From 5c5a287034bf8e7d5d701fd5235990380c586800 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Stefanos=20Carlstr=C3=B6m?= <stefanos.carlstrom@gmail.com>
Date: Wed, 24 Mar 2021 08:55:00 +0100
Subject: [PATCH] Support Unicode occupations/states when parsing
 configurations (#36, #73)

---
 src/configurations.jl  | 16 ++++++++++++----
 test/configurations.jl | 10 ++++++++++
 2 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/src/configurations.jl b/src/configurations.jl
index 4d7ed1a..8f45cb4 100644
--- a/src/configurations.jl
+++ b/src/configurations.jl
@@ -319,9 +319,9 @@ function get_noble_core_name(config::Configuration{O}) where O
 end
 
 function state_sym(state::AbstractString)
-    if state == "c"
+    if state == "c" || state == "ᶜ"
         :closed
-    elseif state == "i"
+    elseif state == "i" || state == "ⁱ"
         :inactive
     else
         :open
@@ -339,13 +339,21 @@ end
 
 function parse_orbital_occupation(::Type{O}, orb_str) where {O<:AbstractOrbital}
     m = match(r"^(([0-9]+|.)([a-z]|\[[0-9]+\])[-]{0,1})([0-9]*)([*ci]{0,1})$", orb_str)
-    parse(O, m[1]) , (m[4] == "") ? 1 : parse(Int, m[4]), state_sym(m[5])
+    m2 = match(r"^(([0-9]+|.)([a-z]|\[[0-9]+\])[-]{0,1})([¹²³⁴⁵⁶⁷⁸⁹⁰]*)([ᶜⁱ]{0,1})$", orb_str)
+    orb,occ,state = if !isnothing(m)
+        m[1], m[4], m[5]
+    elseif !isnothing(m2)
+        m2[1], from_superscript(m2[4]), m2[5]
+    else
+        throw(ArgumentError("Unknown subshell specification $(orb_str)"))
+    end
+    parse(O, orb) , (occ == "") ? 1 : parse(Int, occ), state_sym(state)
 end
 
 function Base.parse(::Type{Configuration{O}}, conf_str; sorted=false) where {O<:AbstractOrbital}
     isempty(conf_str) && return Configuration{O}(sorted=sorted)
     orbs = split(conf_str, r"[\. ]")
-    core_m = match(r"\[([a-zA-Z]+)\]([*ci]{0,1})", first(orbs))
+    core_m = match(r"\[([a-zA-Z]+)\]([*ciᶜⁱ]{0,1})", first(orbs))
     if !isnothing(core_m)
         core_config = core_configuration(O, core_m[1], core_m[2], sorted)
         if length(orbs) > 1
diff --git a/test/configurations.jl b/test/configurations.jl
index 7cfb853..2e5bac1 100644
--- a/test/configurations.jl
+++ b/test/configurations.jl
@@ -42,6 +42,16 @@
         @test_throws ArgumentError parse(Configuration{RelativisticOrbital}, "1s3")
         @test_throws ArgumentError parse(Configuration{Orbital}, "1s2 2p-2")
 
+        @testset "Unicode occupations/states" begin
+            for c in [c"1s2 2p6", c"[He]c 2p6", c"[He]i 2p6", c"1s 3d4", c"[Xe]*"]
+                @test parse(Configuration{Orbital}, string(c)) == c
+            end
+
+            for c in [rc"1s2 2p6", rc"1s 2p- 2p3", rc"[He]c 2p6", rc"[He]i 2p6", rc"1s 3d4", rc"[Xe]*"]
+                @test parse(Configuration{RelativisticOrbital}, string(c)) == c
+            end
+        end
+
         @testset "Spin-configurations" begin
             a = sc""
             @test a isa SpinConfiguration{<:SpinOrbital{<:Orbital}}