Error during db build: "find: -printf: unknown primary or operator"

[kliere]

I'm building a custom db following the instructions on the kraken website using the scripts provided by Mick Watson (http://www.opiniomics.org/building-a-kraken-database-with-new-ftp-structure-and-no-gi-numbers/). Everything works fine up to the moment when I try to build the db using the following command:
karsten$ kraken-build --build --threads 24 --work-on-disk --db kraken_20160720
Kraken build set to minimize RAM usage.
Creating k-mer set (step 1 of 6)...
Found jellyfish v1.1.11
find: -printf: unknown primary or operator

Copied from the script, the line in question looks like:
"KRAKEN_HASH_SIZE=$(find library/ '(' -name '.fna' -o -name '.fa' -o -name '*.ffn' ')' -printf '%s\n' | perl -nle '$sum += $_; END {print int(1.15 * $sum)}')"

I'm using Mac OS X 10.11.6 with Perl v5.18.2.

So, what is wrong with the printf command?

Cheers,

Karsten

[kliere]

Ok, found the fix for the above problem in the Pull request section. However, changing the line in question to "KRAKEN_HASH_SIZE=$(find library/ -name '*.fna' -ls | awk '{print $7}' | perl -nle '$sum += $_; END {print int(1.15 * $sum)}')" results now in the following error:

karsten$ kraken-build --build --threads 24 --work-on-disk --db kraken_20160720
Kraken build set to minimize RAM usage.
Creating k-mer set (step 1 of 6)...
Found jellyfish v1.1.11
Hash size not specified, using '27010052168'
/usr/local/Cellar/kraken/0.10.5-beta/libexec/build_kraken_db.sh: line 96: 49200 Broken pipe: 13 find library/ '(' -name '.fna' -o -name '.fa' -o -name '*.ffn' ')' -print0
49201 Exit 1 | xargs -0 cat
49202 Killed: 9 | jellyfish count -m $KRAKEN_KMER_LEN -s $KRAKEN_HASH_SIZE -C -t $KRAKEN_THREAD_CT -o database /dev/fd/0

So, what's next? ;-)

Cheers,

Karsten

[kliere]

Interestingly, trying to build the library under Linux (Ubuntu 16.04.1 LTS) results in pretty much the same error.

[andrewdavis3]

having the same issue, but I'm not getting a broken pipe Did you ever find a fix?
Creating k-mer set (step 1 of 6)...
Found jellyfish v1.1.11
Hash size not specified, using '23770869521'
/usr/local/Cellar/kraken/0.10.5-beta/libexec/build_kraken_db.sh: line 97: 76250 Done find library/ '(' -name '.fna' -o -name '.fa' -o -name '*.ffn' ')' -print0
76251 Exit 1 | xargs -0 cat
76252 Killed: 9 | jellyfish count -m $KRAKEN_KMER_LEN -s $KRAKEN_HASH_SIZE -C -t $KRAKEN_THREAD_CT -o database /dev/fd/0

[tseemann]

@kliere I see you are using Brew - I am still trying to get the builder working, as it depends on Jellyfish 1.1 not 2.0 as is installed. See https://github.com/Homebrew/homebrew-science/pull/2161

[andrewdavis3]

its the --jellyfish-hash-size that's causing it to exit, not sure what the max should be but probably need to add a cap.

[getopt]

The issue seems to be due macOS version of find not having -printf option. I have changed -printf '%s\n' to -print0 | xargs -0 stat -f '%i ' as per instructions https://stackoverflow.com/questions/752818/find-lacks-the-option-printf-now-what and this worked for me on macOS Sierra

KRAKEN_HASH_SIZE=$(find library/ '(' -name '*.fna' -o -name '*.fa' -o -name '*.ffn' ')' -print0 | xargs -0 stat -f '%i '| perl -nle '$sum += $_; END {print int(1.15 * $sum)}')

[tseemann]

@getopt nice 'find' on the lack of -printf support on macOS ! :-)

[tseemann]

@kliere @getopt @andrewdavis3 the bigger problem is that @DerrickWood has moved to industry and is no longer maintaining this software.

[kliere]

@tseemann, yep indeed - though Kraken is still a great tool, one should look for alternatives such as CLARK...

[jenniferlu717]

We are currently working on updates/fixes to the Kraken software. We haven't abandoned it, promise.