Prepare tag 0.17

.travis.yml

@@ -4,9 +4,8 @@ os:
  - windows
  - osx
 julia:
- - 1.0
- - 1.3
- #- nightly
+ - 1.2
+ - 1.5
 notifications:
  email: false
 addons:

NEWS.md

@@ -1,7 +1,18 @@
 # master (unpublished)
 - no changes
 
-# v0.16.0 (latest)
+# v0.17.0 (latest)
+
+### general
+- Julia versions < 1.2 are no longer supported; this change was necessary for the interleaved parallelism
+overhaul in this version
+- the interleaved parallelism backend got fixed to properly support changes introduced in Julia 1.2;
+it now runs more stably and features better error reporting
+
+### bug fixes
+- fixed occasional hangs when computing networks in parallel via the interleaved mode (issue #9)
+
+# v0.16.0
 
 ### general
 - add explicit parallelism flag (`parallel_mode`) to `learn_network` to avoid having to remove workers when switching from multi-core to single-core computations

Project.toml

@@ -4,7 +4,7 @@ keywords = ["metagenomics", "microbiome", "probabilistic-graphical-models", "hig
 license = "GPL-3.0"
 desc = "Inference of microbial interaction networks from large-scale heterogeneous abundance data"
 authors = ["Janko Tackmann <janko.tackmann@gmail.com>"]
-version = "0.16.0"
+version = "0.17.0"
 
 [deps]
 Combinatorics = "861a8166-3701-5b0c-9a16-15d98fcdc6aa"
@@ -37,7 +37,7 @@ JSON = "0.21"
 LightGraphs = "1"
 SimpleWeightedGraphs = "1"
 StatsBase = "0.32"
-julia = "1"
+julia = "1.2"
 
 [extras]
 DelimitedFiles = "8bb1440f-4735-579b-a4ab-409b98df4dab"

README.md

@@ -9,15 +9,21 @@ FlashWeave predicts ecological interactions between microbes from large-scale co
 
 ## Installation ##
 
-To install Julia, please follow instructions on https://github.com/JuliaLang/julia. The preferred way is to obtain a binary from https://julialang.org/downloads/. Make sure you install Julia 1.0 or above, the versions currently supported by FlashWeave.
+To install Julia, please follow instructions on https://github.com/JuliaLang/julia. The preferred way is to obtain a binary from https://julialang.org/downloads/. Make sure you install Julia 1.2 or above, the versions currently supported by FlashWeave.
 
 In an interactive Julia session, you can then install FlashWeave after typing `]` via
 
 ```julia
-(v1.0) pkg> add FlashWeave
-# to run tests: (v1.0) pkg> test FlashWeave
+(v1.2) pkg> add FlashWeave
+# to run tests: (v1.2) pkg> test FlashWeave
 ```
 
+**Important note: from version 0.17, FlashWeave dropped support for Julia versions < 1.2. If you want to use FlashWeave with older Julia installations, make sure to install FlashWeave 0.16.0 (or lower) via `]` + `add FlashWeave@0.16.0`.**
+
+## Latest changes ##
+
+See [NEWS.md](https://github.com/meringlab/FlashWeave.jl/blob/master/NEWS.md) for the latest features and bugfixes.
+
 ## Basic usage ##
 
 To learn an interaction network from an OTU table and (optionally) a meta data table, you can do

test/io.jl

@@ -60,7 +60,6 @@ end
  end
 
  @testset "unfinished states" begin
- println(tmp_path * "_unchecked.tsv")
  unf_dict_ld = FlashWeave.load_unfinished_variable_info(tmp_path * "_unchecked.tsv")
  @test begin
  all(all(getproperty(unf_dict_ld[i], p) == getproperty(unf_dict[i], p)