Merge pull request #410 from NREL-Sienna/issue-409

Fix performance issue with reading time series
NREL-Sienna · Oct 31, 2024 · 76c1d77 · 76c1d77
2 parents 8c4d57d + d7069b7
commit 76c1d77
Show file tree

Hide file tree

Showing 3 changed files with 42 additions and 16 deletions.
diff --git a/src/single_time_series.jl b/src/single_time_series.jl
@@ -44,7 +44,7 @@ function SingleTimeSeries(;
     return SingleTimeSeries(
         name,
         data,
-        _get_resolution(data),
+        get_resolution(data),
         scaling_factor_multiplier,
         internal,
     )
@@ -72,17 +72,6 @@ function SingleTimeSeries(
     )
 end
 
-function _get_resolution(data::TimeSeries.TimeArray)
-    if length(data) < 2
-        throw(
-            ConflictingInputsError(
-                "Resolution can't be inferred from the data. Please select an appropiate constructor.",
-            ),
-        )
-    end
-    return TimeSeries.timestamp(data)[2] - TimeSeries.timestamp(data)[1]
-end
-
 """
 Construct SingleTimeSeries from a TimeArray or DataFrame.
 

diff --git a/src/utils/utils.jl b/src/utils/utils.jl
@@ -385,11 +385,18 @@ end
 Return the resolution from a TimeArray.
 """
 function get_resolution(ts::TimeSeries.TimeArray)
+    if length(ts) < 2
+        throw(ConflictingInputsError("Resolution can't be inferred from the data."))
+    end
+
+    timestamps = TimeSeries.timestamp(ts)
+    return timestamps[2] - timestamps[1]
+end
+
+function check_resolution(ts::TimeSeries.TimeArray)
     tstamps = TimeSeries.timestamp(ts)
     timediffs = unique([tstamps[ix] - tstamps[ix - 1] for ix in 2:length(tstamps)])
-
     res = []
-
     for timediff in timediffs
         if mod(timediff, Dates.Millisecond(Dates.Day(1))) == Dates.Millisecond(0)
             push!(res, Dates.Day(timediff / Dates.Millisecond(Dates.Day(1))))
@@ -419,14 +426,25 @@ function get_initial_timestamp(data::TimeSeries.TimeArray)
     return TimeSeries.timestamp(data)[1]
 end
 
-function get_module(module_name)
+# Looking up modules with Base.root_module is slow; cache them.
+const g_cached_modules = Dict{String, Module}()
+
+function get_module(module_name::AbstractString)
+    cached_module = get(g_cached_modules, module_name, nothing)
+    if !isnothing(cached_module)
+        return cached_module
+    end
+
     # root_module cannot find InfrastructureSystems if it hasn't been installed by the
     # user (but has been installed as a dependency to another package).
-    return if module_name == "InfrastructureSystems"
+    mod = if module_name == "InfrastructureSystems"
         InfrastructureSystems
     else
         Base.root_module(Base.__toplevel__, Symbol(module_name))
     end
+
+    g_cached_modules[module_name] = mod
+    return mod
 end
 
 get_type_from_strings(module_name, type) =

diff --git a/test/test_utils.jl b/test/test_utils.jl
@@ -53,3 +53,22 @@ IS.get_name(::FakeTimeSeries) = "fake"
         sprint(show, MIME("text/plain"), summary(FakeTimeSeries())),
     )
 end
+
+@testset "Test check_resolution" begin
+    initial_time1 = Dates.DateTime("2020-09-01")
+    len1 = 24
+    resolution1 = Dates.Hour(1)
+    timestamps1 = range(initial_time1; length = len1, step = resolution1)
+    ta1 = TimeSeries.TimeArray(timestamps1, rand(len1))
+    IS.check_resolution(ta1) == resolution1
+
+    initial_time2 = Dates.DateTime("2020-09-02")
+    len2 = 12
+    resolution2 = Dates.Minute(5)
+    timestamps2 = range(initial_time2; length = len2, step = resolution2)
+    ta2 = TimeSeries.TimeArray(timestamps2, rand(len2))
+    IS.check_resolution(ta2) == resolution2
+
+    ta3 = vcat(ta1, ta2)
+    @test_throws IS.DataFormatError IS.check_resolution(ta3)
+end