This isn’t the first time wi-fi has suddenly stopped working, but this was certainly the longest period for which it stopped. I have a hunch it might somehow be connected to the laptop overheating, but I’m yet to find a way to verify this theory of mine.

—
Versions: Linux xxx 2.6.28-14-generic #46-Ubuntu SMP i686 GNU/Linux