Merge branch '44-wrong-emission-ids' into 'main'

Resolve "Wrong emission IDs"

Closes #44

See merge request !53

examples/Getting Started.jl

 ### A Pluto.jl notebook ###
-# v0.19.36
+# v0.19.38
 
 using Markdown
 using InteractiveUtils
@@ -14,8 +14,12 @@ begin
 	# packages will be install for this notebook automatically.
 	using Semirings
 	using TensorFSTs
+	using TensorFSTs.LinearFSTs
 end
 
+# ╔═╡ d5ec1212-e6ee-42eb-8c9b-8f2234ef7f7f
+using TensorFSTs.SpeechRecognitionFSTs
+
 # ╔═╡ a4127267-40fd-4550-9446-6d82b0ae8373
 md"""# TensorFST
 *Lucas ONDEL YANG, LISN CNRS, 24/11/2023*
@@ -55,6 +59,33 @@ md"""
 The function `summary(fst)` return some basic information about your fst. 
 """
 
+# ╔═╡ 2cf7395d-4c6d-4370-aa7f-b57734c38b7b
+s = ProbSemiring{Float64}
+
+# ╔═╡ 68454444-f7ea-48b7-ac59-2aed02b2d31a
+t = tokenfst(
+	s, 
+	[(1, 1, 1), (1, 2, 1), (2, 2, 1), (2, 3, 1), (3, 3, 1)],
+	[1 => 1],
+	[3 => 1],
+	ids(SymbolTable(["a", "b", "c"]))
+)
+
+# ╔═╡ a95b0c6f-5146-4c0c-938d-1c1b7fade296
+draw(t)
+
+# ╔═╡ 5d3c18f3-d466-4c16-b54f-e21ff58b080e
+begin 
+	topo = [(1, 1, 1), (1, 2, 1), (2, 2, 1), (2, 3, 1), (3, 3, 1)]
+	values = [topo_arc[2] for topo_arc in topo]
+end
+
+# ╔═╡ 974cf260-4e3a-44d1-90f4-1b7115d32dc3
+nstates = topo[end][2]
+
+# ╔═╡ 20fadb1b-7883-474b-916e-c48cfb514c32
+
+
 # ╔═╡ f779c388-223c-4299-9523-3f3b3ba21769
 md"""
 To represent the FST as an directed acyclic graph first you can use the `draw(fst)` function.
@@ -504,6 +535,13 @@ draw(
 # ╟─b4279331-1f8c-49b7-9bc4-b46f912ace46
 # ╟─d62946ec-112e-40c4-8b8c-24dd6f1c5103
 # ╠═27d2c155-518c-49fc-b273-0853f7a04c75
+# ╠═d5ec1212-e6ee-42eb-8c9b-8f2234ef7f7f
+# ╠═2cf7395d-4c6d-4370-aa7f-b57734c38b7b
+# ╠═68454444-f7ea-48b7-ac59-2aed02b2d31a
+# ╠═a95b0c6f-5146-4c0c-938d-1c1b7fade296
+# ╠═5d3c18f3-d466-4c16-b54f-e21ff58b080e
+# ╠═974cf260-4e3a-44d1-90f4-1b7115d32dc3
+# ╠═20fadb1b-7883-474b-916e-c48cfb514c32
 # ╟─f779c388-223c-4299-9523-3f3b3ba21769
 # ╠═41df922b-543f-433e-9e30-b6863ca00366
 # ╠═e9365726-92bb-40e8-bce7-9bdbd73466db

lib/SpeechRecognitionFSTs/SpeechRecognitionFSTs.jl

@@ -16,14 +16,14 @@ export EmissionMapping, lexiconfst, tokenfst
 Default emission mapping which assumes the same number of emissions per token.
 """
 struct EmissionMapping <: AbstractMatrix{Int}
-    numstates::Int
+    numpdfs::Int
     numtokens::Int
 end
 
-Base.size(x::EmissionMapping) = (x.numstates, x.numtokens)
-Base.getindex(x::EmissionMapping, i, j) = (j - 1) * x.numstates + i
+Base.size(x::EmissionMapping) = (x.numpdfs, x.numtokens)
+Base.getindex(x::EmissionMapping, i, j) = (j - 1) * x.numpdfs + i
 Base.minimum(x::EmissionMapping) = 1
-Base.maximum(x::EmissionMapping) = (x.numstates * x.num_emission)
+Base.maximum(x::EmissionMapping) = (x.numpdfs * x.numtokens)
 
 
 
@@ -41,36 +41,75 @@ function tokenfst(
     initweights,
     finalweights,
     tokens,
-    mapping = EmissionMapping(length(topo), length(tokens))
+    mapping = nothing
 )
-    emission_count = 0
     states = Set(Int[])
     arcs, init, final = [], [], []
+    
+    if isnothing(mapping)
+        st = Set(Int[]) 
+        for _ in tokens
+            for top in topo
+                push!(st, top[1])
+            end
+        end
+        mapping = EmissionMapping(length(st), length(tokens))   
+    end
 
-    for (i, token) in enumerate(tokens)
-        offset = length(states)
-        for (j, topo_arc) in enumerate(topo)
-            emission_count += 1
+    # Extra state (infinite loop)
+    init_state = length(mapping) + 1
+    push!(states, init_state)
+    
+    for (j, token) in enumerate(tokens)
+        offset = length(states) - 1
+        
+        for topo_arc in topo
+            
             src, dest, weight = offset + topo_arc[1], offset + topo_arc[2], topo_arc[3]
 
             arc = Arc(
                 src = src,
-                isym = mapping[emission_count],
-                osym = j == 1 ? token : 1,
+                isym = mapping[topo_arc[2], j],
+                osym = 0,
                 dest = dest,
                 weight = S(weight)
             )
+            
+            # Init arc to sources
+            if topo_arc[1] == 1
+                init_arc = Arc(
+                    src = init_state,
+                    isym = 0,
+                    osym = token,
+                    dest = src,
+                    weight = S(weight)
+                )
+                push!(arcs, init_arc)
+            end
+
+            # Final arcs to destinations
+            if topo_arc[1] == length(st)
+                final_arc = Arc(
+                    src = src,
+                    isym = 0,
+                    osym = 0,
+                    dest = init_state,
+                    weight = S(weight)
+                ) 
+                push!(arcs, final_arc)
+            end
 
             push!(states, src)
             push!(states, dest)
             push!(arcs, arc)
         end
 
-        for (state, weight) in initweights
-            state = offset + state
-            push!(states, state)
-            push!(init, state => S(weight))
-        end
+        # for (state, weight) in initweights
+        #     print(state)
+        #     state = offset + state
+        #     push!(states, state)
+        #     push!(init, state => S(weight))
+        # end
 
         for (state, weight) in finalweights
             state = offset + state
@@ -79,6 +118,11 @@ function tokenfst(
         end
     end
 
+    # Actually, there is just one init state
+    for (_, weight) in initweights
+        push!(init, init_state => S(weight))
+    end
+
     TensorFST(arcs, init, final)
 end