updated tests

Project.toml

@@ -15,6 +15,6 @@ Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 Unicode = "4ec0a83e-493e-50e2-b9ac-8f72acf5a8f5"
 
 [compat]
-CSV = "0.10.15"
-DataFrames = "1.7.0"
-SyntheticVowelDataset = "0.1.1"
+CSV = "0.10"
+DataFrames = "1.7"
+SyntheticVowelDataset = "0.1"

src/corpora/synthetic_vowel_dataset.jl

@@ -5,7 +5,7 @@
 function prepare(::Val{:SYNTHETIC_VOWEL_DATASET}, synsetdir, odir)
     # Validate the data directory
     ! isdir(synsetdir) && throw(ArgumentError("invalid path $(synsetdir)"))
-    
+
     # Create the output directory.
     dir = mkpath(odir)
     rm(joinpath(odir, "recordings.jsonl"), force=true)
@@ -41,7 +41,7 @@ function synset_recordings(dir::AbstractString)
         for file in files
             lname, ext = splitext(file)
             sname = split(lname, "_")
-            
+
             ext != ".wav" && continue
 
             spk = join(sname[1:3], "_")
@@ -70,14 +70,14 @@ function synset_metadata(dir)
 
     metadata = Dict()
 
-    fpath = "$dir/$(basename(dir))_detailed.csv"
+    fpath = joinpath("$dir", "$(basename(dir))_detailed.csv")
     df = DataFrame(CSV.File(fpath))
 
     # Get number of filter coefficients
     countfilter = count(col -> occursin(r"^a_\d+$", col), names(df))
 
     for row in eachrow(df)
-    
+
         # Get metadata
         spk = row["fname"]
         gender = row["gender"]
@@ -86,10 +86,10 @@ function synset_metadata(dir)
         f₀ = row["f0"]
         ϕ = row["ϕ"]
         vtl = row["vtl"]
-        filter = Dict(["a_$i" => row["a_$i"] for i in 1:countfilter]) 
+        filter = Dict(["a_$i" => row["a_$i"] for i in 1:countfilter])
         angles = Dict(["θ$i" => row["θ$i"] for i in 1:Int(countfilter/2)])
-        magnitudes = Dict(["r$i" => row["r$i"] for i in 1:Int(countfilter/2)])        
-    
+        magnitudes = Dict(["r$i" => row["r$i"] for i in 1:Int(countfilter/2)])
+
         id = "$(spk)_$(gender)_$(vowel)_$(sigid)"
         metadata[id] = Annotation(
             id,  # recording id and annotation id are the same since we have

test/runtests.jl

+
+using JSON
 using SpeechDatasets
+using SyntheticVowelDataset
 using Test
 
 
-
 ## The following tests do not work on the Gitlab CI because the volumes are not mounted on the runner. TODO find better tests
 
-using JSON
-
-PATHS = JSON.parsefile(
-    joinpath(@__DIR__, "src/corpora", "corpora.json")
-)
+#PATHS = JSON.parsefile(
+#    joinpath(@__DIR__, "../src/corpora", "corpora.json")
+#)
 
 
 # rm("/tmp/pfc/recordings.jsonl", force=true)
@@ -45,11 +45,14 @@ PATHS = JSON.parsefile(
 # @test isfile("/tmp/pfc/recordings.jsonl")
 # @test isfile("/tmp/pfc/annotations.jsonl")
 
-using SyntheticVowelDataset
 
-println("Testing SYNTHETIC_VOWEL_DATASET loading")
- SyntheticVowelDataset.generate("/tmp/synthetic_vowel_dataset_dir", "calliope"; classes_number=1, signals_number=1)
-ds = SpeechDatasets.dataset(:SYNTHETIC_VOWEL_DATASET, "/tmp/synthetic_vowel_dataset_dir", "/tmp/synset")
+@testset "SYNTHETIC_VOWEL_DATASET" begin
+    datadir = mktempdir(; cleanup = false)
+    manifestdir = mktempdir(; cleanup = false)
+    SyntheticVowelDataset.generate(datadir, "calliope"; classes_number=1, signals_number=1)
+    ds = SpeechDatasets.dataset(:SYNTHETIC_VOWEL_DATASET, datadir, manifestdir)
+
+    @test isfile(joinpath(manifestdir, "recordings.jsonl"))
+    @test isfile(joinpath(manifestdir, "annotations.jsonl"))
+end
 
-@test isfile("/tmp/synset/recordings.jsonl")
-@test isfile("/tmp/synset/annotations.jsonl")