Merge pull request #41 from pirovc/dev

genome_updater version 0.3.0

.gitmodules

@@ -0,0 +1,12 @@
+[submodule "tests/libs/bats"]
+	path = tests/libs/bats
+	url = https://github.com/bats-core/bats-core.git
+[submodule "tests/libs/bats-assert"]
+	path = tests/libs/bats-assert
+	url = https://github.com/bats-core/bats-assert.git
+[submodule "tests/libs/bats-file"]
+	path = tests/libs/bats-file
+	url = https://github.com/bats-core/bats-file.git
+[submodule "tests/libs/bats-support"]
+	path = tests/libs/bats-support
+	url = https://github.com/bats-core/bats-support.git

.travis.yml

@@ -0,0 +1,10 @@
+language: bash
+
+before_install:
+  - sudo apt-get install parallel
+
+script:
+    - tests/libs/bats/bin/bats tests/integration_offline.bats
+
+notifications:
+  email: false

LICENSE.md

@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2020 - Vitor C. Piro - http://github.com/pirovc
+Copyright (c) 2022 - Vitor C. Piro - http://github.com/pirovc
 All rights reserved.
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
@@ -19,4 +19,4 @@ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
-THE SOFTWARE.
+THE SOFTWARE.

tests/README.md

@@ -0,0 +1,5 @@
+# genome_updater tests
+
+genome_updater uses the [bats](https://github.com/bats-core/bats-core) testing framework for Bash.
+
+Use the `download_test_set.sh` to re-create a random set of offline files to test. Files will be downloaded to `files/genomes`.

tests/download_test_set.sh

@@ -0,0 +1,31 @@
+#!/usr/bin/env bash
+
+entries=20
+outfld="files/"
+mkdir -p ${outfld}
+ext="assembly_report.txt" #,protein.faa.gz"
+db="refseq,genbank"
+og="archaea,fungi"
+
+og="default,${og}"
+for d in ${db//,/ }
+do
+    for o in ${og//,/ }
+    do      
+        if [[ ${o} == "default" ]]; then
+            mkdir -p "${outfld}genomes/${d}/"
+            wget --quiet --show-progress -O "full_assembly_summary.txt" "ftp://ftp.ncbi.nlm.nih.gov/genomes/${d}/assembly_summary_${d}.txt"
+            out_as="${outfld}genomes/${d}/assembly_summary_$d.txt"
+        else
+            mkdir -p "${outfld}genomes/${d}/${o}/"
+            wget --quiet --show-progress -O "full_assembly_summary.txt" "ftp://ftp.ncbi.nlm.nih.gov/genomes/${d}/${o}/assembly_summary.txt"
+            out_as="${outfld}genomes/${d}/${o}/assembly_summary.txt"
+        fi
+        head -n 2 "full_assembly_summary.txt" > "${out_as}"
+        tail -n+3 "full_assembly_summary.txt" | shuf | head -n ${entries} >> "${out_as}"
+        tail -n+3 "${out_as}" | cut -f 20 | sed 's/https:/ftp:/g' | xargs -P ${entries} wget --quiet --show-progress --directory-prefix="${outfld}" --recursive --level 2 --accept "${ext}"
+        cp -r "${outfld}ftp.ncbi.nlm.nih.gov/genomes/" "${outfld}"
+        rm -rf "full_assembly_summary.txt" "${outfld}ftp.ncbi.nlm.nih.gov/" 
+    done
+done
+