Merge pull request #172 from daniel-thom/assert-op

Add assert_op macro

.github/workflows/pr_testing.yml

@@ -10,7 +10,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        julia-version: ['1', 'nightly']
+        julia-version: ['1']
         julia-arch: [x64]
         os: [ubuntu-latest, windows-latest, macOS-latest]
 
@@ -24,7 +24,6 @@ jobs:
         env:
           PYTHON: ""
       - uses: julia-actions/julia-runtest@latest
-        continue-on-error: ${{ matrix.julia-version == 'nightly' }}
         env:
           PYTHON: ""
       - uses: julia-actions/julia-processcoverage@v1

docs/src/style.md

@@ -63,9 +63,22 @@ One exception is the case where one file has all single-line functions.
 an exception and terminate the application.
 * Do not use try/catch to handle retrieving a potentially-missing key from a
 dictionary.
-* Use @assert statements to guard against programming errors. Do not use them
-after detecting bad user input. Note that they may be compiled out in release
-builds.
+
+## Asserts
+* Use `@assert` statements to guard against programming errors. Do not use them
+after detecting bad user input. An assert tripping should indicate that there
+is a bug in the code. Note that they may be compiled out in optimized builds in
+the future.
+* Consider using `InfrastructureSystems.@assert_op` instead of the standard
+`@assert` because it will automatically print the value of the expression.
+Unlike the standard `@assert` the Julia compiler will never exclude
+`@assert_op` in optimized builds.
+
+```julia
+julia> a = 3; b = 4;
+julia> @assert_op a == b
+ERROR: AssertionError: 3 == 4
+```
 
 ## Globals
 

src/InfrastructureSystems.jl

@@ -65,6 +65,7 @@ get_internal(value::InfrastructureSystemsComponent) = value.internal
 
 include("common.jl")
 include("internal.jl")
+include("utils/assert_op.jl")
 include("utils/recorder_events.jl")
 include("utils/flatten_iterator_wrapper.jl")
 include("utils/generate_structs.jl")

src/component.jl

@@ -52,7 +52,7 @@ function _get_columns(start_time, count, ts_metadata::ForecastMetadata)
     if window_count > 1
         index = Int(offset / interval) + 1
     else
-        @assert interval == Dates.Millisecond(0)
+        @assert_op interval == Dates.Millisecond(0)
         index = 1
     end
     if count === nothing

src/components.jl

@@ -247,7 +247,7 @@ function get_components(
         iter = FlattenIteratorWrapper(T, _components)
     end
 
-    @assert eltype(iter) == T
+    @assert_op eltype(iter) == T
     return iter
 end
 

src/deterministic.jl

@@ -260,7 +260,7 @@ get_window(f::Deterministic, initial_time; len = nothing) =
 
 function make_time_array(forecast::Deterministic)
     # Artificial limitation to reduce scope.
-    @assert get_count(forecast) == 1
+    @assert_op get_count(forecast) == 1
     timestamps = range(
         get_initial_timestamp(forecast);
         step = get_resolution(forecast),

src/deterministic_single_time_series.jl

@@ -163,7 +163,7 @@ function _translate_deterministic_offsets(
     s_index = (columns.start - 1) * interval_offset + 1
     e_index = (columns.stop - 1) * interval_offset + horizon
     @debug "translated offsets" horizon columns s_index e_index last_index
-    @assert s_index <= last_index "s_index = $s_index last_index = $last_index"
-    @assert e_index <= last_index "e_index = $e_index last_index = $last_index"
+    @assert_op s_index <= last_index
+    @assert_op e_index <= last_index
     return UnitRange(s_index, e_index)
 end

src/forecasts.jl

@@ -129,7 +129,7 @@ function get_window_common(
     if ndims(data) == 2
         # This is necessary because the Deterministic and Probabilistic are 3D Arrays
         # We need to do this to make the data a 2D TimeArray. In a get_window the data is always count = 1
-        @assert size(data)[1] <= len
+        @assert_op size(data)[1] <= len
         data = @view data[1:len, :]
     else
         data = @view data[1:len]

src/hdf5_time_series_storage.jl

@@ -294,7 +294,7 @@ function deserialize_time_series(
         uuid = get_time_series_uuid(ts_metadata)
         path = _get_time_series_path(root, uuid)
         attributes = _read_time_series_attributes(storage, path, rows, T)
-        @assert attributes["type"] == T
+        @assert_op attributes["type"] == T
         @debug "deserializing a StaticTimeSeries" T
         data_type = attributes["data_type"]
         data = get_hdf_array(path["data"], data_type, rows)
@@ -337,7 +337,7 @@ function deserialize_time_series(
         end
 
         attributes = _read_time_series_attributes(storage, path, rows, T)
-        @assert attributes["type"] == T
+        @assert_op attributes["type"] == T
         @debug "deserializing a Forecast" T
         data_type = attributes["data_type"]
         data = get_hdf_array(path["data"], data_type, attributes, rows, columns)
@@ -510,8 +510,8 @@ function deserialize_time_series(
         uuid = get_time_series_uuid(ts_metadata)
         path = _get_time_series_path(root, uuid)
         attributes = _read_time_series_attributes(storage, path, rows, T)
-        @assert attributes["type"] == T
-        @assert length(attributes["dataset_size"]) == 3
+        @assert_op attributes["type"] == T
+        @assert_op length(attributes["dataset_size"]) == 3
         @debug "deserializing a Forecast" T
         data = SortedDict{Dates.DateTime, Matrix{attributes["data_type"]}}()
         initial_timestamp = attributes["start_time"]
@@ -553,8 +553,8 @@ function deserialize_time_series(
         uuid = get_time_series_uuid(ts_metadata)
         path = _get_time_series_path(root, uuid)
         attributes = _read_time_series_attributes(storage, path, rows, T)
-        @assert attributes["type"] == T
-        @assert length(attributes["dataset_size"]) == 3
+        @assert_op attributes["type"] == T
+        @assert_op length(attributes["dataset_size"]) == 3
         @debug "deserializing a Forecast" T
         data = SortedDict{Dates.DateTime, Matrix{attributes["data_type"]}}()
         initial_timestamp = attributes["start_time"]
@@ -624,7 +624,7 @@ function _append_item!(path::HDF5.HDF5Group, name::AbstractString, value::Abstra
     push!(values, value)
 
     ret = HDF5.o_delete(path, name)
-    @assert ret == 0
+    @assert_op ret == 0
 
     path[name] = values
     @debug "Appended $value to $name" values
@@ -646,7 +646,7 @@ function _remove_item!(path::HDF5.HDF5Group, name::AbstractString, value::Abstra
     end
 
     ret = HDF5.o_delete(path, name)
-    @assert ret == 0
+    @assert_op ret == 0
 
     if isempty(values)
         is_empty = true

src/time_series_formats.jl

@@ -129,9 +129,9 @@ function get_unique_timestamps(::Type{T}, file::CSV.File) where {T <: TimeSeries
         end
     end
 
-    @assert length(timestamps) > 0
+    @assert_op length(timestamps) > 0
     for timestamp in timestamps[2:end]
-        @assert timestamp["count"] == timestamps[1]["count"]
+        @assert_op timestamp["count"] == timestamps[1]["count"]
     end
 
     return timestamps

src/utils/assert_op.jl

@@ -0,0 +1,37 @@
+"""
+Throw an `AssertionError` if conditions like `op(exp1, exp2)` are `false`, where `op` is a conditional infix operator.
+
+# Examples
+
+```
+julia> a = 3; b = 4;
+julia> @assert_op a == b
+ERROR: AssertionError: 3 == 4
+
+julia> @assert_op a + 3 > b + 4
+ERROR: AssertionError: 6 > 8
+```
+"""
+macro assert_op(expr)
+    assert_op(expr)
+end
+
+function assert_op(expr::Expr)
+    # Only special case expressions of the form `expr1 == expr2`
+    if length(expr.args) == 3 && expr.head == :call
+        return assert_op(expr.args[1], expr.args[2], expr.args[3])
+    else
+        return :(@assert $(expr))
+    end
+end
+
+function assert_op(op, exp1, exp2)
+    return :(
+        if !$op($(esc(exp1)), $(esc(exp2)))
+            val1 = $(esc(exp1))
+            val2 = $(esc(exp2))
+            op_str = $(esc(op))
+            throw(AssertionError("$val1 $op_str $val2"))
+        end
+    )
+end

src/utils/utils.jl

@@ -367,7 +367,7 @@ function get_initial_times(
     if count == 0
         return []
     elseif count == 1
-        @assert interval == Dates.Second(0) "interval=$interval"
+        @assert_op interval == Dates.Second(0)
         return range(initial_timestamp; stop = initial_timestamp, step = Dates.Second(1))
     end
     @assert interval != Dates.Second(0) "initial_timestamp=$initial_timestamp interval=$interval count=$count"
@@ -397,7 +397,7 @@ end
 function transform_array_for_hdf(data::SortedDict{Dates.DateTime, Vector{POLYNOMIAL}})
     lin_cost = hcat(values(data)...)
     rows, cols = size(lin_cost)
-    @assert length(first(lin_cost)) == 2
+    @assert_op length(first(lin_cost)) == 2
     t_lin_cost = Array{Float64}(undef, rows, cols, 2)
     for r in 1:rows, c in 1:cols
         tuple = lin_cost[r, c]
@@ -410,7 +410,7 @@ end
 
 function transform_array_for_hdf(data::Vector{POLYNOMIAL})
     rows = length(data)
-    @assert length(first(data)) == 2
+    @assert_op length(first(data)) == 2
     t_lin_cost = Array{Float64}(undef, rows, 1, 2)
     for r in 1:rows
         tuple = data[r]
@@ -425,7 +425,7 @@ function transform_array_for_hdf(data::SortedDict{Dates.DateTime, Vector{PWL}})
     quad_cost = hcat(values(data)...)
     rows, cols = size(quad_cost)
     tuple_length = length(first(quad_cost))
-    @assert length(first(first(quad_cost))) == 2
+    @assert_op length(first(first(quad_cost))) == 2
     t_quad_cost = Array{Float64}(undef, rows, cols, 2, tuple_length)
     for r in 1:rows, c in 1:cols
         tuple_array = quad_cost[r, c]
@@ -441,7 +441,7 @@ end
 function transform_array_for_hdf(data::Vector{PWL})
     rows = length(data)
     tuple_length = length(first(data))
-    @assert length(first(first(data))) == 2
+    @assert_op length(first(first(data))) == 2
     t_quad_cost = Array{Float64}(undef, rows, 1, 2, tuple_length)
     for r in 1:rows
         tuple_array = data[r, 1]

src/validation.jl

@@ -181,7 +181,7 @@ function check_limits(
     field_value,
 ) where {T <: Union{Nothing, NamedTuple}}
     # Validates up/down, min/max, from/to named tuples.
-    @assert length(field_value) == 2
+    @assert_op length(field_value) == 2
     result1 = check_limits_impl(valid_info, field_value[1])
     result2 = check_limits_impl(valid_info, field_value[2])
     return result1 && result2

test/common.jl

@@ -14,7 +14,7 @@ function create_system_data(; with_time_series = false, time_series_in_memory =
             file,
         )
         time_series = get_all_time_series(data)
-        @assert length(time_series) > 0
+        IS.@assert_op length(time_series) > 0
     end
 
     return data

test/test_assert_op.jl

@@ -0,0 +1,10 @@
+@testset "Test assert_op" begin
+    a = 2
+    b = 2
+    IS.@assert_op a == b
+    IS.@assert_op a + 2 == b + 2
+    IS.@assert_op isequal(a + 2, b + 2)
+
+    @test_throws AssertionError IS.@assert_op a + 3 == b + 2
+    @test_throws AssertionError IS.@assert_op isequal(a + 3, b + 2)
+end