update readme and add docker

README.md

@@ -1,15 +1,31 @@
 # slivar: filter/annotate variants in VCF/BCF format with simple expressions
 
 slivar finds all trios in a VCF, PED pair and let's the user specify an expression with indentifiers
-of `kid`, `mom`, `dad` that is applied to each possible trio. samples that pass that filter have the id
-of the kid added to the INFO field.
+of `kid`, `mom`, `dad` that is applied to each possible trio. For example, a simple expression to call
+*de novo* variants:
 
+```javascript
+variant.FILTER == 'PASS' && \                         # 
+variant.call_rate > 0.95 && \                         # genotype must be known for most of cohort.
+INFO.gnomad_af < 0.001 && \                           # rare in gnomad (must be in INFO)
+kid.alts == 1 && mom.alts == 0 && dad.alts == 0 && \  # alts are 0:hom_ref, 1:het, 2:hom_alt, -1:unknown
+kid.DP > 7 && mom.DP > 7 && dad.DP > 7 \              # sufficient depth in all
+&& (mom.AD[0] + dad.AD[0]) < 10                       # no evidence in the parents
 ```
-bpbio slivar \
-   --pass-only \ # output only variants that pass one of the filters.
+
+This requires passing variants that are rare in gnomad that have the expected genotypes and do
+not have any evidence in the parents. If there are 200 trios in the `ped::vcf` given, then this expression
+will be tested on each of those 200 trios.
+
+The expressions are javascript so the user can make these as complex as needed.
+
+
+```
+slivar \
+   --pass-only \ # output only variants that pass one of the filters (default is to output all variants)
    --vcf $vcf \
    --ped $ped \
-   --load functions.js \ 
+   --load functions.js \ # any valid javascript is allowed here.
    --out-vcf annotated.bcf \
    --info "variant.call_rate > 0.9" \ # this filter is applied before the trio filters and can speed evaluation if it is stringent.
    --trio "denovo:kid.alts == 1 && mom.alts == 0 && dad.alts == 0 \
@@ -20,6 +36,8 @@ bpbio slivar \
    --trio "informative:kid.GQ > 20 && dad.GQ > 20 && mom.GQ > 20 && kid.alts == 1 && ((mom.alts == 1 && dad.alts == 0) || (mom.alts == 0 && dad.alts == 1))" \
    --trio "recessive:recessive_func(kid, mom, dad)"
 
+```
+
 Note that `slivar` does not give direct access to the genotypes, instead exposing `alts` where 0 is homozygous reference, 1 is heterozygous, 2 is
 homozygous alternate and -1 when the genotype is unknown. It is recommended to **decompose** a VCF before sending to `slivar`
 
@@ -32,3 +50,11 @@ homozygous alternate and -1 when the genotype is unknown. It is recommended to *
 
  + sample attributes (via `kid`, `mom`, `dad`) include in the FORMAT. available as e.g. kid.AD[1]
  + sample attributes from the ped for `affected`, `sex` are available as, e.g. kid.sex.
+
+
+## How it works
+
+ `slivar` embeds the [duktape javascript engine](https://duktape.org/) to allow the user to specify expressions.
+ For each variant, each trio (and each sample), it fills the appropriate `attributes`. This can be intensive for
+ VCFs with many samples, but this is done **as efficiently as possible** such that `slivar` can evaluate 10's of
+ thousand of variants per second even with dozens of trios.

docker/Dockerfile

@@ -0,0 +1,40 @@
+FROM centos:centos6
+
+RUN yum install -y git curl wget zlib-devel xz-devel bzip2-devel openssl-devel libcurl-devel && \
+    wget http://people.centos.org/tru/devtools-2/devtools-2.repo -O /etc/yum.repos.d/devtools-2.repo && \
+    yum install -y devtoolset-2-gcc devtoolset-2-binutils devtoolset-2-gcc-c++ && \
+    source scl_source enable devtoolset-2 && \
+    echo "source scl_source enable devtoolset-2" >> ~/.bashrc && \
+    echo "source scl_source enable devtoolset-2" >> ~/.bash_profile && \
+    wget --quiet https://ftp.gnu.org/gnu/m4/m4-1.4.18.tar.gz && \
+    tar xzf m4-1.4.18.tar.gz && cd m4* && ./configure && make && make install && cd .. && rm -rf m4* && \
+    wget --quiet http://ftp.gnu.org/gnu/autoconf/autoconf-2.69.tar.gz && \
+    tar xzf autoconf-2.69.tar.gz && \
+    cd autoconf* && ./configure && make && make install && cd .. && rm -rf autoconf* && \
+    git clone --depth 1 https://github.com/ebiggers/libdeflate.git && \
+    cd libdeflate && make -j 2 CFLAGS='-fPIC -O3' libdeflate.a && \
+    cp libdeflate.a /usr/local/lib && cp libdeflate.h /usr/local/include && \
+    cd .. && rm -rf libdeflate && \
+    git clone https://github.com/samtools/htslib && \
+    cd htslib && git checkout 1.9 && autoheader && autoconf && \
+    ./configure --enable-libcurl --with-libdeflate && \
+    cd .. && make -j4 CFLAGS="-fPIC -O3" -C htslib install && \
+    echo "/usr/local/lib" >> etc/ld.so.conf && \
+    ldconfig
+
+RUN cd / && \
+    git clone -b devel --depth 1000 git://github.com/nim-lang/nim nim && \
+    cd nim && \
+    chmod +x ./build_all.sh && \
+    scl enable devtoolset-2 ./build_all.sh && \
+    export PATH=/nim/bin:$PATH && \
+    echo 'PATH=/nim/bin:$PATH' >> ~/.bashrc && \
+    echo 'PATH=/nim/bin:$PATH' >> ~/.bash_profile && \
+    cd / && \
+    git clone --depth 1 git://github.com/brentp/slivar.git && \
+    cd slivar && \
+    nimble install -y hts && \
+    source scl_source enable devtoolset-2  && \
+    nim c -d:release -o:/usr/bin/slivar --passC:-flto src/slivar && \
+    rm -rf /slivar && slivar -h
+